Now On Demand- The Impact of Edge on Key Verticals- CET 2021
The edge will not replace cloud, but it does offer immeasurable benefits for verticals, including fintech, gaming, security, healthcare, and video. This executive discussion will cut across these verticals and how edge cloud is permeating our lives in ways that we cannot yet fully appreciate, but will soon see the impact of.
Executive Speakers:
Rupesh Chokshi – VP, AT&T Cybersecurity
Ashok Ganesan – Chief Product Officer, StackPath
Ryan Peterson – SVP, Strategy & Growth, Reveleer
Rob Bartlett – CEO & Founder, WTFast
Bruno Domingues – Principal Solutions Architect - Financial Services, Intel
Transcription Details
Time: 35:15
SPEAKERS
Rupesh Chokshi, Bruno Domingues, Rob Bartlett, Host, Ashok Ganesan, Ryan Peterson
Host:
The edge will not replace the cloud but it does offer and measurable benefits for verticals including FinTech, gaming security, health care, and video. This executive discussion will cut across these verticals and how edge cloud is permeating our lives in ways that we cannot yet fully appreciate but will soon see the impact of. Joining this Executive Session is Rupesh Chokshi, he's Vice President of Cybersecurity at AT&T. We also have Ashok Ganesan, Chief Product Officer at Stackpath. We also have Ryan Peterson, he's Senior Vice President of Strategy and Growth at Reveleer. And we have Rob Bartlett, he is CEO and founder at WTFast. And last we have Bruno Domingues, he's Principal Solutions Architect, Financial Services, Intel.
Session speakers welcome to the program.
Ashok Ganesan:
Thank you.
Rob Bartlett:
Thank you.
Host:
Well, thanks for being here, before we get into this impact of edge on these key verticals, Ashok, if you could sort of set the stage for us. So, industry has been talking about edge compute for quite some time, so what's the next iteration of this discussion is part of the evolution of edge now deploying VMs and containers at the edge? Or what are the developments should we be expecting soon?
Ashok Ganesan:
So, if you look at the industry, there are some major trends that are kind of coming together and I think people are going to leverage these trends to drive how they develop applications. So, the first one I want to talk about is the customer experience. With the newer apps, like peloton or any of the interactive health fitness type application or pick any industry; every industry is trying to look at how can they be more interactive with the customers and how can they basically grab their attention, keep their attention? What do you need to do to basically give a much better user experience? What that basically means is you need low latency. You need the ability to offer services close to where the consumers are. And so, that certainly is one of the things that we're seeing that is going to drive a lot of enterprise modernization, digitization, app distribution, app modernization, so that's a key trend that we are seeing. The second major trend we are seeing is application [unintelligible: 02:20] are evolving from what used to be a monolithic application to client-server, to [unintelligible: 02:25] now it is all around microservices-based architecture, service mesh. So, it's more easily available, these technologies are more easily available at the place where they need it. And again, if you look at the compute trends, the storage trends, the networking trends, they're all increasing in performance for this, for lower power they can offer the same number of [unintelligible: 02:45]. So, that again is going to give a huge performance boost to applications. So, if you look at all this, and then the cloud consumption model is kind of becoming more common. Now, if you look at all this, what I see will happen is, as we get more compute storage and network to application developers, they're going to write a new wave of applications that's going to take advantage of what's available at the edge, what's available in the center and then give a more uniform experience to the consumer. The goal is basically to give the best experience possible to the consumer, leveraging whatever resources they can grab at wherever they needed. So, I think that's the next major wave that we're going to see and I think edge cloud is going to play a big role in that.
Host:
Rupesh, of course, one of the key verticals that edge will have an impact on is the security sector, as edge support billions of IoT devices that can be compromised, how is encryption patching and the use of artificial intelligence to monitor for detecting and respond to potential threats, now really being introduced?
Rupesh Chokshi:
So, again, Thanks for having me, and fantastic group of folks here. I want to add a little bit to what Ashok mentioned, and then we'll double-click on the security space. So, you Ashok talked about the need and the drivers and we're also seeing 5G, the promise of 5G is all about the low latency, the high-speed networks, etc. So, all of these different things coming together are creating this way for the edge platforms to become very real, and you have two use cases going on. You have a B2B2C, and then you have B2C right. And what I mean by that is more and more applications, and use cases and examples that are going to reside in these platforms that are highly distributed. So, the moment you go to a very highly distributed environment, going to the question that you asked about billions of endpoints, 40 billion-plus depending upon which analyst report you look at. So, now you have to look at what is my threat vectors, what is my surface, what is my perimeter? And security becomes front and center. The amount of data that you are processing and the speed at which you are processing, going back to the low latency, high-speed networks, speed cuts both ways. You have to have an AI engine that will deliver on how do you process all of this information from a security perspective. What endpoints do I allow? What is the identity? What is the management infrastructure that I need to keep all of this information flowing? Because in many cases, you're keeping that information localized, so the data is localized through the app, you process it. And in many cases, you're sending it to a private-public cloud. So, you run the security apparatus looking on both sides to say, what is happening in the hospital campus? What's happening in that financial services, edge cloud, what is happening in that particular Metro? And then simultaneously, you're looking at what are all the streams of data that are coming in and out? And in the security space, we always think about data in motion, data in use, data interest.
Host:
Bruno, I want to go to you what the disruption in the financial services sector, are there significant changes in the way we use data and how our financial institutions leveraging internal and also external data to really enrich and offer new services for your customers?
Bruno Domingues:
Yeah, thanks for the question. Indeed, in financial services industry, the use of the data in the edge, I think that it's disrupting the whole financial solutions, as a whole for example. If you're looking for one is aspect, for example, investment banks, investment banks, we are rely for example, now with alternative data and alternative data bringing a lot more data points that you can process, that give a better forecast what you are, your actual position when looking for the risk. And any other industry, for example, insurance, it's completely changed the rules of the market. Because insurance companies they use rely on proxies to measure the risk. For example, if you look at for [unintelligible: 07:18] insurances, they [unintelligible: 07:09] rely on age, gender, your zip code, your employment, all those elements as a proxy for the risk. And nowadays, using, for example, devices that you can plug it, there is some companies already that you plug in your car, they can measure it how you effectively drive the car. And then, of course, change the way that you can make business, instead of I pay for my proxy’s data. Now I can look in for how I usually drive my car that it's of course, it's really tight with the risk. And then you can look for the same for healthcare, you can look at for the same in [unintelligible: 07:54] industry as well.
Host:
Ryan, I want to move to the healthcare sector again, the impact of edge on key verticals, as AI is at the forefront of edge compute. At a macro level, how can AI help healthcare professionals?
Ryan Peterson:
I appreciate you having me here today. And also, it's nice and kind of refreshing to hear from some other industries in the spectrum beyond just the health care that I normally speak with. I had said earlier when we had been speaking that healthcare is often 15 years behind adopting technology. We are, after all, an industry that still performs tremendous amounts of work using paper medical records and fax machines. But jokes aside, AI is certainly a strong potential to transform the way that healthcare professionals work across the healthcare ecosystem. For example, health care providers and specialists are increasingly using AI to assist with faster, more accurate, and less expensive diagnosis of a number of conditions, including very costly ones like cancers. Moving up a notch, health systems and hospitals are using AI to provide more personal care while someone is in the hospital and once they're released from the hospital, AI is being used to help reduce that patient readmission rate. And finally, health plans at that top tier are using AI and administrative capacities and among others, but I think two examples would be smart chatbots and virtual assistants that help members navigate the health benefits, and their network, and those kinds of items. And even AI augments kind of talking about or aligning with what Bruno was mentioning, fraud, waste, and abuse, and cracking down on false claims and other things that help try to reduce overall costs, not just through those organizations but also to the member level and allow it to run more efficiently.
Host:
And Rob, of course, we can't talk about this subject the impact of edge on key verticals without talking about gaming. So, why are gaming companies partnering with telecoms to deliver cloud gaming services, and how our interconnected data centers the key to the future really of the success of cloud gaming?
Rob Bartlett:
Thanks for having me on this speaking opportunity. If we're talking about cloud gaming with regards to game streaming services, such as Google stadia, as an example. So, the nature of the services, essentially, is adding lag to the game experience because you're having to both render and deliver video from the cloud. So, the key about this is that because you're already introducing latency that has to be close to the user if the data has to travel too far, you're going to add additional latency in the cloud gaming experience isn't going to be fun. If you're talking about traditional games, like massively multiplayer, online games, like League of Legends, Final Fantasy, World of Warcraft, where you have players from all different regions, all competing together in a single spot, these games, you have to have a fast and smooth connection. If you have a single gamer who has a bad connection, they're gonna have a bad experience, but they're also going to negatively impact everyone that you're interacting with in-game as well. And beyond that, beyond making sure that the connection is fast and smooth, you also want to make things as level as possible. So, if you have some gamers with an unfair advantage over another gamer, that's going to create issues, so you want to have the game as fast, as smooth as possible, and everyone as close as possible with regards to their connection, and then everyone will have a good experience.
Host:
So, Ashok back to you. And again, you can talk about the impact of edge without talking about CDN and video streaming. We talked about that earlier today on a different session. So, when it comes to edge-based CDN and streaming video, how can enterprises protect their origin uptime, for example, anycast, routing, caching, and shielding your origin server from traffic spikes?
Ashok Ganesan:
Again, that's a great question. So, when people think about the internet, there is no single point of failure in the internet. Everything works seamlessly, smoothly and there's a lot of collaborating nodes that make it work, right. We want to turn CDN into that. So, there are a lot of tools available to bring resiliency to the CDN architecture. The goal is basically to minimize the blast radius and not have a single point of failure anywhere in the delivery system. That's pretty key, and the tools that we have in our toolkit are things like anycast routing, so the goal is basically if something fails, we can always serve from another place so the impact of the failure is minimized. And we also brought newer technologies to market like [unintelligible; 12:34] caching or things like your origin shield. The goal is, again, we can bring a lot of AI-driven ops so that we can kind of understand what's the best place to server content from and how much load that particular place is incurring. And we can use a lot of variables and parameters to decide how do we basically serve a content and where do we serve the content from? So, I think edge is going to play a pretty key role, and edge content services, caching all that is going to be pretty key role. And we do believe that there is still a lot of room for innovation, and we're going to see a lot more innovation coming in the space.
Host:
Good segue. So, she talked about single points of failure, can you please provide an example of how the challenges of a single points of failure under the traditional centralized cloud computing architecture really still exists? And a current example of how edge will alleviate some of those challenges?
Rupesh Chokshi:
So, it sounds like the example is, I’m watching some streaming videos while I'm playing games that are high speed and I'm sitting in the hospital complex that Ryan talked about, in healthcare while collecting my bitcoins, and what Bruno mentioned. And suddenly, something happens, and now I cannot watch that video or play that game or collect my bitcoins, its life-changing, it's very impactful. So, on a serious note, there's two dimensions to it. So, one is, even though we get closer to the edges and where the applications, the latency, the work is happening, the data flow is happening, that architecture has to be a lot more kind of like redundant, kind of for multiple ways. And Ashok briefly touched upon, that if you think about the networks of today, they kind of have lots of different ways to self-heal and continue the traffic. Now, when you think about security and that going back to that example of, if you have a combination of the data flows in the edge and in your cloud, and what you want to do is bring forward the SASI concepts of security and the framework that allows you to look at it at a user level, at a stream level, at a payload level, and making sure that those things are put together. And then what you do is you remove some of that processing in a way that you're not kind of like, you know... single point of failures. Yes, hardware is going to fail, software is going to fail but we need to find faster ways to reconnect and ensure that the same security controls and policies that are applied when we reconnect when we put all of that back together. So, I'm back watching my video, playing my game, collecting my Bitcoins.
Host:
Interesting, Bruno, what's the importance of identifying on-premise cloud off-premise cloud to ensure that data workloads are being really designated in the right place and that to deliver analytics and AI at scale?
Bruno Domingues:
Yeah, thanks for the question. Indeed, it's a very good question, especially when you're talking about regulated industry like financial services industry. And then, unfortunately, the answer, it's a multi-dimensional. The first thing you have to answer is, [unintelligible: 15:56] is regulated? If it's regulated, the regulator allow you to host that workload outside your premises. For example, if you look in Germany, they have data sovereignty, you must reside in Germany. But not only in Germany, but it's in a German company so there is a lot of nuances of the data. The second one is about data protection itself. Sometimes the regulator don't care about, for example, the protection of the data or the algorithm that you execute on but you are care about, for example, the IP. We haven't seen many, for example, hedge funds or investment banks, they decided to run everything on-premise, even have some benefits, economic benefit to running off-premises, but they care about the IP, the protection of the algorithm they have because we know that it's the secret sauce. And there are some elements as well as economic as well, what we call as data gravity if its data is generated internally, if you start to process outside, maybe you can incur a cost that it's more expensive than the process internally. We have seen some customer that they said, well when you start to move our data to the cloud, you realize after 500 servers, that the clouds become more expensive that to run internally. On the other side, of course, cloud can provide some real differentiation for financial institutions that you're not allowed to create by yourself. For example, imagine chatbots, robot advisor, that you have the speech to text, and imagine a bank in Europe, they have to deal with 16 different languages, you have to develop the capability or you can just use a public cloud with this capability and pipeline in speech to text to NLP. You can use the cloud or you can use internally and then you have your back end system, core bank system that provided the service. So, that it's a lot of elements, regulators, data protection, economic of scale, and as well the capability that you can take advantage of.
Host:
So, Ryan, continuing with the topic of AI, how does the conversion of let's say patient files into electronic health records really coincide with the advancements in artificial intelligence software?
Ryan Peterson:
Well, it shouldn't come as much of a surprise that the increasing prevalence of electronic medical records is coinciding with the advancements of AI, and machine learning, and healthcare. The ability to analyze clinical data, and to combine it with demographic data at the individual level, and at the kind of cohort level as you aggregate the data is providing promising inroads in our understanding of a number of illnesses and especially chronic conditions so that we can set our sights on the reduction and ultimately, the prevention. AI is also helping us to better capture and visualize the interdependence of social determinants of health in an individual's life to better utilize that information while providing treatment and a care plan. Without going too much into it, I mean, social determinants of health is socioeconomic status, geophysical status. So, do people have access to healthy food? Do they have access to good education? Do they live by industry or things that are prevalent and relevant to causing health issues and how you address those? I find the figure a little misleading. The CDC said that 86% of office-based physicians are using an EMR. Now, despite that high adoption rate, the largest impediment frankly, that's hitting our space, kind of jealous of some of you folks of advancing AI, really is around clean interoperability of the data across our ecosystem. Our ecosystem utilizes dozens of software and file formats, many that have been homegrown or kind of orphaned, and then vendors and some of those proprietary pieces. And so, yes, there's a lot of API's and yes, there's a lot of normalization and business intelligence that happens. But that's costly and slow and wonky. So, the sooner that we clear that hurdle in our space, the more valuable this data becomes to help people live their best lives.
Host:
And Rob, again, moving to the gaming space, how does the combination of cloud and edge, again, create a more flexible platform that provides gaming developers and publishers with the ability to scale?
Rob Bartlett:
So, I think right now is a great time to be a gaming company, there's all these cloud available. Years ago, if you wanted your game to be successful, you have to set up your own data center and you had to take a gamble, okay, well, where are my gamers going to be? So, the cloud allows you a lot more agility. So, you can set up maybe a cloud in the US, one in Europe, one in Southeast Asia, and then let it go live and see where your gamers are. And then all of a sudden, you see a bunch of gamers in Turkey so you might decide to set up a server in Germany. So, that's what that code is about, it's about agility and about being able to test things out and adjust accordingly. So, in order to be able to reach critical mass for your game, you actually have to have enough players to play that game, so you can't have too many locations so you might have one location in the US to serve the gamers in the US. So, the challenge with that is that in order to reach critical mass, some of those players might be traveling further to get to the game server. So, if your servers in Chicago, for example, you have gamers from L.A, you have to travel pretty far to actually get to the game server. So, the nice thing about edge is that it helps to fill the gap. So, you set up the cloud locations, and then you use edge to optimize the middle mile from end to end, from the user to the game server to ensure a fast and smooth connection so that way, everyone who's playing your game can have a great connection together.
Host:
So, Ashok back to you again, and walk us through this. So, can you offer an edge compute solution use case designed to automate the provisioning of container and virtual machine instances so they can really quickly maximize the efficiency of workloads and their servers?
Ashok Ganesan:
Yeah, absolutely. Again, there are a lot of good examples that the other speakers covered, and again, I'll pick the same example that I picked earlier, smart fitness. As more people sign up to smart fitness so that they can take care of their health. The companies that offer smart fitness services need a new wave of videoconferencing services that has better video latency, better audio latency. They may have to tie in with your things like your variables; how do you connect all these data together? What's the best place to connect all that data together? How can I basically make sure that people are exercising the right way? You can now tie it into the thing that Ryan spoke about, which is how do I tie that into patient records? How do I tie that into the person's doctor so that if there is an event, that person is notified? The applications are endless in terms of what people can do with compute. So, what I believe will happen is the cloud consumption model, the whole great thing about that is, pooling of resources and allocation of resources on demand. And we're going to do the same thing at the edge with VMs and containers. And not just VMs and containers, it's going to be through every resource that we offer a DHS, and my belief is every industry is going to benefit from that. Again, you can pick other industries, like connected cars, or autonomous vehicles, or even map application. If I'm driving in San Jose Palo Alto area, do I really need to send my data all the way over to Atlanta for decision? No, I don't need to. So, there's going to be a lot of innovation that's going to happen in this industry, because of access to compute. Again, if you think about it, most of our innovation is because people had access to compute and they had a problem, and they just found the best way to solve it. So, I think this whole pooling of resources, allocation of resources, that cloud consumption model, particularly at the edge is going to drive a whole new wave of applications and use cases. It's an exciting period for all of us.
Host:
Rupesh, I had several questions left, that I'm starting to think that I'm going to combine them into one or I'm going to go around the room. If you had to sort of forecast what the subject of discussion should be around the impact of edge on the security sector, let's say 18 months from now, what do you think that discussion would entail?
Rupesh Chokshi:
Well, I think, going back to this kind of construct of users, and devices, and network, and application, and data, and thinking about security and the new threat vectors that we see every day, I would say it's really about how do you do a zero-trust environment, a zero-trust framework. As many of the speakers mentioned, you have lots of different applications, lots of different use cases are connecting into this infrastructure, and data is being moved. How do you validate it? So, just kind of building upon what Ashok said, that if there is a user utilizing something to do with a variable, and smart fitness, and do you share that information with the doctor's office or you don't? You actually have to build a very kind of succinct, which is not just related to that user and their data but also the ecosystem. So, [unintelligible: 25:03] federated zero trust is going to be very important in order for all of us to be successful so that's where I would focus. The cybersecurity space, its never going end, it's never going end. It's a hot topic today and it's gonna be a hot topic 18 months from now because we are a society that thrives on innovation, innovation is what we thrive on. So, that creates opportunity for good and creates platforms for the bad. So, we have to be on the good side of it.
Host:
Absolutely. And, I would imagine, all the verticals that are being discussed today are sort of that bottomless pit of discussion, healthcare is no less. And I'll go to Ryan on this. So, again, 18 months from now, Ryan, if we had to talk about the impact of edge on the healthcare sector, what would that discussion until?
Ryan Peterson:
I think there are a couple things, there's a huge blowback right now to the large tech companies, and that's where your apple and your Google, Facebook, etc., are bringing those wearables. And as my panel colleagues here have talked about, wearables and the Internet of Things is where we're going to really make inroads of having the patient become part of the healthcare process. They're just kind of recipients of care but if we can have the proactivity through the wearables, I mean, you folks know far better than me, the sensors coming down the line, but the 02 sensors, the EKG, the accelerator, the gyroscopes, there's a lot that you can do with that right now. But I think a lot of it has to do with the comfort level of people opting into being able to utilize even more of it. And I think that with that recent blowback and or kind of growing blowback against big data in these large tech firms, something's going to have to give on a public sentiment in order to allow people to be more comfortable without using these devices, so that prevalence can grow. But again, that is our goal with healthcare of trying to get folks to also be a partner in their own care. And I think I've been hearing that across some of the other folks too, so very exciting.
Host:
And Rob gaming, and edge and its impact, let's say in the next 12 to 18 months?
Rob Bartlett:
So, the cloud just keeps getting bigger, edge keeps getting bigger, we're working with a lot of different network partners, and they keep adding more locations. The more locations I see them, the more excited I get, 5G is coming around the corner, I like the idea of 5G network slicing and being able to send the data around how it's supposed to be sent using these new technologies, especially with mobile. Mobile gaming is booming. You got a lot of gamers who are using mobile devices, especially in some remote regions, like India and Southeast Asia are good examples. So, we're really excited about what's happening and we'd like to see continued adoption of mobile and gaming in general. E-Sports is amazing as well, so back before COVID, you'd have eSports competitions happening in an arena, now, they actually have to compete online. So, you got to figure out well, how do we create a level playing field for those gamers and for the esports competitions using the networking and edges that's available. There's a lot of really cool ways to use this tech, and it's just gonna keep getting better and more exciting, I think.
Host:
And Bruno, edge on the financial industry, if we were sitting here 18 months from now, what do you think that discussion would be about?
Bruno Domingues:
Well, there are many factors but one that I would like to highlight, it's about confidential computing. Confidential computing, basically, it's a way that you can process the data and protect it in use. So, usually, we can protect on rest, you can protect in transit and use it something new. And a use a case that you're seeing happening in the financial services industry, ft's how for example, do French financial solutions, they can collaborate in [unintelligible: 29:39] or in machine learning. I cannot give you data but I can process the data, give some insights, or some model. It's good, for example, for antimony laundry, that you have to collaborate but you cannot give your data. And the same for example, that it's happening now especially the last year, especially in the US, it's what you call synthetic identity fraud. It was identity that was produced, it was not a real person that of course, gest in the financial sector, gets in financial institutions and they start makes transactions, but it's not real person. And then how to detect this? That's something that the financial institutions cannot do by themselves, they needed to collaborate with telco companies and other institutions, so you can correlate the data. And you cannot share data so you have to rely on some sort of federated learning.
Host:
And Ashoka, CDN video streaming, again, we have this discussion let's say this time next year, what do you think that topic should be about?
Ashok Ganesan:
So again, I think a lot of great points. As Rob said, latency is going to be pretty key for next-generation gaming, next-generation video streaming, live streaming. And the whole idea behind having everything brought to the center to distribute, I think that's kind of going to give room for other innovation to happen where people should be able to stream from wherever the action is happening. So, again, the point that we talked about 5G and what 5G enables. Again, it's creating a lot of new technologies that people can leverage to give better services to gamers, to people who are consuming video, live stream video. So, it's going to change the whole nature of the game and again, a lot of tools are available already and you will see that's continuing to improve. Things like large file downloads, particularly when you have a lot of gamers and when you want to do an image upgrade or a file upgrade. So, things that they can leverage, like accountability network for its security, to the point that like [unintelligible: 31:52] was making with confidential computing. How do we make sure that only the people who have access to certain things are accessing it, and how do we make sure that people don't steal other people's data? So, I think there's gonna be more tools in the toolkit and the main thing that I'm expecting is basically, the new obligations and the new gaming services, the new video streaming services will take advantage of everything we offer to them in terms of its delivery, its content delivery, and its security, and tie it with edge compute so that they can customize the experience for the end-user.
Host:
That's interesting. It's very interesting today to have these very siloed verticals all on one session at one time, but also have a lot of commonalities as far as the impact of edge on those on those verticals. So, again, really interesting to just start this conversation, I'm sure we'll have more conversations like this, as edge continues to sort of permeate the industry if you will. I wanted to say thank you to all of you for your time. I know everyone's busy in their respective areas. I want to say a special thank you to Ashok over at Stackpath for making today's session possible. So, Ashok, thank you to you and your team for participating but also supporting this session today.
Ashok Ganesan:
Thank you. Thank you very much.
Host:
Thank you, Ashok. And again, to our viewers out there, thank you to the speakers on the impact of edge on key verticals for this session on-demand, April 26. Please log on to thenetworkmediagroup.com. So long.