 All right, I wanted to thank all of you for waiting to eat lunch and join us today. I am Jonathan King, I lead cloud strategy for Ericsson, and I am joined by esteemed panel today. We'll talk for about 30 to 40 minutes and then adjourn. Maybe I can just have each of the panelists introduce themselves, starting with their Hesh there at the end. Hi there, Radesh Balakrishnan, general manager for OpenStack at Red Hat. I'm Doss Kamhout, senior principal engineer at Intel, working on software-defined infrastructure. I'm Roz Roseboro, senior analyst at Heavy Reading, covering the Telco data center. And Susan James heading up the product line for NFE infrastructure at Ericsson. So this is an open panel to talk about current market trends, challenges and opportunities. And we had an announcement last week that we, Red Hat and Ericsson, formed a partnership. And I'm curious where Hesh, start with you, get your perspective on the announcement and how it fits in with trends that you're seeing in the marketplace today. So the core thinking behind the partnership centers around the fact that in the Telco carrier space, there's a burning platform moment. There is a need to make sure that NFE or network function virtualization as a focus area needs to be made real and into production with truly upstream contribution. So the partnership centers around three areas within that context. One is making sure that there is a 100% open source based NFEI solution that's built on Red Hat OpenStack platform, made available by Ericsson. So this is an extension of the Ericsson strategy in this place to become a multi-wim solution. I'm sure Susan can add more context on that one. The second pillar of that is around SDN solution, given the fact that both Ericsson and Red Hat are partnering around Open Daylight as a way to drive networking innovation. Ericsson are bringing to the market a solution that's certified around Red Hat OpenStack platform. And last but not the least, the HDS 8000, which is a hyperscale solution which is available in the market, will certify Red Hat Enterprise Linux as well as Red Hat OpenStack platform. So all in all, an end to end comprehensive alignment of portfolio across Red Hat and Ericsson to make sure that we are fueling the communication revolution that's going on. And maybe Susan, building on that, you can similarly add. Yes, I think for us it's very much about providing operators choice. You know, we don't want to prescribe to them how they should take their journey, and we want to be able to support them in that journey. And as Ridesh said, you know, that's why we have said that we want to take a multi-wim strategy. I think one of the things we see happening in the operators is that, you know, there will be pockets of different parts of the operator taking the journey to cloud and to NFV. And they will make different choices. And we want to be able to then provide a partner that can harmonize those choices in the longer term. And of course, we also see that the opportunity of partnering with Red Hat beyond just the NFVI space. As Ridesh said, you know, we're doing certification on our hardware platforms. And I think also it's about being able to support onboarding of other applications, not just Ericsson applications. So particularly seeing, you know, Realm has such a great market penetration when it comes to applications, not just in the NFV space, but in the anti-space, that it becomes critical for us to be a good infrastructure provider. We need to have support for Realm going forward. So I think, you know, it's about being open and giving choices. And does Intel sits in a unique position where you partner with a lot of people and you're looking at the roadmap of the industry out there when it's coming. So maybe your perspective on this partnership and what you're seeing in the industry. Yeah, well, I mean, fundamentally NFV and everything is happening in the calm space is where we've been supporting for quite some time ever since the start of even Etsy. Red Hat, we've been partners since, you know, Linux, right? So it's been a little while. And we've also been, you know, massive supporters together of OpenStack and into the Kubernetes space. So naturally from a software perspective, focusing on a market that's growing heavily towards Intel architecture is pretty interesting for us. And obviously with Ericsson, you guys are one of the first adopters, or the first adopter of our rack-scale design. So being able to bring, you know, all these parties together is pretty awesome to, you know, go faster. We're always interested in things that, you know, make the industry move. Excellent. Ross, how do you see this announcement? I mean, do you think of this as big news? Well, I think it's big primarily because it's bringing two industry leaders together. Obviously Red Hat has a history of being, you know, being a leader in open source and cloud and IT technologies. And Ericsson clearly has the history of supporting the telcos. So I think, you know, combining those two assets together, I think makes for a really powerful combination. Awesome. Well, I mean, Susan, building on this, you know, how does this fit within Ericsson's existing strategy? Do you see this as a change of Ericsson's strategy and evolution? I mean, I see it very much as an evolution of our strategy. I think, you know, when we started this journey some four, five years ago, we saw the market playing out in a different way than it actually has. So, of course, as we take, you know, steps further, we hold our strategy. And I think, as I said before, operators need to have solutions from rail-based workloads. And we also see, you know, these different islands coming forward. So we do see that most operators will have more than one of them in infrastructure. And we want to shift our focus then on how to then manage this multi-vim environment and how to make, you know, open stack easy to use and consume in a multi-vim context. And that comes from both a management and orchestration perspective and how you onboard VNFs and different applications. But also from a hardware infrastructure, and I think that that fits really very nicely with our HDS 8000 product, where we have the capability of creating virtual pods where we can then allocate dynamically the amount of resources required for running those different VMC fill-on. Maybe, Red Hatch, if you look at the landscape beyond NFE, we look at SDN, we look at, you know, beyond their business models that are coming, you know, how does Red Hat see the operator space beyond NFE? And what are things that you're thinking about in that space? So, if you pull back and look at what's happening in the carrier landscape, clearly on the IT side there's more maturing of adoption of virtualization that's been ongoing for a while. And then now with NFE, introducing the same sense of technologies with different workload focus, I think we are marching towards a unified fabric that's going to be there for both the IT side and the network side and the carrier from a carrier perspective, right? While the journeys might have started at different points in time. Given that our focus has been making sure that we're not creating a forked solution when it comes to the carrier side, but how make open stack carrier grade rather than have a carrier grade open stack, for example, right? So, I think that fits in with the philosophy or the approach that Susan outlined as well, which is the market dynamics have changed. There was at one point in time a felt need to have certain capabilities which are required for telco specifically, but now the community has come around to realize and innovate to make sure that those capabilities are in open stack itself, right? So, that's kind of the long way of saying that we believe 100% upstream first innovation is going to be the model that's going to prove itself out. And we're pretty excited about being in the position to drive that in partnership with Edison. That's great. And, you know, Roz, you've studied this space. You've done recent surveys, I know, on adoption rates and timing. And, you know, what would you see as... We've been talking about it a few for a number of years now. What would you see as the stake today? Well, funny that you mentioned the research. I just completed the survey for Red Hat. And we found for the first time we asked this question quite a bit as you can imagine. And this was the first time where over half of the operators that we surveyed said they were currently executing on their NFV strategy, which I think is quite noteworthy. And similarly, we found that nobody said they didn't need an NFV strategy. And that has happened before. So I think that the question of should we do NFV, that's pretty well decided. I think where we're moving to is now how do we make this all happen? And so I think that's, you know, from an industry standpoint, the more people we have kind of getting their hands dirty, they're starting to expose some of the issues. And of course, the sooner we can expose them, the sooner we can get them fixed. So, like I said, I was quite encouraged to see that. And of course, you know, people that are trying to take advantage of this trend. That's also very exciting. That's great. And does I know you said earlier, you know, have been partnering with Red Hat and Erickson for a long time and will they see a big opportunity with carriers around NFV? Maybe just some more perspective from Intel on what you're seeing in the market in that space, in Intel's perspective. Yeah, so fortunately, you know, it all is moving forward. So that's fundamentally what we want to see. And also, as Redesh noted, we're really interested in having one code base that can work across multiple types of workloads. And if you fundamentally actually look at NFV versus, you know, some other, even a big data analytics workload, they're all processes that are doing computing of some sort with different characteristics. But they can all actually run on the same type of fabric. If I go back a couple years, I think I sat in room with Redesh and we went over rolling upgrades, all these key elements that enterprise needed. And actually, if you look at comms in telcos, it's the same things. So I think, you know, all this work that's been going in for a few years now. And by the way, I was building OpenStack Clouds in Diablo Day. So, you know, a lot of all that work, I think, is really starting to pay off. And so now that the adoption is happening, I think enough of a groundwork has been there that it's not, you know, you need some special OpenStack. It's, you know, you have something that's actually good enough. It's working and it's just going to get better as we get more laser focused on, you know, NFV workloads plus additional workloads that have, you know, very similar characteristics. Thanks for dating yourself. Yeah, absolutely. I was going to add one thing. You know, you're mentioning, Redesh was talking about carrier greatness and part of that same survey, we asked people, you know, what was important as they're looking at an NFVI platform and for the first time ever, something besides security was number one and it was reliability. But what's interesting was that we also asked them how confident are you in the existing NFVI platforms to provide all those characteristics. And I'm sorry to say they're not particularly confident that the existing platforms can do it. And so that's another reason I think it's important. I think it's interesting that Erickson and Ryan are coming together because the telcos want to make sure that the people developing platforms to them really understand what it means to run a network. And I think Erickson, you know, clearly has that kind of heritage so I think that's going to be important. I think that's going to make some, a lot of them feel much better because, you know, you've walked in their shoes and you know what their issues are. Yeah, absolutely. And I think that sets us up for, you know, a good transition, you know, looking at, you know, we have a sense on the state of the industry and why this partnership has come together. And I told the panelists beforehand I was going to make a terrible pun but I couldn't contain myself. We have to talk about containers. And that seems to be like that is the next frontier. And we had, you know, Mirantis in the meeting before next door making some news with directions it's going and taking. And I'm curious to get just an open discussion across the panel. I don't know, maybe Rahesh, if you want to go first because we're looking to do more in the community space around containers. You know, we both are highly interested and see it relevant but I think there's interesting things for the panel to share and just an open perspective on where do we think the intersection between NFE and containers is and where is it headed? Containers? Never heard of them. No, I'm just kidding. So from Red Hat perspective, we are all in on containers. So if you look at Docker as the Docker format, next to Docker the company, Red Hat is the largest contributor. If you look at Kubernetes next to Google, Red Hat is the largest contributor. And we also have a design from ground up fully functional container platform in the form of OpenShift Container Platform that customers such as Prettybond earlier today, they were up on stage talking about their big data implementation. Prettybond is one of the largest deployers of our containers on top of OpenStack as well. So we believe that being in the cockpit, so to say, of driving innovation across OpenStack, KVM, Linux, Kubernetes and Docker gives us an amazing opportunity to present choice in terms of you can run physical, you can run virtual, you can run containers. Our job is to make sure that the experience is seamless. So that's the opportunity that's uniquely in front of Red Hat, and that's our product strategies evolving as well. Now when it comes to probably the carrier space, we see a little bit of maybe a lag in terms of readiness to embrace containers per se, more to do with the applications or the NFEs themselves not being ready to get there. That's a fundamental re-architecting decision, you know, lots of manners of work ahead and our opportunity ahead, if you will. But at the same time from an infrastructure and platform perspective, we are working to make sure that when that ISV momentum does happen, the intersection of OpenStack and containers is ready to operationalize and have a life cycle, supported life cycle around a product. That's great, Susan. I've got to give you the mic. Yes, I've got something to say. I think two weeks ago we were in the SDN OpenFlow World Congress in The Hague, and I think we had a really good demo there that we showed where we had a rabbit-in queue server running in OpenStack and then spinning up the client in containers. And I think, you know, there will be more and more of these sorts of examples where you'll actually want to use a combination of both OpenStack and containers. And the great thing about that is, of course, when you're using a container, the startup time of the container is much, much shorter than it is for OpenStack. So, you know, being able to provide networking via SDN and automating that network becomes really crucially, I would say, going forward. And how we then use containers to be able to complement, I would say, existing applications or existing VNS certainly become something that we see going forward. As Ridesh said, you know, if you look at the bulk of the applications in the NFV space, then very few of them have any containers included at all. But certainly we see that as something that's coming forward. And then when we're building in the infrastructure around both OpenStack and working with SDN, we're making sure that we can manage workloads, again, from a native perspective, a virtualized, and also a container-based perspective. Because applications will tend to dictate what they want to use rather than operators making a choice or vendors making a choice on this is how we should go forward. So, for us to be able to support a myriad of infrastructure becomes really key going forward. Very good. And, Daz, we've talked about containers beyond in other meetings and other settings. It's been something that Intel's been actively involved in, helping, at least stand up, CNCF, contribute early on. So, I know you have a lot to say on this topic. I'm curious to get your perspective, maybe not just on the BNF side, but the industry adoption rate of containers. Sure, yeah, totally. So, I mean, containers have hopefully all known that I've been around for a really long time, heavily in Linux in 2006 and Docker made it easy. So, basically, we've been sharing at least a vision that we'd like to see all data centers, the core of them being containers, which are basically performance isolated processes. And in order to, and the benefit of that is high levels of resource utilization, performance optimizations that you can't do natively in a traditional hypervisor-based environment. If you look at something like Google's Borg, it's built with this concept. And if you have an application that actually has the give me entire server construct, you run it in a VM. But that hypervisor runs in a container. So, it's just, it's a constant container world that it's nice to see. We actually thought it would take longer for the industry from a technology perspective to get to where we are right now, but it's super encouraging to see the advances in Kubernetes, how much movement's going on to it. We love red hats, all over it. And it's allowing us to do some things that haven't been done before. Now, on the application side, though, obviously there's hundreds of different types of apps. NFE is an interesting space because when they took their initial appliances and moved them to NFE, they actually didn't really re-architect them very much. So, the reason that they want this high reliability is they didn't take cloud native concepts, which is resilience, how to design your app in a fashion that allows for failure. Now, when they see the ability to re-architect containers, it brings a whole new interesting space because when you see what the Kubernetes API can do, it's pretty powerful when you think about how you run software. Resilience is built into the constructs of the container orchestrator. So, we're happy to see the pay sets moving. I think there's gonna be a widespread of applications, but I think with what both Red Hat and Ericsson are delivering together, you have that full breadth of bare metal. Give me a VM, give me a container orchestrated, and then I'll go up to you with OpenShift and run across clouds. It's a nice space we're in right now. Yeah, absolutely. And, you know, Roz, you've done a study on NFE, maybe a study in the future on containers. What are your thoughts? We actually did run a survey earlier this year and we asked the question to the telcos about containers. And essentially, I don't think any of them said they'd actually deployed containers. And when I spoke with them, it was more an issue of we're aware of containers. We know they're gonna be a big deal. We're not sure when, we're not sure why. So we're kind of keeping up with it because we know we have to be prepared for it eventually. I think they recognize that it really is gonna be the VNF owners having to do that rearchitecture. And they will be more the consumers of the containers as opposed to creating them. So we just have to know how to manage them. But it's certainly very top of mind for everybody. Susan's tapping my elbow if I'm going to say it. I know I was actually gonna turn it around and ask you the question. Uh... Whoa. It was like the rent that would do this, so, uh... She's called my pun. So I think a couple things to say, I guess, and then re-put my hat back on on the panel piece. I think containers is this next frontier and actually a lot of what does it set in terms of just the fundamental computer science benefits that come with it. But also mindful of the application challenges that you talk about. I think this is an inevitable future, but how we get there has to be really well thought out and staged. And I think that comes with, from an Erickson standpoint, we feel like we've already made investments and have started this journey in certain parts of our business. What we're very interested to do is have partnerships that can help inform where we go when we go. And it's really one of the things that has drawn us to Intel and to Red Hat is that we can partner across multiple areas of our portfolio and then we can better inform and make decisions on where we take investments. And we view that as part of our job for our customers is that, yes, we want them to work with Erickson and they want to work with Erickson, but part of what they expect when they work with Erickson is that they're gonna get this perspective on where things are and where they're going and increasingly in the interconnected world we are in. Part of that means that we have to form relationships like this so that we can start to work together on the lab work, the coding, the development and actually my next topic, which is how do we upstream this in the right way? We have so many different open source projects out there and sometimes there's even contention amongst some of them. So, and then there's also the risk downstream of things happening in Forks and how do we work upstream? So, hopefully that answers your question Susan. But I think we've talked about the Alliance, we've talked about NFE, we've talked about containers, maybe talking more broadly about just open source organizations and upstream activity and you're very active in the open NFE, open daylight and you've mentioned it's Red Hat so you're very active and Intel's active. Maybe you shift the topic to open source and what is Erickson doing and driving right now and what are some of the priority communities that we're focused on? So, at least from my perspective because Erickson as you know is driving lots of different things. So, from my perspective of course I have a high focus on open NFE, obviously open stack, open daylight, open V switch. We're also getting more heavily involved in FDIO. We also have then the open compute platform and ODCCP. So, I haven't been so focused on the harder ones being more of a software girl myself but really we have very much embraced the open source community as a development model. I would say that over the last few years we have very much learned what that means in transforming to be an open source development model and certainly what we've learned along the way with open stack we have then transferred directly into our open daylight engagements where we do take a upstream only approach. So, you know, very much similar to Red Hat. I would say we're in the process of getting there from an open stack perspective. As I said, you know, we've been doing this for a long time and open stack was definitely not ready from a characteristics perspective when we started this journey. So, we still have a legacy of capabilities that we've added to open stack and we're driving the upstream of that content through open NFE as well as directly into open stack with partners working with both Intel and Red Hat to really make sure that that capabilities are upstream because I want to move away from doing any work post-distribution. That has also been our experience around OVS where we took OVS and did improvements. We're now moving away from that model so that we just work directly in the upstream. So, I think we've been doing some very good work with Intel around OVS in the last month or so where really made significant improvements in performance around OVS. So, that's the direction that we will go. You know, we're then using open OPNV as a way of how to glue these different initiatives together and really make some of those choices. It becomes quite confusing. And I think, you know, the whole mano space right now is a good example of that where there is a proliferation of projects ongoing and we need to see that really come together. I don't think it's in anyone's interest to see, you know, too many projects going on. And I think that you will see some more harmonization going in across the networking space as well going forward. And excellent, and active time for this. Excellent, and does Intel has made substantial investments in open source, in carrier space and beyond? I mean, it's a core strategic imperative. It looks like maybe you can share your perspective on Intel open source. Yeah, so pretty much all those projects we're heavily involved with. Even one of my peers, Adrian Hoven, covers a lot of the mono work from the technical perspective. Our goal constantly is drive as much upstream as possible, as quickly as possible because it floats all boats, right? So how to make sure whether it's a cloud service provider that chooses to do everything themselves, whether it's somebody that's working closely with Red Hat, we want to make sure that everybody that is using this technology can move as quickly as possible. We're actually in a weird state at Intel where sometimes the software actually isn't keeping up with the hardware advances, which is, you know, not something that we think is beneficial for the world. We see some of the top players able to utilize those hardware benefits much faster than the rest of the world. And if we, our goal is democratizing how computing happens in the software space and really if only a couple can benefit from this tech because they have deep, deep Linux experience and have their own patches, it's not, we don't see it as necessarily fair for the rest of the world that's trying to push the envelope too. So as fast as we can go upstream as many core projects as possible is pretty beneficial, especially now that we have new tech coming like 3D Crosspoint, which is gonna be a game changer for how people deal with memory and storage. And then what's happening with Silicon Photonics. We want to see core projects move as quickly as possible, get organized around something. Right now there's probably too many container orchestrators out there. We'd like to see maybe people latch on to one, maybe two, but really drive those forward as quickly as possible. There's probably too much strife in the container runtime or formats today. So that's why we care about things like OCI and driving some standards there. So the more that we can lock into a piece of code base, drive it up shame quickly and keep up with hardware, that'd be a fantastic situation. Yeah, we talk so much about software's eating the world, but with 3D Crosspoint, with all the technologies coming at the hardware level, it really is a point where software does need to catch up. And you have to go to an advanced level because everyone knows it's Red Hat, it's open source. So, how do you get the benefits of this upstream first mentality but still maintain the velocity of development to try and catch up to what Daz is talking about? Right, too. First of all, plus one to whatever Susan said and Daz said, right? From a philosophy as well as engagement perspective. To your question about pace of innovation versus sort of the stability, that's what we kind of pride ourselves in being able to parse on a day-to-day basis. For us, upstream communities all are about how do we move the needle on the pace of technology innovation? But at the same time, we quickly turn around and then say, we know intimately what does it take to run in production a workload or an infrastructure and to live with it for a number of years. And if you're from Japan, that years might be in two digits if you will, right? So, that is the product part of it. So, by making sure that we're innovating fast and in an upstream aligned fashion, we can make sure that the pitfall that Daz was talking about can be awarded. In other words, no snowflakes. Everybody gets the benefits of democratized innovation that we are driving at a very fast pace. At the same time, from a product perspective, we make sure that we spend an ordinary amount of energy driving quality engineering and having a support life cycle to make sure that customers can't live with that, right? So, that's the duality that we kind of pride ourselves in being able to do that for the last two decades and we'll continue to do that. That's great. Well, we've had a far ranging number of topics. We're approaching our end here. I wanted to see if there were any burning questions from the audience. I know there's probably hunger pains because we're in the lunch hour. And I have a very important question for Roz if there aren't any questions for the audience. I don't see any questions. We'll be up here afterwards and sort of joining you all in the movement towards lunch. My question for Roz is, like, what hat am I wearing? For those who don't know, our Cubs are going to the World Series starting tonight. Yeah, so this is like actually one of the first times I think I've ever worn a baseball hat at an event. I mean, I've done t-shirts all the time, but so this is the Chicago Cubs baseball hat. So it's the United States thing, baseball, the World Series. And the Cubs have not been to the World Series in 70 years. My mom pointed out that she was three months old the last time they went. Mine too. And they haven't won the World Series in 108 years, which is by some reckoning the longest stretch of any sport in modern history. So hence, I'm missing the opening game tonight to be here with you, so this hat keeps me comfort. So with that, we are a wrap. I wanna thank you for joining us in the lunch hour and please, if you have questions or anything, come up at the end and give our panel a round of applause. Thank you. Thank you.