>> Welcome to this IBM Watson Health client conversation. And we're probing the dynamics of the relationship between IBM and it's clients. We're going to look back at some of the challenges of 2020 and look forward to, you know, present year's priorities. We'll also touch on the future state of healthcare. My name is Dave Vellante. I'll be your host and I'm from theCUBE. And with me are Doctor Alex Towbin, who's Associate Chief Clinical Operations and Informatics at Cincinnati ChilDoctoren's Hospital and John Chrisman of course from IBM Watson health. Welcome gentlemen, Good to see you. Thanks for coming on. >> Thanks for having us. >> Yeah, thanks for having me. >> Yeah I know from talking to many clients around the world, of course virtually this past year, 11 months or so that relationships with technology partners they've been critical over during the pandemic to really help folks get through that. Not that we're through it yet but, we're still through the year now, there's I'm talking professionally and personally and Doctor Towbin, I wonder if you could please talk about 2020 and what role the IBM partnership played in helping Cincinnati children's, you know press on in the face of incredible challenges? >> Yeah, I think our story of 2020 really starts before the pandemic and we were fortunate to be able to plan a disaster and do disaster drill scenarios. And so, as we were going through those disaster drill scenarios, we were trying to build a solution that would enable us to be able to work if all of our systems were down and we worked with IBM Watson Health to design that solution to implement it, it involves using other solutions from our primary one. And we performed that disaster drill in the late January, early February timeframe of 2020. And while that drill had nothing to do with COVID it got us thinking about how to deal with a disaster, how to prepare for a disaster. And so we've just completed that and COVID was coming on the horizon. I'm starting to hear about it coming into the U.S for the first time. And we took that very seriously on our department. And so, because we had prepared for this this disaster drill had gone through the entire exercise and we built out different scenarios for what could happen with COVID what would be our worst case scenarios and how we would deal with them. And so we were able to then bring that to quickly down to two options on how our department and our hospital would handle COVID and deal with that within the radiology department and like many other sites that becomes options of working from home or working in a isolated way and an and an office scenario like where I'm sitting now and we planned out both scenarios and eventually made the decision. Our decision at that point was to work in our offices. We're fortunate to have private offices where we can retreat to and something like that. And so then our relationship with IBM was helpful and that we needed to secure more pieces of hardware. And so even though IBM is our PACS vendor and our enterprise imaging vendor, they also help us to secure the high resolution monitors that are needed. And we needed a large influx of those during the pandemic and IBM was able to help us to get those. >> Wow! So yeah you were able to sort of test your organization resilience before the pandemic. I mean, John, that's quite an accomplishment for last year. I'm sure there are many others. I wonder if one of you could pick it up from here and bring your perspectives into it and, you know maybe ask any questions that you would like to ask them. >> Yeah, sure, Doctor Towbin, that's great that we were able to help you with the hardware and procure things. So I'm just curious before the pandemic how many of the radiologists ever got to read from home, was that a luxury back then? And then post pandemic, are you guys going to shift to how many are on-site versus remote? >> Yeah, so we have a couple of scenarios. We've had talk about it both from our PACS perspective as well as from our VNA enterprise imaging perspective from PACS perspective we always designed our solution to be able to work from a home machine. Our machines, people would access that through a hospital-based VPN. So they would log in directly to VPN and then access the PACS that way. And that worked well. And many of our radiologists do that particularly when they're on call works best for our neuroradiologist who are on call a little bit more frequently. And so they do read from home in that scenario. With enterprise imaging and are used to the enterprise viewer and iConnect access. We always wanted that solution to work over the internet. And so it's set up securely through the internet but not through the VPN. And we have radiologists use that as a way to view studies from home, even not from home, so it can be over one of their mobile devices, such as an iPad and could be at least reviewing studies then. We, for the most part for our radiologist in the hospital that's why we made the decision to stay in the hospital. At COVID time, we have such a strong teaching mission in our department in such a commitment to the education of our trainees. We think that hospital being in the hospital is our best way to do that, it's so hard. We find to do it over something like zoom or other sharing screen-sharing technology. So we've stayed in and I think we'll continue to stay in. There will be some of those needs from a call perspective for example, reading from home, and that will continue. >> And then what's your success been with this with the technology and the efficiency of reading from home? Do you feel like you're just as efficient when you're at home versus onsite? >> The technology is okay. The, our challenges when we're reading from the PACS which is the preferred way to do it rather than the enterprise archive, the challenge is we have to use the PACS So we have to be connected through VPN which limits our bandwidth and that makes it a little bit slower to read. And also the dictation software is a little bit slower when we're doing it. So moving study to study that rapid turnover doesn't happen but we have other ways to make, to accelerate the workflow. We cashed studies through the worklist. So they're on the machine, they load a little bit more rapidly and that works pretty well. So not quite as fast, but not terrible. >> We appreciate your partnership. I know it's been going on 10 years. I think you guys have a policy that you have to look at the market again every 10 years. So what do you think of how the market's changed and how we've evolved with the VNA and with the zero footprint viewers? A lot of that wasn't available when you initially signed up with Amicas years ago, so. >> Yeah, we signed up so we've been on this platform and then, you know now the IBM family starting in 2010, so it's now now 11 years that we're, we've been on as this version of the PACS and about eight, seven or eight years from the iConnect platform. And through that, we've seen quite an evolution. We were one of the first Amicas clients to be on version six and one of the largest enterprises. And that went from, we had trouble at the launch of that product. We've worked very closely with Amicas then to merge. And now IBM from the development side, as well as the support side to have really what we think is a great product that works very well for us and drives our entire workflow all the operations of our department. And so we've really relished that relationship with now IBM. And it's been a very good one, and it's allowed us to do the things like having disaster drill planning that we talked about earlier as far as where I see the market I think PACS in particular is on the verge of the 3.0 version as a marketplace. So PACSS 1 one was about building the packs, I think, and and having electronic imaging digital imaging, PACS 2.0 is more of web-based technology, getting it out of those private networks within a radiology department. And so giving a little bit more to the masses and 3.0 is going to be more about incorporating machine learning. I really see that as the way the market's going to go and to where I think we're at the infancy of that part of the market now about how do you bring books in for machine learning algorithms to help to drive workflow or to drive some image interpretation or analysis, as far as enterprise imaging, we're on the cusp of a lot there as well. So we've been really driving deep with enterprise imaging leading nationally enterprise imaging and I have a role in the MSAM Enterprise Imaging Community. And through all of that work we've been trying to tackle works well from enterprise imaging point of view the challenges are outside of radiology, outside of cardiology and the places where we're trying to deal with medical photos, the photographs taken with a smart device or a digital camera of another type, and trying to have workflow that makes sense for providers not in those specialty to that don't have tools like a DICOM modality workloads store these giant million-dollar MRI scanners that do all the work for you, but dealing with off the shelf, consumer electronics. So making sure the workflow works for them, trying to tie reports in trying to standardize the language around it, so how do we tag photos correctly so that we can identify relevancy all of those things we're working through and are not yet standard within our, within the industry. And so we're doing a lot there and trying and seeing the products in the marketplace continuing to evolve around that on the viewer side, there's really been a big emergence as you mentioned about the zero footprint viewers or the enterprise viewer, allowing easy access easy viewing of images throughout the enterprise of all types of imaging through obtained in the enterprise and will eventually incorporate video pathology. The market is also trying to figure out if there can be one type of viewer that does them all that and so that type of universal viewer, a viewer that cardiologists can use the same as a radiologist the same as a dermatologist, same as a pathologist we're all I think a long way away from that. But that's the Marcus trying to figure those two things out. >> Yeah, I agree with you. I agree with your assessment. You talked about the non DICOM areas, and I know you've you've partnered with us, with ImageMover and you've got some mobile device capture taking place. And you're looking to expand that more to the enterprise. Are you also starting to use the XDS registry? That's part of the iConnect enterprise archive, or as well as wrapping things in DICOM, or are you going to stick with just wrapping things in DICOM? >> Yeah, so far we've been very bunched pro DICOM and using that throughout the enterprise. And we've always thought, or maybe we've evolved to think that there is going to be a role for XDS are I think our early concerns with XDS are the lack of other institutions using it. And so, even though it's designed for portability if no one else reads it, it's not portable. If no one else is using that. But as we move more and more into other specialties things like dermatology, ophthalmology, some of the labeling that's needed in those images and the uses, the secondary uses of those images for education, for publication, for dermatology workflow or ophthalmology workflow, needs to get back to that native file and the DICOM wrap may not make sense for them. And so we've been actively talking about switching towards XDS for some of the non DICOM, such as dermatology. We've not yet done that though. >> Given the era children's hospital has the impact on your patient load, then similar to what regular adult hospitals are, or have you guys had a pretty steady number of studies over the last year? >> In relay through the pandemic, we've had, it has been decreased, but children fortunately have not been as severely affected as adults. There is definitely disease in children and we see a fair amount of that. There are some unique things that happen in kids but that fortunately rare. So there's this severe inflammatory response that kids can get and can cause them to get very sick but it is quite rare. Our volumes are, I think I'm not I think our volumes are stable and our advanced imaging things like CT, MRI, nuclear medicine, they're really most decreased in radiography. And we see some weird patterns, inpatient volumes are relatively stable. So our single view chest x-rays, for example, have been stable. ER, visits are way down because people are either wearing masks, isolating or not wanting to come to the ER. So they're not getting sick with things like the flu or or even common colds or pneumonias. And so they're not coming into the ER as much. So our two view x-rays have dropped by like 30%. And so we were looking at this just yesterday. If you follow the graphs for the two we saw a dip of both around March, but essentially the one view chest were a straight line and the two view chest were a straight line and in March dropped 30 to 50% and then stayed at that lower level. Other x-rays are on the, stay at that low level side. >> Thanks, I know in 2021 we've got a big upgrade coming with you guys soon and you're going to stay in our standalone mode. I understand what the PACSS and not integrate deeply to the VNA. And so you'll have a couple more layers of storage there but can you talk about your excitement about going to 8.1 and what you're looking forward to based on your testimony. >> Yeah we're actually in, we're upgrading as we're talking which is interesting, but it's a good time for talking. I'm not doing that part of the work. And so our testing has worked well. I think we're, we are excited. We, you know, we've been on the product as I mentioned for over 10 years now. And for many of those years we were among the first, at each version. Now we're way behind. And we want to get back up to the latest and greatest and we want to stay cutting edge. There've been a lot of reasons why we haven't moved up to that level, but we do. We're very careful in our testing and we needed a version that would work for us. And there were things about previous versions that just didn't and as you mentioned, we're staying in that standalone mode. We very much want to be on the integrated mode in our future because enterprise imaging is so important and understanding how the comparisons fit in with the comparison in dermatology or chest wall deformity clinic, or other areas how those fit into the radiology story is important and it helped me as a radiologist be a better radiologist to see all those other pictures. So I want them there but we have to have the workflow, right. And so that's the part that we're still working towards and making sure that that fits so we will get there. It'll probably be in the next year or two to get to that immigrating mode. >> As you, look at the number of vendors you have I think you guys prefer to have less vendor partners than than more I know in the cardiology area you guys do some cardiology work. What has been the history or any, any look to the future of that related to enterprise imaging? Do you look to incorporate more of that into a singular solution? >> Cardiology is entirely part of our enterprise imaging solution. We all the cardiology amendments go to our vendor neutral archive on the iConnect platform. All of them are viewed across the enterprise using our enterprise viewer. They have their unique specialty viewer which is, you know, fine. I'm a believer that specialty, different specialties, deserve to have their specialty viewers to do theirs specialty reads. And at this point I don't think the universal viewer works or makes sense until we have that. And so all the cardiology images are there. They're all of our historical cardiology images are migrated and part of our enterprise solution. So they're part of the entire reference the challenge is they're just not all in PACSS. And so that's where, you know, an example, great example, why we need to get to this to the integrated mode to be able to see those. And the reason we didn't do that is the cardiology archive is so large to add a storage to the PACS archive. Didn't make sense if we knew we were going to be in an integrated mode eventually, and we didn't want to double our PACS storage and then get rid of it a couple of years later. >> So once you're on a new version of merge PACS and you're beyond this, what are your other goals in 2021? Are you looking to bring AI in? Are you using anybody else's AI currently? >> Yeah, we do have AI clinical it's phone age, so it's not not a ton of things but we've been using it clinically, fully integrated, it launches. When I open a study, when I opened a bone age study impacts it launches we have a bone age calculator as well that we've been using for almost two decades now. And so that we have to use that still but launching that automatically includes the patient's sex and birth date, which are keys for determining bone age, and all that information is there automatically. But at the same time, the images are sent to the machine learning algorithm. And in the background the machine determines a bone age that in the background it sends it straight to our dictation system and it's there when we opened the study. And so if I agree with that I signed the report and we're done. If I disagree, I copy it from my calculator and put it in until it takes just a couple of clicks. We are working on expanding. We've done a lot of research in artificial intelligence and the department. And so we've been things are sort of in the middle of translation of moving it from the research pure research realm to the clinical realm, something we're actively working on trying to get them in. Others are a little bit more difficult. >> That's the question on that John, Doctor, when you talk about injecting, you know machine intelligence into the equation. >> Yeah. >> What, how do you sort of value that? Does that give you automation? Does it improve your quality? Does it speed the outcome and maybe it's all of those but how do you sort of evaluate the impact to your organisation? >> I there's a lot of ways you can do it. And you touched on one of my favorite one of my favorite talking points, in a lot of what we've been doing and early machine learning is around image interpretation helping me as a radiologist to see a finding. Unfortunately, most of the things are fairly simple tasks that it's asking us to do. Like, is there a broken bone? Yes or no, I'm not trying to sound self-congratulatory or anything, but I'm really good at finding broken bones. I get, I've been doing it for a long time and, and radio, you know so machines doing that, they're going to perform as well as I can perform, you know, and that's the goal. Maybe they'll perform a little bit better maybe a little bit worse but we're talking tiny increments there they're really to me, not much value of that it's not something I would want. I don't value that at a time where I think machine learning can have real value around more on some of the things that you mentioned. So can it make me more efficient? Can it do the things that are so annoying that and they'd take, they're so tedious that they make me unhappy. A lot of little measurements for example are like that an example. So in a patient with cancer, we measure a little tumors everywhere and that's really important for their care, but it's tedious and so if a machine could do that in an automated way and I checked it that, you know, patient when because he or she can get that good quality care and I have a, you know, a workflow efficiency game. So that one's important. Another one that would be important is if the machine can see things I can't see. So I'm really good at finding fractures. I'm not really good at understanding what all the pixels mean and, you know in that same patient with cancer, oh what do all the pixels mean in that tumor? I know it's a tumor. I can see the tumor, I can say it's a tumor but sometimes those pixels have a lot of information in them and may give us prognosis, you know, say that this patient may, maybe this patient will do well with this specific type of chemotherapy or a specific or has a better prognosis with one with one drug compared to another. Those are things that we can't usually pick out. You know, it's beyond the level of that are I can perceive that one is really the cutting edge of machine learning. We're not there yet and then the other thing are things that, you know just the behind the scenes stuff that I don't necessarily need to be doing, or, you know so it's the non interpretive artificial intelligence. >> Dave: Right. >> And that's what I've been also trying to push. So an example of when the algorithms that we've been developing here we check airways. And this is a little bit historical in our department, but we want to make sure we're not missing a severe airway infection. That can be deadly, it's incredibly rare. Vaccines have made it go away completely but we still check airways. And so what happens is the technologist takes the x-ray. They come in to ask us if it's okay, we are interrupted from what we're doing. We open up the study, say yes or no. Okay, not okay, if it's not okay they go back, take another study. Then come back to us again and say, is it okay or not? And we repeat this a couple of times it takes them time that they don't need to spend and takes us time. And so we have, we've built an algorithm where the machine can check that and their machine is as good or a little bit worse than us, but give can give that feedback. >> Dave: Got it. >> The challenge is getting that feedback to the technologist quickly. And so that's, that's I think part for us to work on stuff. >> Thank you for that. So, John, we've probably got three or four minutes left. I'll let you bring it home and appreciate that Doctor Towbin >> I think one of the biggest impacts probably I knew this last year with the pandemic, Doctor Towbin is this, I know you're a big foodie. So having been to some good restaurants and dinners with the hot nurse in a house how's the pandemic affected you personally. And some of the things you like to do outside of work. >> Everything is shut down. And everything has changed. I have not left the house since March besides come to work and my family hasn't either. And so we're hardcore quarantining and staying you know, staying out and keeping it home. So we've not gone out to dinner or done much else. >> So its DoorDash and Uber Eats or just learned to cook at home. >> It's all cooking at home. We're fortunate, my wife loves to cook. My kids love to cook. I enjoy cooking, but I don't have the time as often. So we've done a lot of different are on our own experimenting. Maybe when the silver lining one of the things I've really relished about all this is all this time I get to spend with my family. And that closeness that we've been able to achieve because of being confined in our house the whole time. And so I've played get to play video games with my kids every night. We'd been on a big Fortnite Keck lately since it's been down making. So we've been playing that every night since we've watched movies a lot. And so as a family, we've, I it's something I'll look back fondly even though it's been a very difficult time but it's been an enjoyable time. >> I agree, I've enjoyed more family time this year as well, but final question is in 2021, beyond the PACS upgrade what are the top other two projects that you want to accomplish with us this year? And how can we help you? >> I think our big one is are the big projects are unexpanded enterprise imaging. And so we want to continue rolling out to other areas that will include eventually incorporating scopes, all the images from the operating room. We need to be able to get into pathology. I think the pathology is really going to be a long game. Unfortunately, I've been saying that already for 10 years and it's still probably another 10 years ago but we need to go. We can start with the gross pathology images all the pictures that we take for tumor boards and get those in before we start talking about whole slide scanning and getting in more of the more of the photographs in the institution. So we have a route ambulatory but we need inpatient and ER. >> All right one last question. What can IBM do to be a better partner for you guys? >> I think it's keep listening keep listening and keep innovating. And don't be afraid to be that innovative partner sort of thinking as the small company that startup, rather than the giant bohemoth that can sometimes happen with large companies, it's harder. It is fear to turn quickly, but being a nimble company and making quick decisions, quick innovations. >> Great, quick question. How would you grade IBM, your a tough grader? >> It depends on what I am a tough grader but it depends on what, you know as the overall corporate partnership? >> Yeah the relationship. >> I'd say it's A minus. >> Its pretty good. >> I think, I mean, I, we get a lot of love from IBM. I'm talking specifically in the imaging space. I not, maybe not, I don't know as much on the hardware side but we, yeah, we have a really good relationship. We feel like we're listened to and we're valued. >> All right, well guys, thanks so much. >> So even if it's not an A plus- >> Go ahead. >> I think there's some more to, you know, from the to keep innovating side there's little things that we just let you know we've been asking for that we don't always get but understand the company has to make business decisions not decisions on what's best for me. >> Of course got to hold that carrot out too. Well thanks guys, really appreciate your time. Great conversation. >> Yeah, thank you. >> All right and thank you for spending some time with us. You're watching client conversations with IBM Watson Health.

>> From Boston, Massachusetts. It's theCUBE. Covering Actifio 2019 Data Driven. (upbeat techno music) Brought to you by Actifio. >> Hello everyone and welcome to Boston and theCUBE's special coverage of Actifio Data Driven 19. I'm Dave Vellante. Stu Miniman is here. We've got a special guest, John Furrier is in the house from from Palo Alto. Guys, theCUBE we love to go out on the ground, you know, we go deep. We're here at this data theme, right? We were there in the early days, John, you called me up and say, "Get your butt here, we're going to cover the first of Doop World". And since then things have moved quite fast. Everybody thought, you know, Hadoop Big Data was going to take over the world. Nobody even uses that term anymore, right? It's kind of, now it's AI, and machine intelligence, and block chain, and everything else. So what do you think is happening? Did the early Big Data days fail? You know, Frank Genus this morning called it The experimentation phase. >> I mean, I don't really think Frank has a good handle on what's going on in my opinion, cause I think it's not an experimentation, it's real. That was a wave that was essentially the beginning of, not an experimentation, of realization and reality that data, unstructured data in particular was real and relevant. Hadoop looked good off the tee, mill the fairway as we say, but the thing about the Hadoop ecosystem is that validated big data. Every financial institution jumped on it. Everyone who knew anything about data or had data issues or had a lot of data, knew the value. It's just that the apparatus to build via Hadoop was too expensive. In comes Cloud computing at scale, so, as Cloud was accelerating, you look at the Amazon Web Services Revenue Chart you can almost see the D mark where the inflection point is on the hockey stick of Amazon's revenue numbers. And that is the point in time where Hadoop was on the declining of failure. Hortonworks sold the Cloudera. Cloudera's earnings are at an all-time low. A lot of speculation of their entire strategy, and their venture back company went public, but bet the ranch to be the next data warehouse. That wasn't the business model. The data business was a completely new industry, completely being re-transformed, and, far from experimentation, it is real and definitely growing like a weed, but changing because of the underpinning infrastructure dynamics of Cloud Native, Microservices, and that's only going to get highly accelerated and the people who talk about context of industry like Frank, are going to be off. Their predictions will be off because they don't really see the new picture clear enough, in my opinion, >> So, >> I think he's off. >> So it's not so much of a structural change like it was when we went from, you know, mainframes to PCs, it's more of a sort of flow, evolution into this new area which is being driven, powered by new technologies, we talk about block chain machine intelligence and other things. >> Well, I mean, the make up of companies that were building quote, "Big Data Solutions", were trying to build an apparatus or mechanisms to solve big data problems, but none of them actually had the big data problem. None of them were full of data. None of them had a lot of data. The ones that had problems were the financial institutions, the credit card companies, the people who were doing a lot of large scale, um, with Google, Facebook, and some of the hyperscalers. They were actually dealing with the data tsunami themselves, so the practitioners ended up driving it. You guys at Wikibomb, we pointed this out on theCUBE many times, that the value was going to come from the practitioners not the suppliers of so called technology. So, you know, the Clouderas of the world who thought Hadoop would be relevant and growing as a technology were right on one side, on the other side of the coin was the Cloud decimation of that sector. The Cloud computer just completely blew away that Hadoop market because you didn't have to hire a PhD, you didn't have to hire specialty skills to stand up Hadoop clusters. You could actually throw it in the Cloud and get agile quickly, and get value out of data very very quickly. That has been real, it has not been an experiment. There's been new case studies, new companies born, new brands, so it's not an experiment, it is reality, and it's only going to get more real every day. >> And I add of course now you've got, you mentioned Cloudera and Hortenworks, you also got Matt Bar reeling Stu. Let's talk about Actifio. So they coined the term Copy Data Management, they created the category, of course they do a lot of backup, I mean, everybody in this space does a lot of backup. And then you saw the Silicon Valley companies come in. Particularly Cohesity and Rubric, you know, to a lesser extent he got some other guys like Zerto and Durva, but it was really those two companies, Cohesity and Rubric, they raised more money in their D round than Actifio has since inception. But yet Actifio keeps, you know, plodding along, growing, you know, word is they're profitable, you know, they're not like this really sectioned very East Coast versus kind of West Coast mentality. What's your take on what's going on? >> Yeah, so, Dave right, you look at the early days of Actifio and you say great, Copy Data Management, I have all these copies of data, how do I reduce my cost, get greater utilization than I have and leverage the data? I love the title of the show here, Data Driven. You know, we know at the center of digital transformation if you can't become data driven, like the CMO Brian Regan got up on stage talk about that industrialization of data. How am I going along that journey being this, I collected data versus now, you know, data, you know, is the reason that I make decisions, how I make decisions, I get smarter. The Cloud of course is a huge enabler of this, there's all these services that I can instantly access to be able to get greater insight, and move along with that environment, and if you look underneath all of these backup companies, it's really how I can change that data into business value and drive my business, the metadata underneath and all those pieces, not just the wonky storage and technical solutions that make things better, and I get a faster ROI. It's that data at the core of what we do and how do I get that as a business to accelerate. Because we know IT needs to be able to respond back to the business and data needs to be that rocket fuel. >> Is it the case of data haves and data have-nots? I mean, Amazon has data >> I mean, you're right-- >> and Facebook has data. >> We're talking about Actifio, you brought that up, okay, on this segment, on the inside segment, which is cool, they're here at the event, but they have a good opportunity but they also, they got some challenges. I mean, the thing about Actifio is, to my earlier point, which side of the wave are they on? Are they out too much out front with virtualization and Amazon, the Cloud will take them away, or are they riding the Cloud wave, making that an enabler? And I think what really I like about Actifio is because they have a lot of virtualization capabilities, the question is can they scale that Stu, to containers and microservices, because, the real opportunity in this market, in my opinion, is going to build on the virtualization trend, and make container aware, microservices capabilities because if they don't, then that would be a tell sign. Now either way it's a hot M&A market right now, so I think being in the market, horse on the track as you say. You look at the tableau sales force deal monster numbers we are in clearly a hot IPO market and a major roll up market on the M&A side. I think clearly there's two types of companies, old and new, and that is really what people are looking at, are they part of the old guard, are they the new guard. So, you know, this to me is going to be a tell sign of what they do next, can they make the data driven value proposition, you articulated Stu, actually a reality It's going to come from the technology underneath. >> Well I think it's a really interesting point you're making because, Stu as you probably know, that Amazon announced the Amazon backup service right, and you talked about the backup guys and they're like, "Ah yeah it's backup, but it really doesn't do recovery, it's really not that robust". It's part of me says, "Uh oh"... >> Watch out. >> You better move fast", because Amazon has stated, "Hey if you don't move fast we're going to just keep gobbling", and you've seen Amazon do this. What are your thoughts on that? Can these specialists, can they survive, John's talking about M&A. Can the market support all these guys along with the big, you know, traditional guys like Veritas, and Dell EMC, and IBM and Combol? >> Right, well so Actifio started very much in the data center. They were before this Could wave really took off. It's really only in the last year that they've been sassifying their product. So the question is, does that underlying IP, which wasn't tied to hardware, but, you know, sat at really more of, you know, reminded us of that storage virtualization battles that we talked about for years, Dave, but now they are going in the Cloud. They've got all the partnerships in the Cloud, but they are competing against those new vendors that you talked about like Cohesity and Rubric out there, and there's big money chasing this environment. So, you know, I want to talk to the customers here and find out, you know, where they are using them, and especially some of those first customers using this--. >> Well they clearly need a Cloud play cause that's clearly where the action is. But if you look at what's going on with Amazon, Azure, and Google you see a lot of on premises, Stu, because that's where the customers are. So just because the customers are currently not migrating their existing workloads to the Cloud doesn't mean it's not going to happen. So I think there's an opportunity for any company like Actifio, who may or may not be on the curve on the tech side, one little misfire on a tech bet could cripple the company and also make the company. There's a lot of high risk, reward ratio. How they handle containers. How they build on virtualizations. Virtualization going to to be part of the future with Cloud. These are the kind of the dynamics that are going to be in play, and they got some time on their hands because the on premises growth is because the clients are trying to figure out what to do and they're not going to be migrating, lifting, and shifting workloads all off to the Cloud. New will be Cloud based, but enterprises have proven why we are in multi-Cloud and hybrid-Cloud conversation, that... The enterprise on premises is not going away anytime soon. >> I want to ask you guys, John you specifically, about this sort of new Silicon Valley growth model and how companies are achieving escape velocity. When you and I made our first trip to Barcelona, I was having dinner with David Scott who was the CEO of 3PAR and he said to me, When I came to 3PAR the board said, "Hey we're willing to invest 30 million dollars in this company". And David Scott said to them, "I need way more, I need 80 million dollars". Today 80 million dollars is nothing. You saw, you know, Pure Storage hit escape velocity, was just throwing money, and growing at the problem. You're seeing Cohesity-- >> Well you can debate that. I mean, If you have to build a rocket ship, hit critical mass and you want to fund that, you're going to to need an enterprise. However, there's arguments on the south side that you can actually get fly wheel effect going early with less capital. So again, that's 3PAR-- >> But so that's my point. >> Well so that's 3PAR, that was 2009. >> So, yeah that was early days so that's ancient history. But software is generally supposed to be a capital efficient market, yet these companies are raising many hundreds and hundreds of millions, you know, half a billion dollar raises and they are putting it largely in promotion. Is that the new model, is that sustainable, in your view? >> Well I think you're conflating capital market dynamics with viable companies to invest in. I think there's a robust seed in series A market but the series A market and Silicon Valley is you know, 15 to 25 million, it used to be 3 to 5. So the dynamics are changing on funding. There's just not enough companies, horses on the track, to deploy capital at tranches of 30, 50, 80 million. So the capital markets are clearly going to have the money available so it's a market for the startups and the broke companies. That's separate from actually winning. So you've got slacks going public this weeks, you have other companies who have built business on a sass fly wheel, and then everything else is gravy in terms of the go to market, they got a couple hundred million. I think slack got close to a billion dollars in cash that they've raised. So they're flooded with cash, they'll never spend it all. So there are some companies that can achieve success like that. Others have to buy market share, they got to push and build out a sales force, and it's going to be a function of the role of customer, customization, specialism, and whatnot. But with AI machine leaning there's more efficiencies coming in so I think the modern company can do more with less. >> What do you think of the ride sharing on IPOs, Uber and Lift, do you abol? Do you like 'em or do you think it's just, they're losing too money and can't sustain it? >> I was thinking about that this morning after looking at the article in the Wall Street Journal in our coverage on Silicon angle. You look at Zoom communications, I like models that actually can take a simple concept and an existing mature market and disrupt it by being Cloud efficient and completely sass and data driven. That is an example of success. That to me, Zoom Communications and Zscaler, another company that we talk to, these are companies that were built with a specific value proposition that made the product and they were targeting mature markets with leaders in it. Video conferencing, Webex, Citrix, Zoom came out of nowhere, optimized on simple value proposition, used Cloud scale and data, and crushed it. Uber, Lift, little bit different issue. They're losing money but I would bet on the long term that that is going to be the used case for how people will have transportation. I think that's the long game and I think that without regulatory kind of pressure, without, there's regulatory issues that's really the big risk. But I believe that Uber and Lift absolutely will be long brands and just like Facebook was early on, although they threw off a lot of cash, those guys are building for penetration, and that's where the funding matters. Penetration is critical. Now they're the standard, and people really don't take taxis anymore, but they're really using the ride sharing. And you get the scooters, you get the bikes, they're all sequencing into these adjacent markets which drains more cash but builds the brand, builds the footprint. >> Well that's what I want to ask you. So people compare the early Uber, Lift, Taxi, Ride sharing to Amazon selling books, but there's all these other adjacencies. You have a thought on this? >> Well, just, you know, right, Uber Eats is a huge opportunity for that environment and autonomous vehicles everybody talks about, but it's still quite a ways out. So there are a lot of different- >> Scooters are the same, we're in San Diego, there are 8 gazillion scooters. >> San Diego had fun, you know, going around on their electronic scooters, boy, talk about the gig economy, they pay people at the night, to like go pay by the recharge you do on that, what is the future of work, >> Yeah, that's a great point. >> and how can we have that-- >> Uber going to look a lot like Amazon. You subsidize the front end retail side of the business, but look at the data that they throw up. Uber's data that they're gathering on, not only customer behavior, but just mapping services, 3-D mapping is going to be huge, so you've got these cars that are essentially bots on the road, providing massive mapping and traffic analysis. So you're going to start to see data driven, like Actifio slogan here, be a big part of all design decisions and value proposition from any company out there. And if they're not data driven I think they're going to be toast. >> Probably could because there's that data and that machine learning underneath, that can optimize, you know, where the people are, how I use the system, such a huge wave that we're watching. >> How about one last topic which is heavily data driven, it's Facebook. Facebook is obviously a data driven company, the Facebook crypto play, I love it, I love Facebook. I'm a bull on Facebook, I think it's been beat up. I think, two billion users is hard to replicate, but what's your thoughts on their crypto play? >> Well it's kind of a middle finger to the United States of America but it's a great catalyst for the international market because crypto needed a whale to come in and bring all those users in. Bad timing, in my mind, for Facebook, because given all the anti-trust and regulatory conversations, what better way to show your threat to the world order when you say we're going to run a banking system with a collection of international companies. I think the US is going to look at this and say, "Oh my God! They can't even be trusted to handle personal information and we're going to now let them run a banking system? Run monetary, basically World Bank equivalent infrastructure?" No frickin way! I think this is going to to be a major road to home. I think Facebook has to really make this an ecosystem play if they want to make it work, that's their telegraphic move they're saying, "Hey we want to do for the community but we got our own wallet and we got our own network". But they bring a lot to the table so it's going to be a really interesting dynamic to see the coalescing around Facebook because they could make the market. Look what Instagram did to Snapchat. They literally killed the company, took all their users. That is what's going to happen in the digital money economy when Facebook brings billions of users user experience with money. What happened with Snapchat with Instagram is going to happen to the World Bank if this continues. >> Where do you stand on the government breaking up big tech? >> So Dave, you know, you look in these companies, it's not easy to pull those apart. I don't think our government understands how most of big tech works. You know, take Amazon and AWS, that's one company underneath it. You know, Facebook, Microsoft. You know, Microsoft went through all these issues. Question Dave, we've had lots of debates on Twitter you know, are they breaking the law, are they not doing trust? I have some trust issues with Facebook myself, but most of the big companies up there I don't think the anti-trust kicks in, I don't think it makes sense to pull them apart. >> Stu, the Facebook story and the YouTube story are simply this, they have been hiding under the platform rules, of the Digital Millennium Copyright Act, and they are an editing platform so you can't sue them. Okay, once they become a publisher they could be sued. Just like CNN, Fox News, and everybody else. And we're publishers. So they've been hiding behind the platform. That gig is up. They're going to have to address are you a platform or are you a publisher? You're making editing decisions around what users can see with software, you are essentially editing the feed, that is a publisher role, with that becomes responsibility, and then obviously regulartory. >> Well Facebook is conflicted right now. They're trying to figure out which side of the fence to go on. >> No no no! They want one side! The platform side! They're make billions of dollars! >> Yeah but so they're making decisions about you know, which content to show and whether they monetize it. And when it's controversial content, they'll turn down the ads a little bit but they won't completely eliminate it sometimes. >> So, Dave, the only thing that the partisans in politics seem to agree on though is that big tech has too much power. You know, What's your take on that? >> Well so I think that if they are breaking the law then they should be moderated. But I don't think the answer is to go hard after Elizabeth Warren. Hard after them and break them up. I think you got to start with okay, because you break these companies up what's going to happen is they're going to be worth more, it's going to be AT&T all over again. >> While you guys were at Sysco Live, we covered this at Amazon Web Service and Public Sector Summit. The real issue in government, Stu, is there's too much tech for bad on the PR side, and there's not enough tech for good. Tech is not bad, tech is good. There's not enough promotion around the apps around there. There's real venture funds being created to promote tech for good. That's going to where the tide will turn. When does the tech industry start doing good stuff, not bad stuff. >> All right we've got to wrap. John, thanks for sitting in. Thank you for watching. Be right back, we're here at Actifio Data Driven 2019. From Boston this is theCUBE, be right back. (upbeat techno music)

>> Announcer: Live from San Francisco, it's theCUBE covering Flink Forward, brought to you by data Artisans. (upbeat music) >> This is George Gilbert. We are at Flink Forward, the user conference for the Apache Flink community, sponsored by data Artisans, the company behind Flink. And we are here with Shuyi Chen from Uber, and Shuyi works on a very important project which is the Calcite Query Optimizer, SQL Query Optimizer, that's used in Apache Flink as well as several other projects. Why don't we start with, Shuyi tell us where Calcite's used and its role. >> Calcite is basically used in the Flink Table and SQL API, as the SQL POSSTR and query optimizer in planner for Flink. >> OK. >> Yeah. >> So now let's go to Uber and talk about the pipeline or pipelines you guys have been building and then how you've been using Flink and Calcite to enable the SQL API and the Table API. What workloads are you putting on that platform, or on that pipeline? >> Yeah, so basically I'm the technical lead of the streaming platform, processing platform in Uber, and so we use Apache Flink as the stream processing engine for Uber. Basically we build two different platforms one is the, called AthenaX, which use Flink SQL. So basically enable user to use SQL to compose the stream processing logic. And we have a UI, and with one click, they can just deploy the stream processing job in production. >> When you say UI, did you build a custom UI to take essentially, turn it a business intelligence tool so you have a visual way of constructing your queries? Is that what you're describing, or? >> Yeah, so it's similar to how you compose your, write a SQL query to query database. We have a UI for you to write your SQL query, with all the syntax highlight and all the hint. To write a SQL query so that, even the data scientists and also non engineers in general can actually use that UI to compose stream processing lock jobs. >> Okay, give us an example of some applications 'cause this sounds like it's a high-level API so it makes it more accessible to a wider audience. So what are some of the things they build? >> So for example, in our Uber Eats team, they use the SQL API to, as the stream processing tool to build their Restaurant Manager Dashboard. Restaurant Manager Dashboard. >> Okay. >> So basically, the data log lives in Kafka, get real-time stream into the Flink job, which it's composed using the SQL API and then that got stored in our lab database, P notes, then when the restaurant owners opens the Restaurant Manager, they will see the dashboard of their real-time earnings and everything. And with the SQL API, they no longer need to write the Flink job, they don't need to use Java or skala code, or do any testing or debugging, It's all SQL, so they, yeah. >> And then what's the SQL coverage, the SQL semantics that are implemented in the current Calcite engine? >> So it's about basic transformation, projection, and window hopping and tumbling window and also drawing, and group eye, and having, and also not to mention about the event time and real time, processing time support. >> And you can shuffle from anywhere, you don't have to have two partitions with the same join key on one node. You can have arbitrary, the data placement can be arbitrary for the partitions? >> Well the SQL is the collective, right? And so once the user compose the logic the underlying panel will actually take care of how the key by and group by, everything. >> Okay, 'cause the reason I ask is many of the early Hadoop based MPP sequel engines had the limitation where you had to co-locate the partitions that you were going to join. >> That's the same thing for Flink. >> Oh. >> But it just the SQL part is just take care of that. >> Okay. >> So you do describe what you do, but underlying get translated into a Flink program that actually will do all the co-location. >> Oh it redoes it for you, okay >> Yeah, yeah. So now they don't even need to learn Flink, they just need to learn the SQL, yeah. >> Now you said there a second platform that Uber is building on top of Flink. >> Yeah, the second platform is the, we call it the Flink as a service platform. So the motivation is, we found that SQL actually cannot satisfy all the advanced need in Uber to build stream processing, due to the reason, like for example, they will need to call up RPC services within their stream processing application or even training the RCP call, so which is hard to express in SQL and also when they are having a complicated DAG, like a workflow, it's very difficult to debug individual stages, so they want the control to actually to use delative Flink data stream APL dataset API to build their stream of batch job. >> Is the dataset API the lowest level one? >> No it's on the same level with the data stream, so it's one for streaming, one for batch. >> Okay, data stream and then the other was table? >> Dataset. >> Oh dataset, data stream, data set. >> Yeah. >> And there's one lower than that right? >> Yeah, there's one lower API but it's usually, most people don't use that API. >> So that's system programmers? >> Yeah, yeah. >> So then tell me, who is using, like what type of programmer uses the data stream or the data set API, and what do they build at Uber? >> So for example, in one of the talk later, there's a marketplace team, marketplace dynamics team, it's actually using the platform to do online model update, machinery model update, using Flink, and so basically they need to take in the model that is trained offline and do a few group by, time and location and then apply the model, and then incrementally update the model. >> And so are they taking a window of updates and then updating the model and then somehow promoting it as the candidate or, >> Yeah, yeah, yeah. Something similar, yeah. >> Okay, that's interesting. And what type of, so are these the data scientists who are using this API? >> Well data scientists are not really, it's not designed for data scientists. >> Oh so they're just going the models off, they're preparing the models offline and then they're being updated in line on the stream processing platform. >> Yes. >> And so it's maybe, data engineers who are essentially updating the features that get fed in and are continually training, or updating the models. >> Basically it's a online model update. So as Kafka event comes in, continue to refine the model. >> Okay, and so as Uber looks out couple years, what sorts of things do you see adding to one of these, either of these pipelines, and do you see a shift away from the batch and request response type workloads towards more continuous processing. >> Yes actually there we do see that trend, actually, before becoming entirely of stream processing platform team in Uber, I was in marketplace as well and at that point we always see there's a shift, like people would love to use stream processing technology to actually replace some of the normal backhand service applications. >> Tell me some examples. >> Yeah, for example... So in our dispatch platform, we have the need to actually shard the workload by, for example, writers, to different hosts to process. For example, compute say ETA or compute some of the time average, and this is before done in back hand services and say use our internal distribution system things to do the sharding. But actually with Flink, this can be just done very easily, right. And so actually there's a shift, those people will also want to adopt stream processing technology and, so long as this is not a request response style application. >> So the key thing, just to make sure I understand it's that Flink can take care of the distributed joins, whereas when it was a data base based workload, DBA had to set up the sharding and now it's sort of more transparent like it's more automated? >> I think, it's... More of the support, so if before people writing backhand services they have to write everything: the state management, the sharding, and everything, they need to-- >> George: Oh it's not even data base based-- >> Yeah, it's not data base, it's real time. >> So they have to do the physical data management, and Flink takes care of that now? >> Yeah, yeah. >> Oh got it, got it. >> For some of the application it's real time so we don't really need to store the data all the time in the database, So it's usually keep in memory and somehow gets snapshot, But we have, for normal backhand service writer they have to do everything. But with Flink it has already built in support for state management and all the sharding, partitioning and the time window, aggregation primitive, and it's all built in and they don't need to worry about re-implement the logic and we architect the system again and again. >> So it's a new platform for real time it gives you a whole lot of services, higher abstraction for real time applications. >> Yeah, yeah. >> Okay. Alright with that, Shuyi we're going to have to call it a day. This was Shuyi Chen from Uber talking about how they're building more and more of their real time platforms on Apache Flink and using a whole bunch of services to complement it. We are at Flink Forward, the user conference of data Artisans for the Apache Flink community, we're in San Francisco, this is the second Flink Forward conference and we'll be back in a couple minutes, thanks. (upbeat music)