>>Welcome to this exclusive discussion. IBM, the next three years of life sciences, innovation, precision medicine, advanced clinical data management and beyond. My name is Dave Volante from the Cuban today, we're going to take a deep dive into some of the most important trends impacting the life sciences industry in the next 60 minutes. Yeah, of course. We're going to hear how IBM is utilizing Watson and some really important in life impacting ways, but we'll also bring in real world perspectives from industry and the independent analyst view to better understand how technology and data are changing the nature of precision medicine. Now, the pandemic has created a new reality for everyone, but especially for life sciences companies, one where digital transformation is no longer an option, but a necessity. Now the upside is the events of the past 22 months have presented an accelerated opportunity for innovation technology and real world data are coming together and being applied to support life science, industry trends and improve drug discovery, clinical development, and treatment commercialization throughout the product life cycle cycle. Now I'd like to introduce our esteemed panel. Let me first introduce Lorraine Marshawn, who is general manager of life sciences at IBM Watson health. Lorraine leads the organization dedicated to improving clinical development research, showing greater treatment value in getting treatments to patients faster with differentiated solutions. Welcome Lorraine. Great to see you. >>Dr. Namita LeMay is the research vice-president of IDC, where she leads the life sciences R and D strategy and technology program, which provides research based advisory and consulting services as well as market analysis. The loan to meta thanks for joining us today. And our third panelist is Greg Cunningham. Who's the director of the RWE center of excellence at Eli Lilly and company. Welcome, Greg, you guys are doing some great work. Thanks for being here. Thanks >>Dave. >>Now today's panelists are very passionate about their work. If you'd like to ask them a question, please add it to the chat box located near the bottom of your screen, and we'll do our best to answer them all at the end of the panel. Let's get started. Okay, Greg, and then Lorraine and meta feel free to chime in after one of the game-changers that you're seeing, which are advancing precision medicine. And how do you see this evolving in 2022 and into the next decade? >>I'll give my answer from a life science research perspective. The game changer I see in advancing precision medicine is moving from doing research using kind of a single gene mutation or kind of a single to look at to doing this research using combinations of genes and the potential that this brings is to bring better drug targets forward, but also get the best product to a patient faster. Um, I can give, uh, an example how I see it playing out in the last decade. Non-oncology real-world evidence. We've seen an evolution in precision medicine as we've built out the patient record. Um, as we've done that, uh, the marketplace has evolved rapidly, uh, with, particularly for electronic medical record data and genomic data. And we were pretty happy to get our hands on electronic medical record data in the early days. And then later the genetic test results were combined with this data and we could do research looking at a single mutation leading to better patient outcomes. But I think where we're going to evolve in 2022 and beyond is with genetic testing, growing and oncology, providing us more data about that patient. More genes to look at, uh, researchers can look at groups of genes to analyze, to look at that complex combination of gene mutations. And I think it'll open the door for things like using artificial intelligence to help researchers plow through the complex number of permutations. When you think about all those genes you can look at in combination, right? Lorraine yes. Data and machine intelligence coming together, anything you would add. >>Yeah. Thank you very much. Well, I think that Greg's response really sets us up nicely, particularly when we think about the ability to utilize real-world data in the farm industry across a number of use cases from discovery to development to commercial, and, you know, in particular, I think with real world data and the comments that Greg just made about clinical EMR data linked with genetic or genomic data, a real area of interest in one that, uh, Watson health in particular is focused on the idea of being able to create a data exchange so that we can bring together claims clinical EMR data, genomics data, increasingly wearables and data directly from patients in order to create a digital health record that we like to call an intelligent patient health record that basically gives us the digital equivalent of a real life patient. And these can be used in use cases in randomized controlled clinical trials for synthetic control arms or natural history. They can be used in order to track patients' response to drugs and look at outcomes after they've been on various therapies as, as Greg is speaking to. And so I think that, you know, the promise of data and technology, the AI that we can apply on that is really helping us advance, getting therapies to market faster, with better information, lower sample sizes, and just a much more efficient way to do drug development and to track and monitor outcomes in patients. >>Great. Thank you for that now to meta, when I joined IDC many, many years ago, I really didn't know much about the industry that I was covering, but it's great to see you as a former practitioner now bringing in your views. What do you see as the big game-changers? >>So, um, I would, I would agree with what both Lorraine and Greg said. Um, but one thing that I'd just like to call out is that, you know, everyone's talking about big data, the volume of data is growing. It's growing exponentially actually about, I think 30% of data that exists today is healthcare data. And it's growing at a rate of 36%. That's huge, but then it's not just about the big, it's also about the broad, I think, um, you know, I think great points that, uh, Lorraine and Greg brought out that it's, it's not just specifically genomic data, it's multi omic data. And it's also about things like medical history, social determinants of health, behavioral data. Um, and why, because when you're talking about precision medicine and we know that we moved away from the, the terminology of personalized to position, because you want to talk about disease stratification and you can, it's really about convergence. >>Um, if you look at a recent JAMA paper in 2021, only 1% of EHS actually included genomic data. So you really need to have that ability to look at data holistically and IDC prediction is seeing that investments in AI to fuel in silico, silicone drug discovery will double by 20, 24, but how are you actually going to integrate all the different types of data? Just look at, for example, diabetes, you're on type two diabetes, 40 to 70% of it is genetically inherited and you have over 500 different, uh, genetic low side, which could be involved in playing into causing diabetes. So the earlier strategy, when you are looking at, you know, genetic risk scoring was really single trait. Now it's transitioning to multi rate. And when you say multi trade, you really need to get that integrated view that converging for you to, to be able to drive a precision medicine strategy. So to me, it's a very interesting contrast on one side, you're really trying to make it specific and focused towards an individual. And on the other side, you really have to go wider and bigger as well. >>Uh, great. I mean, the technology is enabling that convergence and the conditions are almost mandating it. Let's talk about some more about data that the data exchange and building an intelligent health record, as it relates to precision medicine, how will the interoperability of real-world data, you know, create that more cohesive picture for the, for the patient maybe Greg, you want to start, or anybody else wants to chime in? >>I think, um, the, the exciting thing from, from my perspective is the potential to gain access to data. You may be weren't aware of an exchange in implies that, uh, some kind of cataloging, so I can see, uh, maybe things that might, I just had no idea and, uh, bringing my own data and maybe linking data. These are concepts that I think are starting to take off in our field, but it, it really opens up those avenues to when you, you were talking about data, the robustness and richness volume isn't, uh, the only thing is Namita said, I think really getting to a rich high-quality data and, and an exchange offers a far bigger, uh, range for all of us to, to use, to get our work done. >>Yeah. And I think, um, just to chime, chime into that, uh, response from Greg, you know, what we hear increasingly, and it's pretty pervasive across the industry right now, because this ability to create an exchange or the intelligent, uh, patient health record, these are new ideas, you know, they're still rather nascent and it always is the operating model. Uh, that, that is the, uh, the difficult challenge here. And certainly that is the case. So we do have data in various silos. Uh, they're in patient claims, they're in electronic medical records, they might be in labs, images, genetic files on your smartphone. And so one of the challenges with this interoperability is being able to tap into these various sources of data, trying to identify quality data, as Greg has said, and the meta is underscoring as well. Uh, we've gotta be able to get to the depth of data that's really meaningful to us, but then we have to have technology that allows us to pull this data together. >>First of all, it has to be de-identified because of security and patient related needs. And then we've gotta be able to link it so that you can create that likeness in terms of the record, it has to be what we call cleaned or curated so that you get the noise and all the missing this out of it, that's a big step. And then it needs to be enriched, which means that the various components that are going to be meaningful, you know, again, are brought together so that you can create that cohort of patients, that individual patient record that now is useful in so many instances across farm, again, from development, all the way through commercial. So the idea of this exchange is to enable that exact process that I just described to have a, a place, a platform where various entities can bring their data in order to have it linked and integrated and cleaned and enriched so that they get something that is a package like a data package that they can actually use. >>And it's easy to plug into their, into their studies or into their use cases. And I think a really important component of this is that it's gotta be a place where various third parties can feel comfortable bringing their data together in order to match it with other third parties. That is a, a real value, uh, that the industry is increasingly saying would be important to them is, is the ability to bring in those third-party data sets and be able to link them and create these, these various data products. So that's really the idea of the data exchange is that you can benefit from accessing data, as Greg mentioned in catalogs that maybe are across these various silos so that you can do the kind of work that you need. And that we take a lot of the hard work out of it. I like to give an example. >>We spoke with one of our clients at one of the large pharma companies. And, uh, I think he expressed it very well. He said, what I'd like to do is have like a complete dataset of lupus. Lupus is an autoimmune condition. And I've just like to have like the quintessential lupus dataset that I can use to run any number of use cases across it. You know, whether it's looking at my phase one trial, whether it's selecting patients and enriching for later stage trials, whether it's understanding patient responses to different therapies as I designed my studies. And so, you know, this idea of adding in therapeutic area indication, specific data sets and being able to create that for the industry in the meta mentioned, being able to do that, for example, in diabetes, that's how pharma clients need to have their needs met is through taking the hard workout, bringing the data together, having it very therapeutically enriched so that they can use it very easily. >>Thank you for that detail and the meta. I mean, you can't do this with humans at scale in technology of all the things that Lorraine was talking about, the enrichment, the provenance, the quality, and of course, it's got to be governed. You've got to protect the privacy privacy humans just can't do all that at massive scale. Can it really tech that's where technology comes in? Doesn't it and automation. >>Absolutely. >>I, couldn't more, I think the biggest, you know, whether you talk about precision medicine or you talk about decentralized trials, I think there's been a lot of hype around these terms, but what is really important to remember is technology is the game changer and bringing all that data together is really going to be the key enabler. So multimodal data integration, looking at things like security or federated learning, or also when you're talking about leveraging AI, you're not talking about things like bias or other aspects around that are, are critical components that need to be addressed. I think the industry is, uh, it's partly, still trying to figure out the right use cases. So it's one part is getting together the data, but also getting together the right data. Um, I think data interoperability is going to be the absolute game changer for enabling this. Uh, but yes, um, absolutely. I can, I can really couldn't agree more with what Lorraine just said, that it's bringing all those different aspects of data together to really drive that precision medicine strategy. >>Excellent. Hey Greg, let's talk about protocols decentralized clinical trials. You know, they're not new to life silences, but, but the adoption of DCTs is of course sped up due to the pandemic we've had to make trade-offs obviously, and the risk is clearly worth it, but you're going to continue to be a primary approach as we enter 2022. What are the opportunities that you see to improve? How DCTs are designed and executed? >>I see a couple opportunities to improve in this area. The first is, uh, back to technology. The infrastructure around clinical trials has, has evolved over the years. Uh, but now you're talking about moving away from kind of site focus to the patient focus. Uh, so with that, you have to build out a new set of tools that would help. So for example, one would be novel trial, recruitment, and screening, you know, how do you, how do you find patients and how do you screen them to see if are they, are they really a fit for, for this protocol? Another example, uh, very important documents that we have to get is, uh, you know, the e-consent that someone's says, yes, I'm, well, I understand this study and I'm willing to do it, have to do that in a more remote way than, than we've done in the past. >>Um, the exciting area, I think, is the use of, uh, eco, uh, E-Pro where we capture data from the patient using apps, devices, sensors. And I think all of these capabilities will bring a new way of, of getting data faster, uh, in, in this kind of model. But the exciting thing from, uh, our perspective at Lily is it's going to bring more data about the patient from the patient, not just from the healthcare provider side, it's going to bring real data from these apps, devices and sensors. The second thing I think is using real-world data to identify patients, to also improve protocols. We run scenarios today, looking at what's the impact. If you change a cut point on a, a lab or a biomarker to see how that would affect, uh, potential enrollment of patients. So it, it definitely the real-world data can be used to, to make decisions, you know, how you improve these protocols. >>But the thing that we've been at the challenge we've been after that this probably offers the biggest is using real-world data to identify patients as we move away from large academic centers that we've used for years as our sites. Um, you can maybe get more patients who are from the rural areas of our countries or not near these large, uh, uh, academic centers. And we think it'll bring a little more diversity to the population, uh, who who's, uh, eligible, but also we have their data, so we can see if they really fit the criteria and the probability they are a fit for the trial is much higher than >>Right. Lorraine. I mean, your clients must be really pushing you to help them improve DCTs what are you seeing in the field? >>Yes, in fact, we just attended the inaugural meeting of the de-central trials research Alliance in, uh, in Boston about two weeks ago where, uh, all of the industry came together, pharma companies, uh, consulting vendors, just everyone who's been in this industry working to help define de-central trials and, um, think through what its potential is. Think through various models in order to enable it, because again, a nascent concept that I think COVID has spurred into action. Um, but it is important to take a look at the definition of DCT. I think there are those entities that describe it as accessing data directly from the patient. I think that is a component of it, but I think it's much broader than that. To me, it's about really looking at workflows and processes of bringing data in from various remote locations and enabling the whole ecosystem to work much more effectively along the data continuum. >>So a DCT is all around being able to make a site more effective, whether it's being able to administer a tele visit or the way that they're getting data into the electronic data captures. So I think we have to take a look at the, the workflows and the operating models for enabling de-central trials and a lot of what we're doing with our own technology. Greg mentioned the idea of electronic consent of being able to do electronic patient reported outcomes, other collection of data directly from the patient wearables tele-health. So these are all data acquisition, methodologies, and technologies that, that we are enabling in order to get the best of the data into the electronic data capture system. So edit can be put together and processed and submitted to the FDA for regulatory use for clinical trial type submission. So we're working on that. I think the other thing that's happening is the ability to be much more flexible and be able to have more cloud-based storage allows you to be much more inter-operable to allow API APIs in order to bring in the various types of data. >>So we're really looking at technology that can make us much more fluid and flexible and accommodating to all the ways that people live and work and manage their health, because we have to reflect that in the way we collect those data types. So that's a lot of what we're, what we're focused on. And in talking with our clients, we spend also a lot of time trying to understand along the, let's say de-central clinical trials continuum, you know, w where are they? And I know Namita is going to talk a little bit about research that they've done in terms of that adoption curve, but because COVID sort of forced us into being able to collect data in more remote fashion in order to allow some of these clinical trials to continue during COVID when a lot of them had to stop. What we want to make sure is that we understand and can codify some of those best practices and that we can help our clients enable that because the worst thing that would happen would be to have made some of that progress in that direction. >>But then when COVID is over to go back to the old ways of doing things and not bring some of those best practices forward, and we actually hear from some of our clients in the pharma industry, that they worry about that as well, because we don't yet have a system for operationalizing a de-central trial. And so we really have to think about the protocol it's designed, the indication, the types of patients, what makes sense to decentralize, what makes sense to still continue to collect data in a more traditional fashion. So we're spending a lot of time advising and consulting with our patients, as well as, I mean, with our clients, as well as CRS, um, on what the best model is in terms of their, their portfolio of studies. And I think that's a really important aspect of trying to accelerate the adoption is making sure that what we're doing is fit for purpose, just because you can use technology doesn't mean you should, it really still does require human beings to think about the problem and solve them in a very practical way. >>Great, thank you for that. Lorraine. I want to pick up on some things that Lorraine was just saying. And then back to what Greg was saying about, uh, uh, DCTs becoming more patient centric, you had a prediction or IDC, did I presume your fingerprints were on it? Uh, that by 20 25, 70 5% of trials will be patient-centric decentralized clinical trials, 90% will be hybrid. So maybe you could help us understand that relationship and what types of innovations are going to be needed to support that evolution of DCT. >>Thanks, Dave. Yeah. Um, you know, sorry, I, I certainly believe that, uh, you know, uh, Lorraine was pointing out of bringing up a very important point. It's about being able to continue what you have learned in over the past two years, I feel this, you know, it was not really a digital revolution. It was an attitude. The revolution that this industry underwent, um, technology existed just as clinical trials exist as drugs exist, but there was a proof of concept that technology works that this model is working. So I think that what, for example, telehealth, um, did for, for healthcare, you know, transition from, from care, anywhere care, anytime, anywhere, and even becoming predictive. That's what the decentralized clinical trials model is doing for clinical trials today. Great points again, that you have to really look at where it's being applied. You just can't randomly apply it across clinical trials. >>And this is where the industry is maturing the complexity. Um, you know, some people think decentralized trials are very simple. You just go and implement these centralized clinical trials, but it's not that simple as it it's being able to define, which are the right technologies for that specific, um, therapeutic area for that specific phase of the study. It's being also a very important point is bringing in the patient's voice into the process. Hey, I had my first telehealth visit sometime last year and I was absolutely thrilled about it. I said, no time wasted. I mean, everything's done in half an hour, but not all patients want that. Some want to consider going back and you, again, need to customize your de-centralized trials model to, to the, to the type of patient population, the demographics that you're dealing with. So there are multiple factors. Um, also stepping back, you know, Lorraine mentioned they're consulting with, uh, with their clients, advising them. >>And I think a lot of, um, a lot of companies are still evolving in their maturity in DCTs though. There's a lot of boys about it. Not everyone is very mature in it. So it's, I think it, one thing everyone's kind of agreeing with is yes, we want to do it, but it's really about how do we go about it? How do we make this a flexible and scalable modern model? How do we integrate the patient's voice into the process? What are the KPIs that we define the key performance indicators that we define? Do we have a playbook to implement this model to make it a scalable model? And, you know, finally, I think what organizations really need to look at is kind of developing a de-centralized mature maturity scoring model, so that I assess where I am today and use that playbook to define, how am I going to move down the line to me reach the next level of maturity. Those were some of my thoughts. Right? >>Excellent. And now remember you, if you have any questions, use the chat box below to submit those questions. We have some questions coming in from the audience. >>At one point to that, I think one common thread between the earlier discussion around precision medicine and around decentralized trials really is data interoperability. It is going to be a big game changer to, to enable both of these pieces. Sorry. Thanks, Dave. >>Yeah. Thank you. Yeah. So again, put your questions in the chat box. I'm actually going to go to one of the questions from the audience. I get some other questions as well, but when you think about all the new data types that are coming in from social media, omics wearables. So the question is with greater access to these new types of data, what trends are you seeing from pharma device as far as developing capabilities to effectively manage and analyze these novel data types? Is there anything that you guys are seeing, um, that you can share in terms of best practice or advice >>I'll offer up? One thing, I think the interoperability isn't quite there today. So, so what's that mean you can take some of those data sources. You mentioned, uh, some Omix data with, uh, some health claims data and it's the, we spend too much time and in our space putting data to gather the behind the scenes, I think the stat is 80% of the time is assembling the data 20% analyzing. And we've had conversations here at Lilly about how do we get to 80% of the time is doing analysis. And it really requires us to think, take a step back and think about when you create a, uh, a health record, you really have to be, have the same plugins so that, you know, data can be put together very easily, like Lorraine mentioned earlier. And that comes back to investing in as an industry and standards so that, you know, you have some of data standard, we all can agree upon. And then those plugs get a lot easier and we can spend our time figuring out how to make, uh, people's lives better with healthcare analysis versus putting data together, which is not a lot of fun behind the scenes. >>Other thoughts on, um, on, on how to take advantage of sort of novel data coming from things like devices in the nose that you guys are seeing. >>I could jump in there on your end. Did you want to go ahead? Okay. So, uh, I mean, I think there's huge value that's being seen, uh, in leveraging those multiple data types. I think one area you're seeing is the growth of prescription digital therapeutics and, um, using those to support, uh, you know, things like behavioral health issues and a lot of other critical conditions it's really taking you again, it is interlinking real-world data cause it's really taking you to the patient's home. Um, and it's, it's, there's a lot of patients in the city out here cause you can really monitor the patient real-time um, without the patient having coming, you know, coming and doing a site visit once in say four weeks or six weeks. So, um, I, and, uh, for example, uh, suicidal behavior and just to take an example, if you can predict well in advance, based on those behavioral parameters, that this is likely to trigger that, uh, the value of it is enormous. Um, again, I think, uh, Greg made a valid point about the industry still trying to deal with resolving the data interoperability issue. And there are so many players that are coming in the industry right now. There are really few that have the maturity and the capability to address these challenges and provide intelligence solutions. >>Yeah. Maybe I'll just, uh, go ahead and, uh, and chime into Nikita's last comment there. I think that's what we're seeing as well. And it's very common, you know, from an innovation standpoint that you have, uh, a nascent industry or a nascent innovation sort of situation that we have right now where it's very fragmented. You have a lot of small players, you have some larger entrenched players that have the capability, um, to help to solve the interoperability challenge, the standards challenge. I mean, I think IBM Watson health is certainly one of the entities that has that ability and is taking a stand in the industry, uh, in order to, to help lead in that way. Others are too. And, uh, but with, with all of the small companies that are trying to find interesting and creative ways to gather that data, it does create a very fragmented, uh, type of environment and ecosystem that we're in. >>And I think as we mature, as we do come forward with the KPIs, the operating models, um, because you know, the devil's in the detail in terms of the operating models, it's really exciting to talk these trends and think about the future state. But as Greg pointed out, if you're spending 80% of your time just under the hood, you know, trying to get the engine, all the spark plugs to line up, um, that's, that's just hard grunt work that has to be done. So I think that's where we need to be focused. And I think bringing all the data in from these disparate tools, you know, that's fine, we need, uh, a platform or the API APIs that can enable that. But I think as we, as we progress, we'll see more consolidation, uh, more standards coming into play, solving the interoperability types of challenges. >>And, um, so I think that's where we should, we should focus on what it's going to take and in three years to really codify this and make it, so it's a, it's a well hum humming machine. And, you know, I do know having also been in pharma that, uh, there's a very pilot oriented approach to this thing, which I think is really healthy. I think large pharma companies tend to place a lot of bets with different programs on different tools and technologies, to some extent to see what's gonna stick and, you know, kind of with an innovation mindset. And I think that's good. I think that's kind of part of the process of figuring out what is going to work and, and helping us when we get to that point of consolidating our model and the technologies going forward. So I think all of the efforts today are definitely driving us to something that feels much more codified in the next three to five years. >>Excellent. We have another question from the audience it's sort of related to the theme of this discussion, given the FDA's recent guidance on using claims and electronic health records, data to support regulatory decision-making what advancements do you think we can expect with regards to regulatory use of real-world data in the coming years? It's kind of a two-parter so maybe you guys can collaborate on this one. What role that, and then what role do you think industry plays in influencing innovation within the regulatory space? >>All right. Well, it looks like you've stumped the panel there. Uh, Dave, >>It's okay to take some time to think about it, right? You want me to repeat it? You guys, >>I, you know, I I'm sure that the group is going to chime into this. I, so the FDA has issued a guidance. Um, it's just, it's, it's exactly that the FDA issues guidances and says that, you know, it's aware and supportive of the fact that we need to be using real-world data. We need to create the interoperability, the standards, the ways to make sure that we can include it in regulatory submissions and the like, um, and, and I sort of think about it akin to the critical path initiative, probably, I don't know, 10 or 12 years ago in pharma, uh, when the FDA also embrace this idea of the critical path and being able to allow more in silico modeling of clinical trial, design and development. And it really took the industry a good 10 years, um, you know, before they were able to actually adopt and apply and take that sort of guidance or openness from the FDA and actually apply it in a way that started to influence the way clinical trials were designed or the in silico modeling. >>So I think the second part of the question is really important because while I think the FDA is saying, yes, we recognize it's important. Uh, we want to be able to encourage and support it. You know, when you look for example, at synthetic control arms, right? The use of real-world data in regulatory submissions over the last five or six years, all of the use cases have been in oncology. I think there've been about maybe somewhere between eight to 10 submissions. And I think only one actually was a successful submission, uh, in all those situations, the real-world data arm of that oncology trial that synthetic control arm was actually rejected by the FDA because of lack of completeness or, you know, equalness in terms of the data. So the FDA is not going to tell us how to do this. So I think the second part of the question, which is what's the role of industry, it's absolutely on industry in order to figure out exactly what we're talking about, how do we figure out the interoperability, how do we apply the standards? >>How do we ensure good quality data? How do we enrich it and create the cohort that is going to be equivalent to the patient in the real world, uh, in the end that would otherwise be in the clinical trial and how do we create something that the FDA can agree with? And we'll certainly we'll want to work with the FDA in order to figure out this model. And I think companies are already doing that, but I think that the onus is going to be on industry in order to figure out how you actually operationalize this and make it real. >>Excellent. Thank you. Um, question on what's the most common misconception that clinical research stakeholders with sites or participants, et cetera might have about DCTs? >>Um, I could jump in there. Right. So, sure. So, um, I think in terms of misconceptions, um, I think the communist misconceptions that sites are going away forever, which I do not think is really happening today. Then the second, second part of it is that, um, I think also the perspective that patients are potentially neglected because they're moving away. So we'll pay when I, when I, what I mean by that neglected, perhaps it was not the appropriate term, but the fact that, uh, will patients will, will, will patient engagement continue, will retention be strong since the patients are not interacting in person with the investigator quite as much. Um, so site retention and patient retention or engagement from both perspectives, I think remains a concern. Um, but actually if you look at, uh, look at, uh, assessments that have been done, I think patients are more than happy. >>Majority of the patients have been really happy about, about the new model. And in fact, sites are, seem to increase, have increased investments in technology by 50% to support this kind of a model. So, and the last thing is that, you know, decentralized trials is a great model and it can be applied to every possible clinical trial. And in another couple of weeks, the whole industry will be implementing only decentralized trials. I think we are far away from that. It's just not something that you would implement across every trial. And we discussed that already. So you have to find the right use cases for that. So I think those were some of the key misconceptions I'd say in the industry right now. Yeah. >>Yeah. And I would add that the misconception I hear the most about is, uh, the, the similar to what Namita said about the sites and healthcare professionals, not being involved to the level that they are today. Uh, when I mentioned earlier in our conversation about being excited about capturing more data, uh, from the patient that was always in context of, in addition to, you know, healthcare professional opinion, because I think both of them bring that enrichment and a broader perspective of that patient experience, whatever disease they're faced with. So I, I think some people think is just an all internet trial with just someone, uh, putting out there their own perspective. And, and it's, it's a combination of both to, to deliver a robust data set. >>Yeah. Maybe I'll just comment on, it reminds me of probably 10 or 15 years ago, maybe even more when, um, really remote monitoring was enabled, right? So you didn't have to have the study coordinator traveled to the investigative site in order to check the temperature of the freezer and make sure that patient records were being completed appropriately because they could have a remote visit and they could, they could send the data in a via electronic data and do the monitoring visit, you know, in real time, just the way we're having this kind of communication here. And there was just so much fear that you were going to replace or supplant the personal relationship between the sites between the study coordinators that you were going to, you know, have to supplant the role of the monitor, which was always a very important role in clinical trials. >>And I think people that really want to do embrace the technology and the advantages that it provided quickly saw that what it allowed was the monitor to do higher value work, you know, instead of going in and checking the temperature on a freezer, when they did have their visit, they were able to sit and have a quality discussion for example, about how patient recruitment was going or what was coming up in terms of the consent. And so it created a much more high touch, high quality type of interaction between the monitor and the investigative site. And I think we should be looking for the same advantages from DCT. We shouldn't fear it. We shouldn't think that it's going to supplant the site or the investigator or the relationship. It's our job to figure out where the technology fits and clinical sciences always got to be high touch combined with high-tech, but the high touch has to lead. And so getting that balance right? And so that's going to happen here as well. We will figure out other high value work, meaningful work for the site staff to do while they let the technology take care of the lower quality work, if you will, or the lower value work, >>That's not an, or it's an, and, and you're talking about the higher value work. And it, it leads me to something that Greg said earlier about the 80, 20, 80% is assembly. 20% is actually doing the analysis and that's not unique to, to, to life sciences, but, but sort of question is it's an organizational question in terms of how we think about data and how we approach data in the future. So Bamyan historically big data in life sciences in any industry really is required highly centralized and specialized teams to do things that the rain was talking about, the enrichment, the provenance, the data quality, the governance, the PR highly hyper specialized teams to do that. And they serve different constituencies. You know, not necessarily with that, with, with context, they're just kind of data people. Um, so they have responsibility for doing all those things. Greg, for instance, within literally, are you seeing a move to, to, to democratize data access? We've talked about data interoperability, part of that state of sharing, um, that kind of breaks that centralized hold, or is that just too far in the future? It's too risky in this industry? >>Uh, it's actually happening now. Uh, it's a great point. We, we try to classify what people can do. And, uh, the example would be you give someone who's less analytically qualified, uh, give them a dashboard, let them interact with the data, let them better understand, uh, what, what we're seeing out in the real world. Uh, there's a middle user, someone who you could give them, they can do some analysis with the tool. And the nice thing with that is you have some guardrails around that and you keep them in their lane, but it allows them to do some of their work without having to go ask those centralized experts that, that you mentioned their precious resources. And that's the third group is those, uh, highly analytical folks that can, can really deliver, uh, just value beyond. But when they're doing all those other things, uh, it really hinders them from doing what we've been talking about is the high value stuff. So we've, we've kind of split into those. We look at people using data in one of those three lanes and it, and it has helped I think, uh, us better not try to make a one fit solution for, for how we deliver data and analytic tools for people. Right. >>Okay. I mean, DCT hot topic with the, the, the audience here. Another question, um, what capabilities do sponsors and CRS need to develop in-house to pivot toward DCT? >>Should I jump in here? Yeah, I mean, um, I think, you know, when, when we speak about DCTs and when I speak with, uh, folks around in the industry, I, it takes me back to the days of risk-based monitoring. When it was first being implemented, it was a huge organizational change from the conventional monitoring models to centralize monitoring and risk-based monitoring, it needs a mental reset. It needs as Lorraine had pointed out a little while ago, restructuring workflows, re redefining processes. And I think that is one big piece. That is, I think the first piece, when, you know, when you're implementing a new model, I think organizational change management is a big piece of it because you are disturbing existing structures, existing methods. So getting that buy-in across the organization towards the new model, seeing what the value add in it. And where do you personally fit into that story? >>How do your workflows change, or how was your role impacted? I think without that this industry will struggle. So I see organizations, I think, first trying to work on that piece to build that in. And then of course, I also want to step back for the second to the, uh, to the point that you brought out about data democratization. And I think Greg Greg gave an excellent point, uh, input about how it's happening in the industry. But I would also say that the data democratization really empowerment of, of, of the stakeholders also includes the sites, the investigators. So what is the level of access to data that you know, that they have now, and is it, uh, as well as patients? So see increasingly more and more companies trying to provide access to patients finally, it's their data. So why shouldn't they have some insights to it, right. So access to patients and, uh, you know, the 80, 20 part of it. Uh, yes, he's absolutely right that, uh, we want to see that flip from, uh, 20%, um, you know, focusing on, on actually integrating the data 80% of analytics, but the real future will be coming in when actually the 20 and 18 has gone. And you actually have analysts the insights out on a silver platter. That's kind of wishful thinking, some of the industries is getting there in small pieces, but yeah, then that's just why I should, why we share >>Great points. >>And I think that we're, we're there in terms that like, I really appreciate the point around democratizing the data and giving the patient access ownership and control over their own data. I mean, you know, we see the health portals that are now available for patients to view their own records, images, and labs, and claims and EMR. We have blockchain technology, which is really critical here in terms of the patient, being able to pull all of their own data together, you know, in the blockchain and immutable record that they can own and control if they want to use that to transact clinical trial types of opportunities based on their data, they can, or other real world scenarios. But if they want to just manage their own data because they're traveling and if they're in a risky health situation, they've got their own record of their health, their health history, uh, which can avoid, you know, medical errors occurring. So, you know, even going beyond life sciences, I think this idea of democratizing data is just good for health. It's just good for people. And we definitely have the technology that can make it a reality. Now >>You're here. We have just about 10 minutes left and now of course, now all the questions are rolling in like crazy from the crowd. Would it be curious to know if there would be any comments from the panel on cost comparison analysis between traditional clinical trials in DCTs and how could the outcome effect the implementation of DCTs any sort of high-level framework you can share? >>I would say these are still early days to, to drive that analysis because I think many companies are, um, are still in the early stages of implementation. They've done a couple of trials. The other part of it that's important to keep in mind is, um, is for organizations it's, they're at a stage of, uh, of being on the learning curve. So when you're, you're calculating the cost efficiencies, if ideally you should have had two stakeholders involved, you could have potentially 20 stakeholders involved because everyone's trying to learn the process and see how it's going to be implemented. So, um, I don't think, and the third part of it, I think is organizations are still defining their KPIs. How do you measure it? What do you measure? So, um, and even still plugging in the pieces of technology that they need to fit in, who are they partnering with? >>What are the pieces of technology they're implementing? So I don't think there is a clear cut as answered at this stage. I think as you scale this model, the efficiencies will be seen. It's like any new technology or any new solution that's implemented in the first stages. It's always a little more complex and in fact sometimes costs extra. But as, as you start scaling it, as you establish your workflows, as you streamline it, the cost efficiencies will start becoming evident. That's why the industry is moving there. And I think that's how it turned out on the long run. >>Yeah. Just make it maybe out a comment. If you don't mind, the clinical trials are, have traditionally been costed are budgeted is on a per patient basis. And so, you know, based on the difficulty of the therapeutic area to recruit a rare oncology or neuromuscular disease, there's an average that it costs in order to find that patient and then execute the various procedures throughout the clinical trial on that patient. And so the difficulty of reaching the patient and then the complexity of the trial has led to what we might call a per patient stipend, which is just the metric that we use to sort of figure out what the average cost of a trial will be. So I think to point, we're going to have to see where the ability to adjust workflows, get to patients faster, collect data more easily in order to make the burden on the site, less onerous. I think once we start to see that work eases up because of technology, then I think we'll start to see those cost equations change. But I think right now the system isn't designed in order to really measure the economic benefit of de-central models. And I think we're going to have to sort of figure out what that looks like as we go along and since it's patient oriented right now, we'll have to say, well, you know, how does that work, ease up? And to those costs actually come down and then >>Just scale, it's going to be more, more clear as the media was saying, next question from the audiences, it's kind of a best fit question. You all have touched on this, but let me just ask it is what examples in which, in which phases suit DCT in its current form, be it fully DCT or hybrid models, none of our horses for courses question. >>Well, I think it's kind of, uh, it's, it's it's has its efficiencies, obviously on the later phases, then the absolute early phase trials, those are not the ideal models for DCTs I would say so. And again, the logic is also the fact that, you know, when you're, you're going into the later phase trials, the volume of number of patients is increasing considerably to the point that Lorraine brought up about access to the patients about patient selection. The fact, I think what one should look at is really the advantages that it brings in, in terms of, you know, patient access in terms of patient diversity, which is a big piece that, um, the cities are enabling. So, um, if you, if, if you, if you look at the spectrum of, of these advantages and, and just to step back for a moment, if you, if you're looking at costs, like you're looking at things like remote site monitoring, um, is, is a big, big plus, right? >>I mean, uh, site monitoring alone accounts for around a third of the trial costs. So there are so many pieces that fall in together. The challenge actually that comes when you're in defining DCTs and there are, as Rick pointed out multiple definitions of DCTs that are existing, uh, you know, in the industry right now, whether you're talking of what Detroit is doing, or you're talking about acro or Citi or others. But the point is it's a continuum, it's a continuum of different pieces that have been woven together. And so how do you decide which pieces you're plugging in and how does that impact the total cost or the solution that you're implementing? >>Great, thank you. Last question we have in the audience, excuse me. What changes have you seen? Are there others that you can share from the FDA EU APAC, regulators and supporting DCTs precision medicine for approval processes, anything you guys would highlight that we should be aware of? >>Um, I could quickly just add that. I think, um, I'm just publishing a report on de-centralized clinical trials should be published shortly, uh, perspective on that. But I would say that right now, um, there, there was a, in the FDA agenda, there was a plan for a decentralized clinical trials guidance, as far as I'm aware, one has not yet been published. There have been significant guidances that have been published both by email and by, uh, the FDA that, um, you know, around the implementation of clinical trials during the COVID pandemic, which incorporate various technology pieces, which support the DCD model. Um, but I, and again, I think one of the reasons why it's not easy to publish a well-defined guidance on that is because there are so many moving pieces in it. I think it's the Danish, uh, regulatory agency, which has per se published a guidance and revised it as well on decentralized clinical trials. >>Right. Okay. Uh, we're pretty much out of time, but I, I wonder Lorraine, if you could give us some, some final thoughts and bring us home things that we should be watching or how you see the future. >>Well, I think first of all, let me, let me thank the panel. Uh, we really appreciate Greg from Lily and the meta from IDC bringing their perspectives to this conversation. And, uh, I hope that the audience has enjoyed the, uh, the discussion that we've had around the future state of real world data as, as well as DCT. And I think, you know, some of the themes that we've talked about, number one, I think we have a vision and I think we have the right strategies in terms of the future promise of real-world data in any number of different applications. We certainly have talked about the promise of DCT to be more efficient, to get us closer to the patient. I think that what we have to focus on is how we come together as an industry to really work through these very vexing operational issues, because those are always the things that hang us up and whether it's clinical research or whether it's later stage, uh, applications of data. >>We, the healthcare system is still very fragmented, particularly in the us. Um, it's still very, state-based, uh, you know, different states can have different kinds of, uh, of, of cultures and geographic, uh, delineations. And so I think that, you know, figuring out a way that we can sort of harmonize and bring all of the data together, bring some of the models together. I think that's what you need to look to us to do both industry consulting organizations, such as IBM Watson health. And we are, you know, through DTRA and, and other, uh, consortia and different bodies. I think we're all identifying what the challenges are in terms of making this a reality and working systematically on those. >>It's always a pleasure to work with such great panelists. Thank you, Lorraine Marshawn, Dr. Namita LeMay, and Greg Cunningham really appreciate your participation today and your insights. The next three years of life sciences, innovation, precision medicine, advanced clinical data management and beyond has been brought to you by IBM in the cube. You're a global leader in high tech coverage. And while this discussion has concluded, the conversation continues. So please take a moment to answer a few questions about today's panel on behalf of the entire IBM life sciences team and the cube decks for your time and your feedback. And we'll see you next time.

from Yorktown Heights New York it's the queue coverings IBM cloud innovation be brought to you by IBM hi I'm Peter Burris and we have wrapped our the cubes coverage of IBM innovation day here at the Thomas J Watson Research Center in Yorktown Heights now for anybody that's been in the industry for a while you know that this is one of the mecca's of the computing industry this is where an enormous number of innovations have taken place innovations about relating to semiconductor processes and you know CPU architectures innovations relating to middleware and innovations relating to database management and very importantly innovations relating to how customers and companies engage to be more successful with technology and in many respects that's really what's happening with the overall drive to cloud is to bring closer together that invention that takes place and pushes forward what technology can do and then a delivery model that's focused on ensuring the customers can actually more easily do it and IBM is absolutely part of that conversation we'll be going forward especially as we think about how those high-value legacy applications are going to be employed within a cloud context to further drive transaction capabilities with event capabilities in the cloud we've had some great conversations we've heard for example from Hilary hunter who's a CTO here at cloud infrastructure about the new role that opend plays within innovation how IBM is trying to further leverage that with the Red Hat acquisition we've had great conversations with Jason McGee talking about how the developer mindsets evolving in response to some new innovations with cloud we've heard from a number of other individuals I won't list them all but if I were trying to summarize the three points that I think kept coming through it's number one the cloud does force changes to the way you think about business problems and methods tooling and approaches for doing that are starting to mature very rapidly Micro services for an example for example is not just a technology it's also an approach to thinking about a problem and that informs everything I did the second thing that we've heard is that can't just talk about greenfield applications we've had this enormous investment in applications have been running businesses for a long time of those applications tend to be very stateful they tend to be very database driven they tend to be very operational in nature those applications have to move forward if nothing more from at least from a management standpoint how can we bring a management mindset an operating model of the cloud to start to channel or structure change and evolve how we manage those applications but ultimately bring new classes of services to those applications I think the last one that we've heard over and over and over it that this really is gonna require a strong community we have to take a community approach to invention you have to take a community approach to innovation and the social change is required to take advantage of technology and achieve the business outcomes that we want and if one thing has come through loud and clear through all the conversations is that that this year IBM think or I didn't think 2019 San Francisco is gonna be a great place to be able to get together with peers and have those conversations and think about the outcomes that enterprises want to achieve and then talk to people that can actually help you get there and one of the things that I find interesting about think this year is that the industry's changing we're seeing new rules or evolution of roles and an evolution of how those roles work together and think is actually starting to reflect that it's manifesting itself itself there's a couple of campuses one that's focused more on data and AI a very very natural binding or combining and one that's focused more on infrastructure and cloud again very natural so I hope to see be able to carry on and continue these conversations we've had today at IBM think and hope to see you there as well so once again this is Peter Burris Ricky bond the cube from the IBM from the from the illustrious from the vaunted Thomas J Watson Research Center in Yorktown Heights thanks very much for watching the cube today [Music]

>> From Yorktown Heights New York It's theCUBE, covering IBM Cloud Innovation Day. Brought to you by IBM. >> Hi, I'm Peter Burris and welcome theCUBE. We're broadcasting today from IBM innovation day at the Thomas J. Watson research labs in Yorktown New York. Having a number of great conversations about what's going on with the industry, what's going on with the cloud, and to bring that further, Judith Hurwitz, president of Hurwitz Associates, longtime analyst. Judith, welcome to theCUBE. >> Thank you, Peter. Great to be here. So, Judith, I'll just open it up. What do you think are the two or three most important things that people should be thinking about right now? >> Well, I think as we look at the maturation of cloud and computing and the changes that we see, I think one of the most important things is the movement towards open and standards, because what customers really want is computing. They don't really care if you tell them "Well, that service runs over there and this one runs over here." They don't care about that. What they care about all of the workloads, all of the applications they need to get their jobs done just work. So if a workload needs to move, it should be able to move because it's less expensive or more efficient or it handles a workload better in terms of performance or security. Customers want the freedom to be able to do what they want when they want it, and not to be locked in. So openness is really becoming the battlecry for the cloud. >> You're talking about two things there. Let me parse them out. You're talking about the breaking of the natural relationship between where the resources are and where the value of the work is provided. >> Yes. >> And there is a degree of openness to that, but then there's also this notion of openness which is how fast innovation, what model are we going to use? Let's break those apart. Let's start with the idea of the cloud breaking the traditional mold of this workload here, that workload there. How is cloud doing that and what's the future for that flexibility look like? >> I think if we were having this conversation ten years from now we wouldn't be talking about cloud. We would be talking about the elasticity and the way we do computing so it really meets the needs of whatever business change you're experiencing. What's held companies back and what's held IT back is the idea that you're stuck with the platform or the application or the technology that you've always been using, and it makes change really hard. So, the more flexibility you can have, and the cloud in terms of elasticity, the way you can create new workloads using cloud native and microservices and leveraging containers, all of these techniques will lead us into a world where you can create a bunch of services and choose and pick the ones you want to get your job done and it really adds a level of innovation and speed that we've never seen before with IT. >> So let's build on that. One of the things we tell our clients is to focus on what we call plasticity. It's a physics term. Elasticity is a single workload, scale it up and down. Plasticity is new workload changes, transforms, leads, perturbs the infrastructure, the infrastructure reforms around it. One of the reasons why that concept becomes so important is precisely because of the rate of increase in innovation, as you said. So now tie open back to that. What is it about open, that's not just about making sure we have system software standards, but is actually doing a better job of turning business into software at a higher level. >> In a sense, it's what I would call service as software. If you can take the business process or how you want to interact with your customers, and you can turn those into software services that are malleable, that you can change and innovate on without having to go from top to bottom and recode everything, which is what's held companies back for probably 40 or 50 years. As you modularize things, you can, for example, simple idea like the way you would calculate a 30 year mortgage. In most companies over the years there were 30 different ways you could do that and each application had its own way. What if you could have a single service that did that that you could apply it no matter what the use case and what the business case was, apply that same concept to any business logic or any business strategy, that's where you get what you're calling- something that's very plastic, very malleable, and allows you to change, because in the past we've always written applications or written systems as though they were based on how we do business right now. And when you do that, you can't change. >> So one of the ways, again, if I were to describe some of the big changes and let me test this on you, is that I say for the first 50 years of computing it was known process unknown technology. We knew we were going to do accounting, we knew we were going to exchange titles, became supply chain, et cetera, we knew we were going to do HR. But we didn't know if it was going to run on a mainframe or how to run on a mainframe, or client server or the internet or whatever else it was. We're entering into a world now where it's unknown process, relatively known computing, or technology. We know it's going to be a cloud or cloudlike thing. When we think about that unknown process, more data first applications, data driven applications, where do you foresee some of these magnificent changes that are on the horizon? >> So, I think one of the most important changes is that we start leading with data rather than process, because if you lead with process, that's the past. If you lead with data, data will lead you to process. So if we have data driven organizations where the data, using it in a predictive analytics way, really using machine learning, algorithms, and some of the emerging AI techniques, we can begin to have data drive us to process. >> So, Judith, I know you've gone to IBM Think every year for a number of years now. Probably almost as long as I have. If you step back and say San Francisco, 2019, February, 30,000 plus people, what are you looking to get out of Think this year that builds upon what you've gotten out of it in the past? >> Well, what I really like about Think and about IBM events is that it brings together so many people, both IBMs fantastic technical leadership with business leadership, and it brings together the programmers. It brings together the IT leaders with business leaders, so it's a really coming together of the minds across business organizations, really collaborating together to really get to the heart of key business problems. >> Excellent. Judith Hurwitz, president of Hurwitz and Associates, thanks for being on theCUBE. >> Thank you. >> And this is Peter Burris, we'll be back with more of theCUBE from IBM Innovation Day in a few minutes. (upbeat techno music)

>> From Yorktown Heights, New York, it's theCUBE, covering IBM Cloud Innovation Day. Brought to you by IBM. >> Hi, I'm Peter Burris from Wikibon, and you're watching theCUBE being broadcast from IBM Innovation Day at the Thomas J. Watson Research Lab in beautiful Yorktown, New York. And we've had a number of great conversations thus far, we've got some more on the horizon, stay with us. Now, we've got Jim Comfort. Jim Comfort is the General Manager of Hybrid Cloud Services at IBM. Jim, welcome to theCUBE. >> Thank you, Peter, glad to be here. >> So, Jim, what does Hybrid Cloud Services as a group do? >> Actually, we run infrastructure for clients. That's our business, but we help you advise, build and manage private cloud. Advise, build and manage consumption of public cloud, Azure, Google, IBM, and we help you manage and stitch all of that together. >> So a lot of people think of cloud and they think of this monolithic thing. "If I go to the cloud, suddenly my business has changed." But there's more to it than that. There's a number of different things that a business has to be successful at to succeed at getting to the cloud. What is your perspective on that? >> Well, I completely agree. And this is kind of my first conversation with clients is, you need a business strategy, but to execute that strategy you have to realize it will touch most everything in your business. It'll touch infrastructure, it'll touch applications, it'll touch your dev ops, or your development process morph to dev ops. It'll touch your operations very profoundly, this whole SRE thought, and it will test your data governance and management as well as your security and compliance. So that's the scope that you have to comprehend. >> But most people, they start with perhaps the infrastructure first and end up with the data last. Is that the right way to think about this? >> I agree, many do, and actually I have not seen many build-it-they-will-come strategies succeed. And so what I really look for is, do you understand the business drivers? Top-line revenue growth, new markets, new insights, new data, and from that can you derive a technology strategy? What I've seen happen in many cases is, if you start from the bottom up you'll be trapped in what I call the religious wars of technology that never end. >> And most people, a lot of folks start from the bottom up, because they start from the technology side of the business. >> Correct. >> Are you seeing more business people getting engaged, and conceptualizing what the strategy needs to be? >> I am, and it starts on both sides. The business people will say, "I need to move faster than you can move, so I'm going to do something different," and the IT people will say, "I can do that for you, here's what you need." The two signatures of the most successful transformations are does the line of business and the IT have the relationship to collaborate so they actually learn together? And then if they have that, have they actually created a team that understands the new as well as they understand the existing or the old, so they can actually understand what's real, what's not, where's the hype, what really happens. And then they get into the rational, real planning decision. >> So as you think about some of the assessment challenges, because you said you go through the assessment process, what are some of the key questions that a client should start with as they think about undertaking this journey? >> Well, number one is start with the business driver. I said that already, but you have to start with understanding what you're trying to accomplish so you can make choices. And the other is, start small enough and get to the end of something so that you know what the reality is, and that's where our, this is where we bring in our methods. When you hear us talk about the garage method, you hear us talk about MVPs and all the language everyone wants to use. We like to start with something, and start that iterative cycle of learning. That's the key. >> So with an iterative cycle of learning, in many respects this whole notion of agility is predicated on this idea of being agile or iterative. But it's also empirical, knowing what the data is, knowing what the data says, and being opportunistic. How does a customer balance that as they get going, say early on in the cloud journey? >> I think, again, most of what we're talking about in digital transformations is new insights that will help your business. That could be from data that you had, it could be new data. And if they think about it, what insights am I looking for? What new experience am I trying to create, and what do I need to do that? Then you start to get people to step back and think, well, what are all the possibilities? And now, how do we tackle that? So it starts from realizing, what insight am I looking for? >> So there's a lot of invention happening in the industry. >> Oh, yeah. >> And enormous new things being created. Customers are being overwhelmed at trying to adopt them. The innovation side, the social side of effecting a change in the business. You mentioned some of the markers for success and putting together the strategy. Go forward a little bit. What are some of the companies that have successfully gotten to that end stage maturity doing differently? >> We have a number of very good ones. I mean, a very clear one in my mind is American Airlines, where they were really trying to change the experience. They had three distinct things that had grown up over time, the mobile experience, the kiosk experience and the Web experience. Three completely different things. They brought it together, converged it, modernized it, and now completely changed the experience and the speed with which they can now act on what they see for their clients or for their customers, all of us. But also as they get new ideas, the speed and the velocity that they can bring those in is phenomenal. >> And that improves their ecosystem, their ability to work with a lot of others as well. >> Their ecosystem, how to work with others, how to bring in new ideas. And this is all, for them it's all about client satisfaction and service to their end client, to the end user. That's what it was. It had a lot of technology dimensions, but they were very clear the experience they were trying to attack. >> So next February, IBM Think, 30-plus thousand people descending upon San Francisco. You guys are taking it over. What kind of conversations are going to be on your agenda as you work with customers and partners to get this message out? >> Well, it's really two things. I often joke the blessing and curse of IBM is the breadth of our portfolio. It's a very large place, but we actually have a very simple, clear way to talk to, advise, move, build and manage. Those are the steps you need in your journey. Now, which journey for you, which type of thing. But that, we have clarity on that, and I think you'll see that displayed at Think and get to understand it. The other thing is that we have a lot of experiential and real practical, we've made this happen for many large clients at scale, and I think that what we want people to understand is we can help you that same way. It's really pretty simple. >> Jim Comfort, General Manager Hybrid Cloud Services at IBM. Thanks for being on theCUBE. >> Thank you, Peter. >> And we'll be back momentarily with more from theCUBE at IBM Innovation Day here at the Thomas J. Watson Research Center in New York.

>> From York Town Heights, New York, it's theCUBE covering IBM Cloud Analyst Summit, brought to you by IBM. (techy music) >> Hi, welcome back, I'm Peter Burris of theCUBE, and we're having conversations here at the IBM Innovation Day at the Thomas J. Watson Research Lab in York Town Heights, New York. We've got a great conversation. Don Bolia is the general manager of cloud developer services at IBM, welcome to theCUBE, Don. >> Thank you very much. >> Or should I say welcome back to theCUBE? >> (chuckling) Yes, thank you. >> So, Don, we were talking with one of your colleagues, Hillery Hunter, who's the CTO-- >> Mm-hm. >> Of here at the cloud infrastructure team, and about the fact that everybody's talking about the rate of growth of data, and nobody's really discussing the rate of growth of software, which is perhaps even more important, ultimately, to business. What is that rate of growth look like, and how is it related to the role of cloud? >> Yeah, so it's a great question. I mean, with my role as kind of owner of our platform services from the cloud perspective, one of the things we've noticed over the last probably five or 10 years is just a massive rate and pace change with respect to iteration on the software development cycle. So, they started with mobile, I would say, and then has moved to cloud since then, where you know, the expectation is everything is updating all the time, you know, everyday, all times of the day. Within our own Kubernetes and container service, as an example, we push over 500 updates a week to that software stack on behalf of our customers, and so I think there's a rate and pace of how things are changing from that perspective, but then there's also the fact that everybody's leveraging those services to then build the next generation of software. So, in our case we have a set of base services that I provide for things like containers that then the Watson team, for example, uses to build their microservices, which are then, you know, realized as machine learning and other types of services that they provide. So, you see the stacking of software, if you will, from you know, the high iteration rate at the bottom all the way to the next level and the next level, and the ability to unlock value now is something that happens in, you know, hours in some cases, or a couple of days, whereas before just provisioning the software would've taken months, and so we're really seeing just a whole change in the way people can develop things and how quickly they can get to the end result. >> Now, we're here at the Thomas J. Watson Research Lab, and downstairs is this wall of all IBM fellows, and one of them E.F. Codd, the famous originator of database and the role that SQL played, et cetera-- >> Mm-hm. >> In relational database technology. He wrote a seminal paper back in the early 1970s about how the notion of developer was going to evolve over time, and he might've been a little aggressive in thinking that we were going to end up with these citizens developers than we actually happened, but we are seeing the role of developer changing, and we are seeing new classes of professionals become more developer-like. >> Mm-hm. >> How is that relationship changing the way that we think of developer services that you serve? >> Yeah, it's a great question. I think, first of all, software is sort of invading almost every single industry, and so, you know, people have got to have some amount of those skills to be able to function in kind of the optimal way for whatever industry they're in. So, what we're seeing is that as we've built more and more foundational services, the act of actually creating something new is more about stitching together, composing, orchestrating a set of things, as opposed to really building from scratch everything from the ground up, and you know, things like our Watson services are a great example, right? The ability to tap into something like that with a couple lines of code in an hour, as opposed to what would've taken, you know, months, years, whatever, and even really, frankly, been out of the reach of most developers to begin with is now something you can have somebody come in and do, you know, with a fairly low level of skill and get a good result on the outside. >> So, we've got more demand for code as we move to digital business, more people participating in that process, cloud also enables paths, a lot of new classes of tools that are going to increase the productivity-- >> Yep. >> Including automated code generation. How is the process, how is that tool set evolving, especially as it pertains to the cloud? >> Yeah, so I think one of the mantras of cloud is automation, and in order to standardize and automate, that's really how you get to the kind of scale that we would see in, say, a public cloud like the IBM cloud. So, it really is kind of a fundamental premise of anything you do has to be something that you automate, and so we've seen a whole class of tools, to your points, really start to emerge, which allow people to get that kind of, you know, automated capability. So, nobody thinks of, for example, creating a, you know, a build pipeline these days without using a set of tools. You know, often they're opensource tools, and there's a lot of choice within that whole spectrum of tools, and we support a bunch of different varieties, but you would never think today of having a build process that isn't totally automated, right, that can't be instantly recreated. Even the whole process of how you deploy code in a cloud these days is sort of an assumption that you can destroy that and restart at any point, and in order to do that, you really need the automation behind that, so I think it's a base premise now. I don't think you can really be at the velocity that people are expecting out of software without having a totally automated process to go through that. >> So, any digital business strategy presumes that data's an asset, and things that are related to data are assets, including software in many... Well, software is data when you come right down to it. >> Mm-hm. >> And we want to exploit that data and generate new sources of value out of that data, and that's one of the predicates of digital business, but at the same time we also want to protect those attributes of data-- >> Mm-hm. >> That are our IP, our enterprise's distinction. As we move forward with software, how do we reconcile that tension between more openness and generating a community that's capable of improving things, while at the same time ensuring that we've got good control over our IP where it actually does create a business differentiation? >> Now, that's right, and you're right, data's king. So, you know, the software can do, you know, a set of things, but most of the time it's operating on a set of that data, and that data's where the true value that you can unlock comes from. Our policy, from an IBM perspective, has always been that, you know, your data is yours, and to your point, this IP that you may want to protect, and we try to give you the tools to do that, and so a lot of our philosophy, within the cloud in particular, is around things like Bring Your Own Key, where you have control of the keys that encrypt that data that's in the cloud. In fact, we would like to be totally out of that loop, quite frankly, and have it be something that is controlled by our clients, and that they can, you know, get the value they're looking for, and so we'll never have a situation where one of our services is, you know, using or acting on data that is really, you know, not ours to use, and so that's been a fundamental premise of the cloud as we go forward, and again, we continue to provide a set of tools that really let you manage that, and to your point, you know, not everything gets managed at the same level. Some things are highly protected, and therefore have, you know, layers and layers of security policy around them, and there's other examples where, you know, you're relatively able to make that open through a set of APIs, for example, and let everybody have access it. From our perspective, though, that's really a client choice, and so for us it's about giving the right tools so that they can do the job they need to do. >> February 2019, San Francisco, IBM's taking over San Francisco with the IBM THINK show. What types of conversations are you looking forward to having with customers? What excites you about the 2019 version? >> Yeah, so I mean it's a great venue. It is absolutely, you know, something that I look forward to every year. I know my team looks forward to it, as well. I mean, the amount of interaction we get with clients... I mean, it's really all about the client stories, so you know, what are they able to do, in my case, with our cloud services. What can I learn about what they've done, and how, you know, can we then leverage that to make our services better, and so, you know, to me it's all about, you know, what you can learn from others, and it's a great form to be able to do that and there's a lot of great things that, you know, you can dive deep on. You get access to a lot of the IBM technical experts, so I have all of my, you know, fellows and distinguished engineers there, you know, on hand, and just great conversations. There's always great insights that you get from it, highly recommend it. >> Don Bolia, IBM general manager of cloud developer services, thanks very much for being on theCUBE. >> Thank you. >> Once again, we'll be back from IBM Innovation Day here at Thomas J. Watson Research Center in York Town Heights, talk to you soon. (techy music)

>> From Yorktown Heights, New York, it's theCUBE, covering IBM Cloud Innovation Day, brought to you by IBM. >> I'm Peter Burris of Wikibon. Welcome back to IBM Innovation Day, covered by theCUBE, from beautiful Yorktown Heights, New York, Thomas J. Watson Research Center. A lot of great conversations about the journey to the cloud and what it means, and we're going to have another one here with Denis Kennelly, who is the General Manager of Cloud Integration in IBM. Denis, welcome to theCUBE. >> Thank you, Peter, and welcome to Yorktown also. >> I love it here. So, very quickly, what does the GM of Cloud Integration do? >> Yeah, so, I suppose we start from the beginning, right? So I am responsible for a lot of what we call the traditional IBM middleware. So these are brands that are known to the industry and to our customers, things like WebSphere, Message Queue, or MQ, as we know it, which is kind of the core foundation stones for a lot of IT today that's out there in the industry. And it's not just about, you know, sometimes people talk about this legacy, but this is what all the systems run on today. And also, I'm involved in the whole journey of moving that middleware to the cloud and enabling customers to get on that journey to cloud. And it's not just to a cloud, because your typical enterprise today has probably on average about five different clouds, and clouds, as we know them as the IS players of the past, but also when we talk about cloud, we also think about things like SaaS properties and applications of that regard. So it's helping customers go from that traditional IT infrastructure and on their journey to the cloud. That's what I do. >> So utilizing these enterprise-ready technologies that have driven the enterprise, bringing them to the cloud as services, but also making sure that the stuff that's currently installed can engage and integrate the cloud from a management service standpoint as well. >> Absolutely, because customers have made a huge investment in this middleware, and a lot of the transactions, and a lot of the security, and a lot of the risks set in these systems, and they have served us very well for many decades. Now, as we start to move to the cloud, it isn't a binary switch. It's going to be a transition over time, and today, I think we're about 20% into that journey. I would say we've done some of the easier parts. Now we're getting into some of the more complex and some of the more difficult problems. And kind of one of the underlying pieces of technology we're using to enable customers to do that is container technology. So we've made the decision to use containers right across our middleware, our software. So what I mean by that is we've taken all our software and it's running on containers today, and that's a key enabler to make this happen, because containers give you that flexibility and that openness to run on different targeted environments and be able to run on different clouds at the end of the day. >> The model by which developers thought about integration would be through a transaction. Generally pretty stateful. So, I'll put something in a queue, I'll wait for a response, guaranteed delivery. Now we're moving to a world, containers, a lot more reliance on stateless interactions. It means we're being driven mainly by events. I'm thinking in terms of events. Talk about how that is changing the way we think about the role of middleware or the role of integration amongst all these different possible services. >> Yeah, it's a great point. I mean, so if you think about containers, we think about stateless, and we think about microservices, and we talk about event-based applications, so a lot of those front ends are on that today and building on those technologies. So you've got to enable the new developers to build in that way. Now, how do you integrate that with that backend, right? Because at the end of the day, these transactions are running in the backend, and you really want to enable, as part of the transformation, you want to open up those backends to those new developers and to those new customer insights, because what is digital transformation? It's about putting the customer at the middle and enable insights on those customers, and enable rapid development of those applications. So at the core of that is integration, and integration is not just message-based integration. It's being able to take those backend transactions and surface them up through APIs, not just the standard APIs as we think of maybe as web services, but event-based probability models, and event-based APIs also, and doing that in a consistent and a secure manner, because if you have all these complex transactional systems, who has access to that data? Who has access to make those transactions? Who can, at certain levels, et cetera, and we really have to do that in a secure and a consistent manner across these environments is critical to what we do. >> So, can you give us some examples of some customers that are successfully transitioning their backend systems to these new technologies in a way that protects the backend system, makes it economical to do so, in other words, doesn't force change, but can utilize some of these new integration technologies to make both the new investments more valuable but also the backends more valuable too. >> Yeah, I mean, if you think of, I'll give you an example of a customer, American Airlines, in the airline industry, right? So, if you think about travel and airline travel in times past, you know, you made a reservation maybe through an agent and you booked the flight from A to B. Today, you have your cellphone, you get regular updates on your flights. If you're delayed, you're possibly offered re-routing options, et cetera, right, so there's a classic example of how digital has transformed the airline industry and the airline booking industry. If your flight, you know, if there's weather patterns, et cetera, how you can get real time updates on your flights. So, okay, that's all happening on the front end, on your cellphone, or your tablet, or whatever, but the backend booking system is still a transactional-based system that says, Peter is on this flight going from A to B at this time, et cetera. So, that's an example of how we have modernized an application and we have worked with American Airlines to make that happen, to give you that kind of 360 view as a customer, where you bring in together flight information, weather information, rating information, because we'll offer you different alternatives in terms of if you need to rebook in the event of something going on, and at the backend, there's still a transaction that says, book Peter on this flight from A to B, and that's a real life example of a transformation, how we've integrated those two worlds there. >> So if we go back five or six, or more than that, say 10, 15 years, in the days of MQ, for example, the people who were developing, and setting up those systems, and administering and managing those systems were a relatively specialized group. Today, the whole concept of DevOps in many respects is borrowing from much of the stuff that those folks did many, many years ago as infrastructure builders, or developers, as I call them. How does that group move into this new world of integration in the cloud? >> Yeah, so, I think first of all, the rate and pace has multiplied, right, so the rate and pace of which we make changes to the system has multiplied. I mean, maybe traditionally, we run in changes maybe once a month. We have things like change control windows. Things were very well controlled, et cetera, right? But at the end of the day, it doesn't meet the needs of today and what we need to do in a digital world. So today, we're running in changes on the hour. So now, you're faced with a challenge, right? So when you make changes, how do you know that the system is still performing, is still operating at the level you need it to operate on? You start to think about security and you start to think about, okay, I've made a change, have I introduced vulnerabilities into the system? You've got to, you know, in the past, these were all separate groups and almost islands within the operation center, where you have the developer, who kind of over to all the code, and then operations looked at it and see how it's performed, and security checked for compliance, et cetera, and they were kind of three different islands of personas or groups within the organization. Today, that's really collapsing into one organization. The developer is responsible for making sure the change gets in, for making sure the change performs, and is also security compliant. And we call this the role of the SRE, or the systems reliability engineer, and really bringing those two worlds together into one persona, and it's not just one persona but having the systems on the inside to make that happen. And that's critical in how management is changing and the management of these systems is changing, and how the skill level is needed in this new world. >> So Denis, one more question. In a few months, IBM Think is going to take over San Francisco, February 2019, >> Looking forward to it. >> 3,000 people. Talk to us a little bit about what gets you excited about Think, and what kind of conversations you hope to be having while you're there. >> Yeah, well, you know, this is the one time of the year where all of IBM comes together, and it's new this year that we're going to San Francisco, and in particular, in our cloud business, which I'll talk about, which really encompasses everything we're talking about here, which is our middleware business and also how we move customers to the cloud, and really engaging with customers in those conversations. And this is the one time of the year where all of IBM comes together, and where you can see the full breadth of our capabilities all the ways from our systems, and the hardware, down at that level, at the chip level, right through to the middleware and the software to our cloud, and actually engaging with customers, and really understanding what the customer needs are, and making sure that what we are working on is meeting those customer needs, and of course, if we need to adapt or change, and take that feedback back into the organization, so we do that in real time. It's a very exciting time for us. It's a week in the year that I really look forward to, because that's where all of IBM comes together, including our services, et cetera, and where we actually have conversations with key customers and partners and really understanding what's going on in the industry and how we can help people on this journey to the cloud that I talked about. >> Denis Kennelly, IBM General Manager of Cloud Integration, thanks very much for being on theCUBE. >> Thank you, Peter. And once again, this is Peter Burris. We're signing off from the IBM Innovation Day, here at the Thomas J. Watson Research Center in Yorktown Heights. Thank you very much for watching. Let's carry on these conversations about cloud and the future of computing.

>> From Yorktown Heights, New York, it's theCUBE, covering IBM Cloud Analyst Summit, brought to you by IBM. >> Hi, I'm Wikibon's Peter Burris. Welcome back to theCUBE coverage of IBM Innovation Day, here at the Thomas J. Watson Research Center in Yorktown Heights, New York. Great series of conversations, and this next one also is going to be a great conversation, with Jason McGee, who's an IBM Fellow, VP and CTO of Cloud Platform here at IBM. Jason, welcome to theCUBE. >> Thanks for having me. >> So, we've had a lot of great conversations about what does open mean, where is the cloud going, what is the role of developers in this whole thing, but I want to dig a little bit deeper into this kind of core question. The cloud suggests a new model for computing. I would also think that would mean that there's a new model for development on the horizon. >> Yeah, absolutely. >> Can you talk to us a little bit about that? >> Yeah, sure, I mean, I think that's absolutely true. I think one of the core things that people are trying to get out of cloud these days is development velocity, you know. For many years, of course, one of the key pressures in IT has been how do I do stuff more quickly, and that's gone through many iterations over time, but I think cloud today, people are really trying to figure out how to leverage cloud as a platform for speed of development, and the combination of services on cloud, and new development models like microservices, and new technologies like containers are all kind of contributing elements in helping people solve this problem, how do I build stuff more quickly. >> So, with all that new technology, is a new mindset required? Does somebody think about the problem differently, does somebody break the problem down differently? How do you start with that notion of looking at a business requirement or business outcome, and translate it into the technology? We used to just create code. Now we're doing something different. >> Yeah, I think the first thing you have to do is think about how to organize people. You know, software development at the end of the day is a sport amongst people and you have to think about how to break up the problem, and so, like microservices, a lot of us think of microservices as a technology. It's not really a technology, it's really a philosophy about how to attack a problem with a group of people, it's about how to organize, and its fundamental idea is break it into independent parts, and allow a small team of people to not only develop that part but to own it end-to-end, you know, like the old development model was development, test, production, hand it over the wall to operations. The new model is break it into small problems and then have a team own the whole thing end-to-end, and with that new organizational philosophy comes new architectures for apps, new technologies to help you do that, and new platforms to run things on. >> So, as we think about that, that suggests that the approach to software from a licensing standpoint, from what are you buying, what are you installing is also going to change. How do you foresee, and what is IBM preparing customers for in this kind of new world where software is a service coming from a lot of different places as opposed to a license with, you know, 800 million lines of code or eight billion lines of code behind it? >> Yeah, it's interesting. I think these new ideas are enabled by things like cloud. Part of the reason that cloud has enabled this new model to be feasible is because you get, for example, consumption-based pricing. You can use a wide variety of technologies, you can pick the right tool for the job, you can pay for just what you use, and therefore, the old models of static software licensing and big platforms can start to fade away as these small teams are able to kind of pick the right tool for the job, and that wouldn't be possible in a world without like, as a service delivery, and meter pricing, and things like that, because you would have to consolidate to fewer choices and buy bigger chunks of things. >> As you said, microservices is more of a philosophical approach to how you think about software, and it's also predicated on that wonderful notion of REST. A great paper was written a number of years ago on APIs. IBM has kind of an interesting role in the industry, though. IBM has got to bring a whole bunch of customers with highly stateful applications forward into the cloud. Kubernetes, great for stateful. How are we going to address that tension between the stateless world of greenfield applications and the stateful legacy that has to move into this new world? >> Yeah, I'm glad you brought that up. I mean, I think a lot of times new trends emerge and it's easy to ignore the past, but the lesson I've learnt in over 20 years in IT is like, nothing ever goes away, right, and so you have to not only define the future, but you have to figure out how to help people get there. I actually think part of the reason technologies like Kubernetes are so dominant right now is because they actually do a reasonable job at both. You know, Kubernetes and containers are a great platform for the kind of new architectures and for adopting these new methodologies we're talking about, but they can also accommodate the existing apps, and you can move existing apps into these new platforms, and so, that helps give people a path. They can move something they have and then slowly re-factor it, or they can move something they have and build new things around it, and they could do all that with platforms like Kubernetes as an enabler, right? And it's been interesting to watch. Like, at IBM, we obviously make Kubernetes available, both in our public and private clouds, but we're also big users, and we run all of our cloud services on that platform. Stateful databases, AI and machine learning workloads, analytics platforms, stateless web apps, like, the whole lot, we've been able to run on a platform like that. >> Talk to me a little bit about this notion of cloud operating model and how we manage that, because it seems to me as though the user adoption of a lot of these new technologies are going to be facilitated if we can put forward a management platform that uses those technologies to manage those technologies. What's the relationship there between the evolution of management? Is that a leading edge of how we are going to see people adopt some of these technologies? >> It's certainly a very kind of critical component of the story. I mean, if you really believe in the idea that where we want to move to is this kind of microservice model of small teams that run things themselves, then you get into the question of, all right, well, if you have eight people whose job is to run something in production, they need to be able to do that efficiently, right? You can't have complex operational processes, you need a lot of really good tools, it needs to be really easy for them, 'cause you're asking people to have a really vast set of knowledge, and so, it's driving the evolution of management philosophies. You're seeing new technologies, like SDO, for example, emerge, which are allowing like an application person to define policy about security, and access, and networking that normally would've required like a network expert to go to. >> And more, which makes it a very powerful platform. >> Powerful platform, right, but I think it's coming out of this realization that like, if that small team of people ever want to sleep, and when they have to run things, they're going to need tools to help them do that. So it's been interesting to watch that kind of circular evolution of these different domains. >> So, 20 years of experience from web-sphere forward. Let's think about the next five years. Where is the biggest innovation going to happen in software? >> Well, I mean, there's the obvious stuff around the application of AI, but the part that I'm most excited about is I think we've been on an arc over the last 20 years, to make the application the center of IT. You know, historically, infrastructure has been the center of IT. You start a project, you buy a server, you install an operating system, you set up management tools. >> That's been a big asset. >> The center has been the infrastructure and you build your way up. And I think as velocity has become dominant, we've been trying to flip it and say, I'm building an app. Let me focus on the app, and focus on what the app needs, and drive the requirements down, and I don't think we're done yet. I think there's a lot more to do there, but that's the path we're on. I think over the next five years, we'll really get there, where as an app team, I don't really have to think about infrastructure, and I can have the system adapt to the needs of the application. >> Do you foresee a point where the data and the application are increasingly and further broken apart? >> The data and the application? I don't know that they're going to be further broken apart, but I think we'll see more kind of intelligent scheduling and combinations of those things, like there are cases where the data needs to be king, and the application needs to come to the data, and vice versa, and historically, the data world and the app world have been pretty separate, right, and you know, again, if we think teams are going to run their things, then just like they're doing ops and dev, they're going to have to do apps and data, right, and so, there's an opportunity there to bring those worlds closer. I see some of it, but, you know, Kubernetes as an example, as a common operational platform for both kinds of systems, but there's more, for sure. >> So bring it together when it makes the most amount of sense, keep it separate when other people need to use the data. >> Stop assuming you have specialists in every technology, and assume you have a multi-disciplinary team that has to run it all. >> All right, Jason, one more question. February, San Francisco, IBM takes it over with IBM Think. A lot of users, a lot of new questions being raised, a lot of opportunity for learning, a lot of opportunity for networking. What are you hoping to accomplish? What conversations do you want to have at Think? >> Yeah, I'm really excited, I think, to have conversations with clients about how they're actually adapting to this new world. I think sometimes the biggest challenge is not technology, but how organizations assimilate these ideas, and so, I'm excited for the conversations with customers about what problems they're solving, sharing those experiences with each other, and also practitioners. I think we've moved into a world where IT is dominated by the people who actually do the work, by the practitioners, and I really hope to see a lot of them show up at Think in February and share with us what they're doing. >> Jason McGee, IBM Fellow, VP, CTO, Cloud Platform here at IBM. Thanks very much for being on theCUBE. >> Thank you. >> And once again, this is Peter Burris from the Thomas J. Watson Research Center in Yorktown Heights. You've been watching theCUBE. Stay tuned. (techno music)

(technological music) >> From Yorktown Heights, New York, it's theCUBE covering IBM Cloud Innovation Day, brought to you by IBM. >> Hi, I'm Peter Burris and we are broadcasting theCUBE from IBM Innovation Day at the Thomas J. Watson Research Lab in Yorktown, New York. We've got a great number of guests to talk about. We're going to start with Hillery Hunter, who's the CTO and vice-president of cloud infrastructure at IBM. Hillery, welcome to theCUBE. >> Thank you very much. Pleasure to be here. >> So, you're relatively new in your role. Tell us about some of the things that you're focusing on as the CTO of cloud infrastructure here at IBM. >> As CTO for cloud infrastructure, I'm focused on making our cloud the best possible place that it can be for people to bring their data, bring their applications, and overall, come into that modernization journey with us, the process of transforming to become a digital enterprise. >> So, one of the things that people talk about all the time is how fast data's being generated. Nobody seems to be talking about how fast software is being generated, and yet, that seems to be one of the advantages and potentially the liabilities of doing cloud wrong. Talk to us a little bit about how IBM sees the world of software changing as we move forward with the cloud. >> [Hillery] There are parts that are consistent with what we've seen for about the past 20 years in open source, and there are parts that certainly, we feel like are accelerating and changing. With regard to the pace of software and its change today, open source is clearly this innovation space. It's this playground where lots of people can go and can contribute. We can take... We're here at the IBM research facility. We can take the latest in innovations and math that helps us accomplish great AI and AI insights. We can take that into open source. We can take microservice integration capabilities and take it into open source and work there collaboratively with people across the industry. What we see, therefore, is a tremendous rate and pace in change of software and the capability of software and its ability to analyze data and bring insights to data and realize the promises of big data, of getting insight out of that data, is just really on a tremendous growth rate. When you move to cloud, you're not just doing what they used to say of converting capital expense on premises into opex and renting a server in the cloud. You're bringing your overall workload and modernizing it and bringing it into this era where you're able to apply through microservices and cloud-based programming methodology, you're able to bring the latest of software capability to your data and get more insights out of it. >> You're really able to alter the operating model of how not only your technology group works, but also how your business works. >> Absolutely. >> How does Red Hat play a role in this? >> We have shared principles with Red Hat. We both have been active in the open source communities. IBM famously had billion dollars of investment in LINUX going back 20 years ago, and Red Hat is a prominent name in open source. We have a shared understanding of the value of open source and the value of rate and pace of innovation that's commensurate with what open source provides. We have a shared value around what enterprises need and a shared client-centric view that you need support on your software, that you need certifications, that you expect security, those kind of things. There's tremendous amount of shared value proposition in what we see as the rate and pace of innovation as well as then moving that into an enterprise context. Enterprises make these choices very carefully. As consumers of enterprise capabilities, we expect them to guard our data, we expect them to do things on our data in a secure way, and there are many foundational elements in philosophy that are similar between the two of us. >> You mentioned that cloud started out as this notion of capex to opex, move all your data to a single place, let somebody else deal with it. Increasingly, enterprise is starting to recognize that their data may sometimes have to remain in place. We start talking about innovation, open source, these new classes of services. What is it going to mean to bring the cloud experience to the data from IBM's perspective? >> We really see that the data today exists in multiple places, that largely because of that, people are partway through their journey to overall modernization. They're partway through their journey to the cloud. We really think that the world is going to be hybrid, meaning that... Or, the world is hybrid, I guess I would say, meaning that there is data and there is cloud function needed on premises and in public clouds. There's a need for private, dedicated environments in the public cloud as well. There's a significant amount of IT that is currently traditional in that people are in the process of modernizing, and that may initially be through a private cloud context on the journey to overall workload modernization. We also see that the world is multi-cloud. People are using upwards of 9 clouds or more in many cases, and that, in a lot of cases, has to do with this intersection of function and data residency and being able to bring together all of those pieces of where the data needs to be or where the data currently is, and then bring software function to the data is something that we see as critically important. >> Without being too specific in the use of the word binding, today, the idea is you bring your data to a cloud supplier and then, you can run the services of that cloud supplier supplies on that data. Do you and IBM foresee a world in which the customer's going to be able to control their own data and then acquire the services from the cloud and bring it to their data? Is that the direction you think it's going to go? >> Not only do we see that it will be possible, we think that it is possible and we're putting things in market already today that enable people to bring cloud function to their data. The IBM Cloud private offerings and IBM Cloud private for data enable people to, in their environment, where their data resides, bring sophisticated data, warehousing data analytics and AI capabilities. Fundamentally, that process of workload modernization is a set of steps and it starts with data and it starts with modernization of that environment and it matures then into being able to get deep insights through the power of AI on that data. >> Let me ask you one more question. In February, IBM's going to host 30,000+ people in San Francisco. Unbelievable opportunity for networking, learning, and IBM Think. What kind of conversations do you expect that you're going to be having in Think in 2019? >> I think you hit at the heart of the conversations that we're going to be having at Think and our positioning of the hybrid multi-cloud environment. Our other core tenets there are open and open source and keeping up with the rate and pace of open source as an innovation stream, providing choice in how folks are deploying cloud and deploying systems. We also are going to be having conversations around security. That's a core enterprise value proposition and ultimately, management. You want to not just declare that the world is hybrid and multi-cloud, but provide solutions to that and we believe we have strong answers to how to bring these pieces together and enable people to successfully move at the rate and pace of innovation that they need, yet in a secure context, and leverage the ability to deploy cloud capabilities where their data currently is, be that on private or public context. >> Hillery Hunter, CTO and vice president of cloud infrastructure at IBM, thanks for talking to theCUBE here today at the IBM Innovation Day. >> Thank you so much for having me. It was a pleasure. >> And, we will be back momentarily with more conversations at IBM Innovation Day.

(digital music) >> From Yorktown Heights, New York It's theCUBE! Covering IBM Cloud Innovation Day. Brought to you by IBM. >> Hi, it's Wikibon's Peter Burris again. We're broadcasting on The Cube from IBM Innovation Day at the Thomas J Watson Research Laboratory in Yorktown Heights, New York. Have a number of great conversations, and we got a great one right now. Rob Thomas, who's the General Manager of IBM Analytics, welcome back to theCUBE. >> Thanks Peter, great to see you. Thanks for coming out here to the woods. >> Oh, well it's not that bad. I actually live not to far from here. Interesting Rob, I was driving up the Taconic Parkway and I realized I hadn't been on it in 40 years, so. >> Is that right? (laugh) >> Very exciting. So Rob let's talk IBM analytics and some of the changes that are taking place. Specifically, how are customers thinking about achieving their AI outcomes. What's that ladder look like? >> Yeah. We call it the AI ladder. Which is basically all the steps that a client has to take to get to get to an AI future, is the best way I would describe it. From how you collect data, to how you organize your data. How you analyze your data, start to put machine learning into motion. How you infuse your data, meaning you can take any insights, infuse it into other applications. Those are the basic building blocks of this laddered AI. 81 percent of clients that start to do something with AI, they realize their first issue is a data issue. They can't find the data, they don't have the data. The AI ladder's about taking care of the data problem so you can focus on where the value is, the AI pieces. >> So, AI is a pretty broad, hairy topic today. What are customers learning about AI? What kind of experience are they gaining? How is it sharpening their thoughts and their pencils, as they think about what kind of outcomes they want to achieve? >> You know, its... For some reason, it's a bit of a mystical topic, but to me AI is actually quite simple. I'd like to say AI is not magic. Some people think it's a magical black box. You just, you know, put a few inputs in, you sit around and magic happens. It's not that, it's real work, it's real computer science. It's about how do I put, you know, how do I build models? Put models into production? Most models, when they go into production, are not that good, so how do I continually train and retrain those models? Then the AI aspect is about how do I bring human features to that? How do I integrate that with natural language, or with speech recognition, or with image recognition. So, when you get under the covers, it's actually not that mystical. It's about basic building blocks that help you start to achieve business outcomes. >> It's got to be very practical, otherwise the business has a hard time ultimately adopting it, but you mentioned a number of different... I especially like the 'add the human features' to it of the natural language. It also suggests that the skill set of AI starts to evolve as companies mature up this ladder. How is that starting to change? >> That's still one of the biggest gaps, I would say. Skill sets around the modern languages of data science that lead to AI: Python, AR, Scala, as an example of a few. That's still a bit of a gap. Our focus has been how do we make tools that anybody can use. So if you've grown up doing SPSS or SaaS, something like that, how do you adopt those skills for the open world of data science? That can make a big difference. On the human features point, we've actually built applications to try to make that piece easy. Great example is with Royal Bank of Scotland where we've created a solution called Watson Assistant which is basically how do we arm their call center representatives to be much more intelligent and engaging with clients, predicting what clients may do. Those types of applications package up the human features and the components I talked about, makes it really easy to get AI into production. >> Now many years ago, the genius Turing, noted the notion of the Turing machine where you couldn't tell the difference between the human and a machine from an engagement standpoint. We're actually starting to see that happen in some important ways. You mentioned the call center. >> Yep. >> How are technologies and agency coming together? By that I mean, the rate at which businesses are actually applying AI to act as an agent for them in front of customers? >> I think it's slow. What I encourage clients to do is, you have to do a massive number of experiments. So don't talk to me about the one or two AI projects you're doing, I'm thinking like hundreds. I was with a bank last week in Japan, and they're comment was in the last year they've done a hundred different AI projects. These are not one year long projects with hundreds of people. It's like, let's do a bunch of small experiments. You have to be comfortable that probably half of your experiments are going to fail, that's okay. The goal is how do you increase your win rate. Do you learn from the ones that work, and from the ones that don't work, so that you can apply those. This is all, to me at this stage, is about experimentation. Any enterprise right now, has to be thinking in terms of hundreds of experiments, not one, not two or 'Hey, should we do that project?' Think in terms of hundreds of experiments. You're going to learn a lot when you do that. >> But as you said earlier, AI is not magic and it's grounded in something, and it's increasingly obvious that it's grounded in analytics. So what is the relationship between AI analytics, and what types of analytics are capable of creating value independent of AI? >> So if you think about how I kind of decomposed AI, talked about human features, I talked about, it kind of starts with a model, you train the model. The model is only as good as the data that you feed it. So, that assumes that one, that your data's not locked into a bunch of different silos. It assumes that your data is actually governed. You have a data catalog or that type of capability. If you have those basics in place, once you have a single instantiation of your data, it becomes very easy to train models, and you can find that the more that you feed it, the better the model's going to get, the better your business outcomes are going to get. That's our whole strategy around IBM Cloud Private for Data. Basically, one environment, a console for all your data, build a model here, train it in all your data, no matter where it is, it's pretty powerful. >> Let me pick up on that where it is, 'cause it's becoming increasingly obvious, at least to us and our clients, that the world is not going to move all the data over to a central location. The data is going to be increasingly distributed closer to the sources, closer to where the action is. How does AI and that notion of increasing distributed data going to work together for clients. >> So we've just released what's called IBM Data Virtualization this month, and it is a leapfrog in terms of data virtualization technology. So the idea is leave your data where ever it is, it could be in a data center, it could be on a different data center, it could be on an automobile if you're an automobile manufacturer. We can federate data from anywhere, take advantage of processing power on the edge. So we're breaking down that problem. Which is, the initial analytics problem was before I do this I've got to bring all my data to one place. It's not a good use of money. It's a lot of time and it's a lot of money. So we're saying leave your data where it is, we will virtualize your data from wherever it may be. >> That's really cool. What was it called again? >> IBM Data Virtualization and it's part of IBM Cloud Private for Data. It's a feature in that. >> Excellent, so one last question Rob. February's coming up, IBM Think San Francisco thirty plus thousand people, what kind of conversations do you anticipate having with you customers, your partners, as they try to learn, experiment, take away actions that they can take to achieve their outcomes? >> I want to have this AI experimentation discussion. I will be encouraging every client, let's talk about hundreds of experiments not 5. Let's talk about what we can get started on now. Technology's incredibly cheap to get started and do something, and it's all about rate and pace, and trying a bunch of things. That's what I'm going to be encouraging. The clients that you're going to see on stage there are the ones that have adopted this mentality in the last year and they've got some great successes to show. >> Rob Thomas, general manager IBM Analytics, thanks again for being on theCUBE. >> Thanks Peter. >> Once again this is Peter Buriss of Wikibon, from IBM Innovation Day, Thomas J Watson Research Center. We'll be back in a moment. (techno beat)

>> Narrator: Live from Munich, Germany, it's theCube, covering IBM, Fast Track Your Data. Brought to you by IBM. >> Welcome to Fast Track Your Data, everybody, welcome to Munich, Germany, this is theCube, the leader in live tech coverage, I'm Dave Vellante with my co-host Jim Kobielus. Dinesh Nirmal is here, he's the vice president of IBM Analytics Development, of course, at IBM, and he's joined by Roland Voelskow, who is the Portfolio Executive at T-Systems, which is a division of Deutche Telekom. Gentlemen, welcome to theCube, Dinesh, good to see you again. >> Thank you. Roland, let me start with you. So your role inside T-Systems, talk about that a little bit. >> Yeah, so thank you for being here, at T-Systems we serve our customers with all kinds of informal hosting services, from infrastructure up to application services, and we have recently, I'd say, about five years ago started to standardize our offerings as a product portfolio and are now focusing on coming from the infrastructure and infrastructure as a service offerings. We are now putting a strong effort in the virtualization container, virtualization to be able to move complete application landscapes from different platforms from, to T-Systems or between T-Systems platforms. The goal is to make, to enable customers to talk with us about their application needs, their business process needs, and have everything which is related to the right place to run the application will be managed automatically by our intelligent platform, which will decide in a multi-platform environment if an application, particularly a business application runs on high available private cloud or a test dev environment, for example, could run on a public cloud, so the customer should not need to deal with this kind of technology questions anymore, so we want to cover the application needs and have the rest automated. >> Yeah, we're seeing a massive trend in our community for organizations like yours to try to eliminate wherever possible undifferentiated infrastructure management, and provisioning of hardware, and Lund management and those things that really don't add value to the business trying to support their digital transformations and raise it up a little bit, and that's clearly what you just described, right? >> Roland: Exactly. >> Okay, and one of those areas that companies want to invest, of course, is data, you guys here in Munich, you chose this for a reason, but Dinesh, give us the update in what's going on in your world and what you're doing here, in Fast Track Your Data. >> Right, so actually myself and Roland was talking about this yesterday. One of the challenges our clients, customers have is the hybrid data management. So how do you make sure your data, whether it's on-premise or on the cloud, you have a seamless way to interact with that data, manage the data, govern the data, and that's the biggest challenge. I mean, lot of customers want to move to the cloud, but the critical, transactional data sits still on-prem. So that's one area that we are focusing in Munich here, is, especially with GDPR coming in 2018, how do we help our customers manage the data and govern the data all through that life cycle of the data? >> Okay, well, how do you do that? I mean, it's a multi-cloud world, most customers have, they might have some Bluemix, they might have some Amazon, they have a lot of on-prem, they got mainframe, they got all kinds of new things happening, like containers, and microservices, some are in the cloud, some are on-prem, but generally speaking, what I just described is a series of stovepipes, they each have their different lifecycle and data lifecycle and management frameworks. Is it your vision to bring all of those together in a single management framework and maybe share with us where you are on that journey and where you're going. >> Exactly, that's exactly our effort right now to bring every application service which we provide to our customers into containerized version which we can move across our platforms or which we can also transform from the external platforms from competition platforms, and onboard them into T-Systems when we acquire new customers. Is also a reality that customers work with different platforms, so we want to be the integrator, and so we would like to expand our product portfolio as an application portfolio and bring new applications, new, attractive applications into our application catalog, which is the containerized application catalog, and so here comes the part, the cooperation with IBM, so we are already a partner with IBM DB2, and we are now happy to talk about expanding the partnership into hosting the analytics portfolio of IBM, so we bring the strength of both companies together the marked excess credibility, security, in terms of European data law for T-Systems, from T-Systems, and the very attractive analytics portfolio of IBM so we can bring the best pieces together and have a very attractive offering to the market. >> So Dinesh, how does IBM fulfill that vision? Is it a product, is it a set of services, is it a framework, series of products, maybe you could describe in some more depth. >> Yeah, it all has to start with the platform. So you have the underlying platform, and then you build what you talked about, that container services on top of it, to meet the need of our enterprise customers, and then the biggest challenge is that how do you govern the data through the lifecycle of that data, right? Because that data could be sitting on-prem, data could be sitting on cloud, on a private cloud, how do you make sure that you can take that data, who touched the data, where that tech data went, and not just the data, but the analytical asset, right, so if your model's built, when was it deployed, where was it deployed? Was it deployed in QA, was it deployed in development? All those things have to be governed, so you have one governance policy, one governance console that you can go as a CDO to make sure that you can see where the data is moving and where the data is managed. So that's the biggest challenge, and that's what we are trying to make sure that, to our enterprise customers, we solve that problem. >> So IBM has announced at this show a unified governance catalog. Is that an enabler for this-- >> Dinesh: Oh, yeah. >> capability you're describing here? >> Oh yeah, I mean, that is the key piece of all of this would be the unified governance, >> Jim: Right. >> which is, you have one place to go govern that data as the CDO. >> And you've mentioned, as has Roland, the containerization of applications, now, I know that DB2 Developer Community Edition, the latest version, announced at this show, has the ability to orchestrate containerized applications, through Kubernetes, can you describe how that particular tool might be useful in this context? And how you might play DB2 Developer Community Edition in an environment where you're using the catalog to manage all the layers of data or metadata or so forth associated with these applications. >> Right, so it goes back to Dave's question, How do you manage the new products that's coming, so our goal is to make every product a container. A containerized way to deliver, so that way you have a doc or registry where you can go see what the updates are, you can update it when you're ready, all those things, but once you containerize the product and put it out there, then you can obviously have the governing infrastructures that sits on top of it to make sure all those containerized products are being managed. So that's one step towards that, but to go back to your DB2 Community Edition, our goal here is how do we simplify our product for our customers? So if you're a developer, how can we make it easy enough for you to assemble your application in matter of minutes, so that's our goal, simplify, be seamless, and be able to scale, so those are the three things we focused on the DB2 Community Edition. >> So in terms of the simplicity aspect of the tool, can you describe a few features or capabilities of the developer edition, the community edition, that are simpler than in the previous version, because I believe you've had a community edition for DB2 for developers for at least a year or two. Describe the simplifications that are introduced in this latest version. >> So one, I will give you is the JSON support. >> Okay. >> So today you want to combine the unstructured data with structured data? >> Yeah. >> I mean, it's simple, what we have a demo coming up in our main tent, where asset dialup, where you can easily go, get a JSON document put it in there, combined with your structured data, unstructured data, and you are ready to go, so that's a great example, where we are making it really easy, simple. The other example is download and go, where you can easily download in less than five clicks, less than 10 minutes, the product is up and running. So those are a couple of the things that we are doing to make sure that it is much more simpler, seamless and scalable for our customers. >> And what is Project Event Store, share with us whatever you can about that. >> Dinesh: Right. >> You're giving a demo here, I think, >> Dinesh: Yeah, yeah. >> So what is it, and why is it important? >> Yeah, so we are going to do a demo at the main tent on Project Event Store. It's about combining the strength of IBM Innovation with the power of open source. So it's about how do we do fast ingest, inserts into a object store, for example, and be able to do analytics on it. So now you have the strength of not only bringing data at very high speed or volume, but now you can do analytics on it. So for example, just to give you a very high level number we can do more than one million inserts per second. More than one million. And our closest competition is at 30,000 inserts per second. So that's huge for us. >> So use cases at the edge, obviously, could take advantage of something like this. Is that sort of where it's targeted? >> Well, yeah, so let's say, I'll give you a couple of examples. Let's say you're a hospital chain, you want the patient data coming in real time, streaming the data coming in, you want to do analytics on it, that's one example, or let's say you are a department store, you want to see all the traffic that goes into your stores and you want to do analytics on how well your campaign did on the traffic that came in. Or let's say you're an airline, right? You have IOT data that's streaming or coming in, millions of inserts per second, how do you do analytics, so this is, I would say this is a great innovation that will help all kinds of industries. >> Dinesh, I've had streaming price for quite awhile and fairly mature ones like IBM Streams, but also the structured streaming capability of Spark, and you've got a strong Spark portfolio. Is there any connection between Product Event Store and these other established IBM offerings? >> No, so what we have done is, like I said, took the power of open source, so Spark becomes obviously the execution engine, we're going to use something called the Parquet format where the data can be stored, and then we obviously have our own proprietary ingest Mechanism that brings in. So some similarity, but this is a brand new work that we have done between IBM research and it has been in the works for the last 12 to 18 months, now we are ready to bring it into the market. >> So we're about out of time, but Roland, I want to end with you and give us the perspective on Europe and European customers, particular, Rob Thomas was saying to us that part of the reason why IBM came here is because they noticed that 10 of the top companies that were out-performing the S&P 500 were US companies. And they were data-driven. And IBM kind of wanted to shake up Europe a little bit and say, "Hey guys, time to get on board." What do you see here in Europe? Obviously there are companies like Spotify which are European-based that are very data-driven, but from your perspective, what are you seeing in Europe, in terms of adoption of these data-driven technologies and to use that buzzword. >> Yes, so I think we are in an early stage of adoption of these data-driven applications and analytics, and the European companies are certainly very careful, cautious about, and sensitive about their data security. So whenever there's news about another data leakage, everyone is becoming more cautious and so here comes the unique, one of the unique positions of T-Systems, which has history and credibility in the market for data protection and uninterrupted service for our customers, so that's, we have achieved a number of cooperations, especially also with the American companies, where we do a giant approach to the European markets. So as I said, we bring the strength of T-Systems to the table, as the very competitive application portfolio, analytics portfolio, in this case, from our partner IBM, and the best worlds together for our customers. >> All right, we have to leave it there. Thank you, Roland, very much for coming on. Dinesh, great to see you again. >> Dinesh: Thank you. >> All right, you're welcome. Keep it right there, buddy. Jim and I will be back with our next guests on theCube. We're live from Munich, Germany, at Fast Track Your Data. Be right back.