>> Hey welcome back everybody, Jeff Frick here with theCUBE we're at the data privacy day at Twitter's world headquarters in downtown San Fransciso and we're really excited to get into it with our next guest Dr. Andreas Weigend, he is now at the Social Data Lab, used to be at Amazon, recently published author. Welcome. >> Good to be here, morning. >> Absolutely, so give us a little about what is Social Data Lab for people who aren't that familiar with it and what are you doing over at Berkeley? >> Alright, so let's start with what is social data? Social data is a data people create and share whether they know it or not and what that means is Twitter is explicit but also a geo location or maybe even just having photos about you. I was in Russia all day during the election day in the United States with Putin, and I have to say that people now share on Facebook what the KGB wouldn't have gotten out of them under torture. >> So did you ever see the Saturday Night Live sketch where they had a congressional hearing and the guy the CIA guy says, Facebook is the most successful project that we've ever launched, people tell us where they are who they're with and what they're going to do, share pictures, location, it's a pretty interesting sketch. >> Only be taught by Black Mirror, some of these episodes are absolutely amazing. >> People can't even watch is it what I have not seen I have to see but they're like that's just too crazy. Too real, too close to home. >> Yeah, so what was the question? >> So let's talk about your new book. >> Oh that was social data. >> Yeah social data >> Yeah, and so I call it actually social data revolution. Because if you think back, 10, 20 years ago we absolutely we doesn't mean just you and me, it means a billion people. They think about who they are, differently from 20 years ago, think Facebook as you mentioned. How we buy things, we buy things based on social data we buy things based on what other people say. Not on what some marketing department says. And even you know, the way we think about information I mean could you do a day without Google? >> No >> No. >> Could you go an hour without Google? >> An hour, yes, when I sleep. But some people actually they Google in their sleep. >> Well and they have their health tracker turned on while they sleep to tell them if they slept well. >> I actually find this super interesting. How dependent I am to know in the morning when I wake up before I can push a smiley face or the okay face or the frowny face, to first see how did I sleep? And if the cycles were nice up and down, then it must have been a good night. >> So it's interesting because the concept from all of these kind of biometric feedback loops is if you have the data, you can change your behavior based on the data, but on the other hand there is so much data and do we really change our behaivor based on the data? >> I think the question is a different one. The question is alright, we have all this data but how can we make sure that this data is used for us, not against us. Within a few hundred meters of here there's a company where employees were asked to wear a fit bit or tracking devices which retain more generally. And then one morning one employee came in after you know not having had an exactly solid night of sleep shall we say and his boss said I'm sorry but I just looked at your fit bit you know this is an important meeting, we can't have you at that meeting. Sorry about that. >> True story? >> Yeah >> Now that's interesting. So I think the fit bit angle is interesting when that is a requirement to have company issued health insurance and they see you've been sitting on your couch too much. Now how does that then run into the HIPPA regulations. >> You know, they have dog walkers here. I'm not sure where you live in San Francisco. But in the area many people have dogs. And I know that a couple of my neighbors they give when the dog walker comes to take the dog, they also give their phone to the dog walker so now it looks like they are taking regular walks and they're waiting for the discount from health insurance. >> Yeah, it's interesting. Works great for the person that does walk or gives their phone to the dog walker. But what about the person that doesn't, what about the person that doesn't stop at stop signs. What happens in a world on business models based on aggregated risk pooling when you can segment the individual? >> That is a very very very biased question. It's a question of fairness. So if we know everything about everybody what would it mean to be fair? As you said, insurance is built on pooling risk and that means by nature that there are things that we don't know about people. So maybe, we should propose lbotomy data lobotomy. So people actually have some part chopped off out of the data chopped off. So now we can pool again. >> Interesting >> Of course not, the answer is that we as society should come up with ways of coming up with objective functions, how do we weigh the person you know taking a walk and then it's easy to agree on the function then get the data and rank whatever insurance premium whatever you're talking about here rank that accordingly. So I really think it's a really important concept which actually goes back to my time at Amazon. Where we came up with fitness functions as we call it. And it takes a lot of work to have probably spent 50 hours on that with me going through groups and groups and groups figuring out, what do we want the fitness function to be like? You have to have the buy in of the groups you know it they just think you know that is some random management thing imposed on us, it's not going to happen. But if they understand that's the output they're managing for, then not bad. >> So I want to follow up on the Amazon piece because we're big fans of Jeff Hamilton and Jeff Bezzos who we go to AWS and it's interesting excuse me, James Hamilton when he talks about the resources that EWS can bring to bear around privacy and security and networking and all this massive infrastructure being built in terms of being able to protect privacy once you're in the quote un-quote public cloud versus people trying to execute that at the individual company level and you know RSA is in a couple of weeks the amount of crazy scary stuff that is coming in for people that want interviews around some of this crazy security stuff. When you look at kind of public cloud versus private cloud and privacy you know supported by a big heavy infrastructure like what EWS has versus a Joe Blow company you know trying to implement them themselves, how do you see that challenge. I mean I don't know how the person can compete with having the resourses again the aggregated resources pool that James Hamilton has to bring to barrel this problem. >> So I think we really need to distinguish two things. Which is security versus privacy. So for security there's no question in my mind that Joe Blow, with this little PC has not a chance against our Chinese or Russian friends. Is no question for me that Amazon or Google have way better security teams than anybody else can afford. Because it is really their bread and butter. And if there's a breach on that level then I think it is terrible for them. Just think about the Sony breach on a much smaller scale. That's a very different point from the point of privacy. And from the point about companies deliberately giving the data about you for targeting purposes for instance. And targeting purposes to other companies So I think for the cloud there I trust, I trust Google, I trust Amazon that they are doing hopefully a better job than the Russian hackers. I am more interested in the discussion on the value of data. Over the privacy discussion after all this is the world privacy day and there the question is what do people understand as the trade off they have, what they give in order to get something. People have talked about Google having this impossible irresistible value proposition that for all of those little data you get for instance I took Google Maps to get here, of course Google needs to know where I am to tell me to turn left at the intersection. And of course Google has to know where I want to be going. And Google knows that a bunch of other people are going there today, and you probably figure out that something interesting is happening here. >> Right >> And so those are the interesting questions from me. What do we do with data? What is the value of data? >> But A I don't really think people understand the amount of data that they're giving over and B I really don't think that they understand I mean now maybe they're starting to understand the value because of the value of companies like Google and Facebook that have the data. But do you see a shifting in A the awareness, and I think it's even worse with younger kids who just have lived on their mobile phones since the day they were conscious practically these days. Or will there be a value to >> Or will they even mobile before they were born? Children now come pre-loaded, because the parents take pictures of their children before they are born >> That's true. And you're right and the sonogram et cetera. But and then how has mobile changed this whole conversation because when I was on Facebook on my PC at home very different set of information than when it's connected to all the sensors in my mobile phone when Facebook is on my mobile phone really changes where I am how fast I'm moving, who I'm in proximity to it completely changed the privacy game. >> Yes so geo location and the ACLU here in Northern California chapter has a very good quote on that. "Geo location is really extremely powerful variable" Now what was the question? >> How has this whole privacy thing changed now with the proliferation of the mobile, and the other thing I would say, when you have kids that grew up with mobile and sharing on the young ones don't use Facebook anymore, Instagram, Snap Chat just kind of the notion of sharing and privacy relative to folks that you know wouldn't even give their credit card over the telephone not that long ago, much less type it into a keyboard, um do they really know the value do they really understand the value do they really get the implications when that's the world in which they've lived in. Most of them, you know they're just starting to enter the work force and haven't really felt the implications of that. >> So for me the value of data is how much the data impacts a decision. So for the side of the individual, if I have data about the restaurant, and that makes me decide whether to go there or to not go there. That is having an impact on my decision thus the data is valuable. For a company a decision whether to show me this offer or that offer that is how data is valued from the company. So that kind of should be quantified The value of the picture of my dog when I was a child. That is you know so valuable, I'm not talking about this. I'm very sort of rational here in terms of value of data as the impact is has on decisions. >> Do you see companies giving back more of that value to the providers of that data? Instead of you know just simple access to useful applications but obviously the value exceeds the value of the application they're giving you. >> So you use the term giving back and before you talked about kids giving up data. So I don't think that it is quite the right metaphor. So I know that metaphor come from the physical world. That sometimes has been data is in your oil and that indeed is a good metaphor when it comes to it needs to be refined to have value. But there are other elements where data is very different from oil and that is that I don't really give up data when I share and the company doesn't really give something back to me but it is much interesting exchange like a refinery that I put things in and now I get something not necessarily back I typically get something which is very different from what I gave because it has been combined with the data of a billion other people. And that is where the value lies, that my data gets combined with other peoples data in some cases it's impossible to actually take it out it's like a drop of ink, a drop in the ocean and it spreads out and you cannot say, oh I want my ink back. No, it's too late for that. But it's now spread out and that is a metaphor I think I have for data. So people say, you know I want to be in control of my data. I often think they don't have deep enough thought of what they mean by that. I want to change the conversation of people saying You what can I get by giving you the data? How can you help me make better decisions? How can I be empowered by the data which you are grabbing or which you are listening to that I produce. That is a conversation which I want to ask here at the Privacy Day. >> And that's happening with like Google Maps obviously you're exchanging the information, you're walking down the street, you're headed here they're telling you that there's a Starbucks on the corner if you want to pick up a coffee on the way. So that is already kind of happening right and that's why obviously Google has been so successful. Because they're giving you enough and you're giving them more and you get in this kind of virtuous cycle in terms of the information flow but clearly they're getting a lot more value than you are in terms of their you know based on their market capitalization you know, it's a very valuable thing in the aggregation. So it's almost like a one plus one makes three >> Yes. >> On their side. >> Yes, but it's a one trick pony ultimately. All of the money we make is rats. >> Right, right that's true. But in-- >> It's a good one to point out-- >> But then it begs the question too when we no longer ask but are just delivered that information. >> Yes, I have a friend Gam Dias and he runs a company called First Retail, and he makes the point that there will be no search anymore in a couple of years from now. What are you talking about? I search every day, but is it. Yes. But You know, you will get the things before you even think about it and with Google now a few years ago when other things, I think he is quite right. >> We're starting to see that, right where the cards come to you with a guess as to-- >> And it's not so complicated If let's see you go to the symphony you know, my phone knows that I'm at the symphony even if I turn it off, it know where I turned it off. And it knows when the symphony ends because there are like a thousand other people, so why not get Ubers, Lyfts closer there and amaze people by wow, your car is there already. You know that is always a joke what we have in Germany. In Germany we have a joke that says, Hey go for vacation in Poland your car is there already. But maybe I shouldn't tell those jokes. >> Let's talk about your book. So you've got a new book that came out >> Yeah >> Just recently released, it's called Data for the People. What's in it what should people expect, what motivated you to write the book? >> Well, I'm actually excited yesterday I got my first free copies not from the publisher and not from Amazon. Because they are going by the embargo by which is out next week. But Barnes and Noble-- >> They broke the embargo-- Barnes and Noble. Breaking news >> But three years of work and basically it is about trying to get people to embrace the data they create and to be empowered by the data they create. Lots of stories from companies I've worked with Lots of stories also from China, I have a house in China I spend a month or two months there every year for the last 15 years and the Chinese ecosystem is quite different from the US ecosystem and you of course know that the EU regulations are quite different from the US regulations. So, I wrote on what I think is interesting and I'm looking forward to actually rereading it because they told me I should reread it before I talk about it. >> Because when did you submit it? You probably submitted it-- >> Half a year >> Half a year ago, so yeah. Yeah. So it's available at Barnes and Noble and now Amazon >> It is available. I mean if you order it now, you'll get it by Monday. >> Alright, well Dr. Andreas Weigin thanks for taking a few minutes, we could go forever and ever but I think we've got to let you go back to the rest of the sessions. >> Thank you for having me. >> Alright, pleasure Jeff Frick, you're watching theCUBE see you next time.

(upbeat music) >> Announcer: Live from Barcelona, Spain it's theCUBE covering Cisco Live 2020, brought to you by Cisco and its ecosystem partners. >> Welcome back. This is theCUBE's live coverage of Cisco Live 2020 here in Barcelona, Spain. I'm Stu Miniman. My co-host for this segment is Dave Volante. John Furrier is also in the house. We're doing a little more than three days wall-to-wall coverage. One of the big themes we're talking about this week is in this complicated world, networking, containerization, applications going through transformation. Future work simplification is something that is very important and helping us to really tease through and understand some of the integration, some of the announcements where Cisco is helping to simplify the environment, happy to welcome back to the program one of our Cube alumni, Kaustubh Das who is a Vice President of Product Management at Cisco. KD, thanks so much for joining us. >> Oh, I'm delighted to be here, it's great to be here. >> All right. So but up on the main stage, they walk through a number of the announcement. Listen Tony, I was talking about some of the pieces and two of the announcements from the main stage are under your purview. So why don't we start there, walk us through the news. >> Yeah, so there's two two major announcements. The first one's called Cisco Intersight Workload Optimizer. And what it is, it's a way to have visibility into your data center, all the way from the applications and in fact, the user journeys within those applications, all the way down through the virtualization there, through the app servers, through the container platforms down into the servers, the networks, storage lands. So you have a map of the data center. You have a common data set that the application owner and the infrastructure owner can both look at and you finally have a common vocabulary so that it helps them to troubleshoot faster so on a fast reactor way, they talking the same language not pointing fingers at each other or do things proactively to prevent problems from happening when you see a server running hot, a virtual machine running hot, an application server running hot. You can diagnose it and have that conversation before it happens. >> My understanding is that Intersight and there's also some integrations with AppDynamics there, AppD which of course we know we talk to that team at the Amazon Cloud shows a lot. So that common vocabulary spans between my hybrid and multi cloud environments. Am I getting that right? >> Correct and there's two pieces even within that. So certainly that's integrations with AppD so from AppD we get information about the application performance. We get information about the business metrics associated with the application performance. We get information about the journeys that user take within the application and then we take that data then we stitch it together with infrastructure data to map how many applications are dependent on which application servers, how many VMs are those dependent on, what does those VMs run on? What hosts are they dependent on, what networks do they Traverse, what lands do they run on? And each one of these is an API call into that element in the infrastructure stack. Each API call gives us a little bit of data and then we piece together this data to create this map of the of the entire data center. There's a multi cloud aspect to it obviously and so we also make API calls into AWS and Azure and clouds out there and we get data about utilization of the various instance types. We get data about performance from the cloud as well. >> So two announcements. Insight Workload Optimizer and HyperFlex AppDynamics, is that right or they are separate? >> HyperFlex application platform. >> Okay. >> So if we look at the, let me just put these two in context. Every enterprise is doing two things. It's trying to run application that it already hosts and then it's writing some bespoke new applications. So the first announcement, the Cisco Intersight Workload Optimizer and the integration of the AppD, that helps us be more performant for applications we're running, to have troubleshoot faster, to have reduced cost in a multiply cloud environment. The second announcement Dave, the HyperFlex application platform, it's really targeted towards developers who are writing new applications on a container platform. And for those developers, IT needs to give them a simple appliance like easy to use container as a service platform. So what HX AP HyperFlex application platform is is a container as a service platform driven from the cloud so that the developer gets the same experience that they get when they go to an AWS and and request a pod. But they get it on-prem and it's fully 100% upstream Kubernetes compliant. It's curated by us so it's very simple appliance like feel for development environments on container. >> Okay. So Insight Workload Optimizer, it really attacks the problem of sort of the mystery of what goes on inside VMs and the application team, the infrastructure team, they're not talking to each other. You're bringing a common, like you said parlance together. >> Kaustubh: Correct. >> Really so they can solve problems and that that trickles down to cost optimization as well as performance. >> It does, aha. >> And I understand hyper HyperFlex app platform it's really bringing that cloud experience to on-prem for hybrid environments. >> For our new development. So if you're developing on containers, you're probably using Kubernetes but you're probably using this entire kind of ecosystem of open source tools. >> Yeah. >> And we make that simple. >> Okay. >> We make it simple for developers to use that and variety to provide that to developers. >> Okay. since underneath, there's HyperFlex. is there still virtualization involved in there and how does this tie in with the rest of the Kubernete solutions that we were talking about with your cloud partner? >> Great, great. Great question. So yes, there is HyperFlex underneath this. So to develop, you need a platform. The best platform we think is the elastic platform that is hyper-convergence. And with type of flex, we took storage networking and compute, packaged it together, made it super simple. We're doing the same thing with Kubernetes. So it's the same concept that how do you take complex things, package it together and make it almost appliance like. We said we're doing the same thing with Kubernetes. Now Stu, the point about virtualization is a good one. A lot of container deployments today are run in virtual machines. And they run in virtual machines for good reason, for isolation, for multi-tenancy, for all these kinds of ignition. However, the promise of containers was to sort of get rid of the tax that you pay when you deploy a virtualization environment. And what we're giving out right now is no tax, no virtualization tax virtualization environment. So we have a layer over transition in there. It's designed for this use case so it does give the isolation, it does give the multi-tenancy benefits but you don't need to need to pay additionally for it if you're deploying on containers-- >> Job wise it is some KB and base type solution >> Kaustubh: Correct. >> Underneath, it makes a lot of sense if you look at the large virtualization player out there. It's been talking about how do I enable the infrastructure that's all virtualized and everything and bring them along to that journey >> Correct. >> For that bridge if you will to the environment? Sure containerization sometimes I want to be able to spin it up super fast. It leaves, it dies, but if I'm putting something in my data center, probably the characteristics I'm looking at are a little bit different. >> Correct, correct. The other thing it does and you touched on it a little bit was we have a homogeneous environment with the major clouds out there. So one of the things developers want to do is they want to develop in one place and they want to deploy in another place so develop on Amazon and deploy on-prem or Azure. We've got an environment with very native integrations so that it's natively integrated into EKS and AKS. And we facilitate that develop anywhere, deploy anywhere motion for developers who are trying to build on this. >> So okay. What does the customer have to do to consume these solutions? >> So our customer right now for this one is IT operations. It maybe helps to bit back a little bit on why we did this. I had a lot of customers come to me and they said listen, I'm IT, I'm in the business of taking shrink-wrap software, taking enterprise-grade resilient infrastructure, putting that together. I'm not in the business of getting open source drops, every week, every day, every month, putting them together by making sure all the versions line up and doing that again and again and again. So the putting together an Ikea piece part of open source software has not been traditionally the IT operator's business. So our customer is that IT operator. What they need to do is they buy a, if they may have a HyperFlex system already, or they buy a HyperFlex effect system. They add on a license for the HyperFlex application platform. They have an Intersight license. This is delivered from the cloud so Intersight manages that deployment, manages the lifecycle, manages the upgrades and so forth. If they have a state that spreads across multiple sites, Intersight is cloud-based so it can actually reach all those sites and so they're in business. >> Okay, so very low prerequisite. You just got to have the product and you can add on to it. >> Yeah, I have the HyperFlex system, add on to the license, you're done. >> So I'm curious. How unique do you see this in the marketplace? I think the keynotes this morning is that there's no other company that can actually do this. I wonder if you can sort of add some color to that and just help our viewers understand the uniqueness of Cisco's offer. >> Sure. So I think it's unique on a number of different dimensions. The first dimension is HyperFlex itself. We've had an appliance mentality to this for a long time and we really co-designed the software and the hardware to build the most performance hyper-converged system out there. We took the same approach when we went down the path of Kubernetes and building this container platform. And so it's called design software and infrastructure together. The second thing is we said we're going to be 100% upstream Kubernetes compliant right, so if you look at the major offerings out there in this space, they're often several months actually behind where the open source is, where the upstream of the sources and developers don't want that. They want the latest and greatest, they want they want to be current, right. So we are far ahead of most of the other offerings out there in terms of how close they are to their upstream commodities. The final piece is Intersight. Intersight gives us immense ability to have scale where especially if you're developing on containers and micro services, you're talking tens of thousands, many tens of thousands of N nodes, maybe more. And being in the cloud, we have the scale and we have reached so a lot of our customers have distributed assets and branches and you know, hotel chains with hotels and so forth. Intersight allows us the ability to actually deploy across a distributed asset class with with the centralized kind of provisioning. >> You see a huge uptake right now and containers generally Kubernetes, specifically. It's sort of across the board but I wonder if you could comment on how much of that demand and activity is coming from sort of the traditional IT roles versus with other hoody developers? >> Yeah, that's that's a great question. So yes, there is a on a hype cycle it's at the top of the hype cycle. Everybody's in actual adoption. I think it's pretty good as well right. So that is every company I talk to is doing something in containers, every company. But usually, it starts at the developers. It starts with like you described with the folks in the hoodies and that's great. I mean they're experimenting, they're getting this thing. What hasn't happened is it hasn't gotten mainstream. And things can mainstream is when IT picks it up. It certifies hey this is resilient, this is enterprise-grade, I can stand behind it, I can manage the lifecycle of it. That's what we're enabling here. I'm giving IT a path to mainstream containers, to mainstream Kubernetes so that the adoption kind of takes it from that pipe cycle to mainstream adoption. >> Do you see K.D. new sort of data protection approaches or thinking as containers come into play? I mean they're ephemeral, you know microservices sometimes aren't so micro. Like you say, they're running often times inside a VM. So how are people thinking about protecting containers? >> Yeah, yeah, that's a big topic in itself. I mean one of the things that we found is even though they were supposed to be ephemeral, they require persistent storage so we've implemented within hyperflex a CSI plugin that provides that persistent storage layer to containers. Then once you do that, all of the data protection mechanism of HyperFlex come into play. So within the cluster, the resiliency, the triple replication, the backups, the partnerships we have with their other data protection pairs, all of those mechanisms become available instantly and those are enterprise-grade. Those are ones that IT knows and can stand behind. Those become available to containers right away >> Great. >> But it's great, great question. >> Awesome. >> Just want to go back to when you were talking about Intersight and the reach and the scale of the solution reminds me that Cisco has a strong legacy in global environment. What I'm curious about, we've talked a little bit about Edge computing in the past. >> Kaustubh: Yes. >> Where are you seeing Edge today? Where is that going? What should we be looking at in that space when it comes to Edge? >> Yeah, no, it's a big part of our customer demand. In fact, we haven't seen I think all flash was the other technology that took place so fast but Edge has been really phenomenal in its growth rate. Over the last year, we've seen I think probably up to 15% to 20% of my engagements are in this space on at least the hyper convert side. So we see that as a big growth area. More and more deployments are happening. They're being centrally managed, deployed at the edges and so the only solution that scales to something like that is something that's based on the cloud. But it's not just enough to be based in the cloud. You've got to maintain that entire lifecycle right? You've got to make sure you can do installs, upgrades, you know OS installs, health monitoring and so as we built that Intersight platform, we've added all those capabilities to it over time So we started with hey this is a SAS-based management platform and then we added telemetry and then we said if we can actually match signatures, now machines can manage machines. So a good amount of my support calls are now machines calling each other and then fixing themselves. So that's just path-breaking from an informant Edge environment. You don't have an IT person, add an Edge location. You want to drop, ship an appliance there, and you want to be able to see it remotely. So I think it's a completely new operating model. >> I know we got to go but I want to run your scenario by K.D.'s. Do share with me from one of my breaking analysis. Look Dave, you mentioned Flash, that's what triggered me. (laughing) So think of containers and Kubernetes, think of like Flash. Remember Flash used to be the separate thing which we used to think it was a separate market and now it's just everywhere, it's embedded in everything. >> Kaustubh: Yes. >> So the same thing is going to happen with Kubernetes. It's going to be embedded in solutions. This is exactly what it is. By 2023, we're probably not going to be talking about it as a separate thing, maybe that's sooner. It's really just going to be ubiquitous, yeah. >> No, I totally agree. I think the underpinnings that you need for that future, you need a common infrastructure platform and a common management platform. So you don't want to have a new Silo creator and this has been our philosophy even for hyperconvergence. We said hey, there's going to be converging infrastructure that will be hyper converted. But they need to be the same management system, they need to be the same fabric. And so if it's Silo is not going to work. Same thing for containers you know. It's got to be the same platform in this case, it's HyperFlex. Hyperflex runs virtualization, it runs containers with HXAP. You get all of those benefits that I've talked about. It's all management insights, it's a common management platform across both of those. At some point, these are all tools in somebody's tool kit and you pick the right one for the job. >> Kaustubh, it is wonderful to hear the company that has been dominant in one of the silos for so long of course helping to bring the silos together work across the domains. Congratulations on that good news, always great to have you. >> Yeah, always great to be here, thank you. >> Dave: Thank you. >> For Dave Folante, I'm Stu Miniman back from lunch where we hear more from Cisco live in Barcelona 2020. Thank you for watching theCUBE.