 Good morning everyone and as Russ said, I mean this is a milestone being the first user summit We recently have formally joined the the Zen project advisory board. I'm second seat behind my boss Jay Williams Who sits on the board? I am the senior principal product manager for a product called AppLogic who AppLogic Computer Associates acquired a small startup company named 3TERRA and That was the the nexus for Then moving into the cloud arena But wanted to start out and open this talk with some of the motivation behind the title of this thing I'm not your dad's hypervisor not your father's hypervisor. There seems to be you know a huge misconception out there and you know we're all I think we're all working hard to Dispel some of the rumors and the myths about Zen and One of the things I was actually watching a presentation for an open stack it was an architectural overview of grizzly and great overview at the end of it The speaker was taking questions and the second person that came up a bit indignant Walked up grabbed the microphone was like why is Zen on those slides Yeah, isn't Zen dead To which the speaker replied absolutely not it's absolutely still active. It's still a viable hypervisor. It's still open source and Absolutely still important to Not only to the Zen community itself and and cloud in general But for the future of open stack as it provides yet another alternative platform to deploy upon The other alternative title I had for this was the little hypervisor that could You know as you see Russ's t-shirt that says you know Zen Zen's been around for ten years and that's you know a millennium In in this industry And it really did start to have small beginnings and it's come quite a long way Let me quickly start just through this obligatory slide up here I've kind of come full circle. I started out working in the software space being a Primarily, you know technologist hacker tanker and in my own right Linux on that go way back. I remember installing Debian and floppies Back in the day, but product manager by day. I actually do enjoy building products and Infusing them with the features that customers want But I started out network planning and design as an engineer right at a college Really looked at helping, you know a large company do the network planning and design for their network upgrade, but Part of that was looking at network management specifically, you know the software that manages The infrastructure so in the early beginning similar similar to what the cloud is doing today I was really focusing on how software managed the infrastructure of corporate data networks in that sense, but you know obviously Looking at other elements network elements compute elements within the infrastructure And then became a product manager moved to the Silicon Valley became a product manager specifically related three switches so I ended up getting into the hardware business and Actually built ethernet switches from the ground up from chipset design PCB layouts and all that And then ironically ended up back into the software space working in the cloud space with with AppLogic And then simply because this is a Linux Linux Zan show. I actually do manage to code from time to time Some people think that it just because I wear a fancy shirt and I've got a product manager in my title. I Can't do that Anyhow, so let's go back to the beginning. I think you know, I'd rather than go through the entire history I think you know everybody and Again, I don't know, you know what everyone here What your experience has been with Zen or the length of experience you've had with Zen Obviously this began as a research project back in University of Cambridge Ian Pratt The point to be made there about that is that although it started small as a research project within a university environment It was designed very well and it evolved very quickly through in that environment. I've heard some people say it was It was in a clean environment As I guess in contrast to a very public open, you know Hashing about of of commits and whatnot in any open source committee. Obviously it was open sourced and it was Shepherded along through its history and I think that you know Lars gave a great speech Yesterday or the day before about, you know, the lessons learned through that evolutionary process and that history of Zen But again, it was designed tested and improved in a research environment Obviously still many of us many of us in this room still maintain that it's the best design hypervisor that's available today And the proof of that is really that it continues to To provide the ability to scale and an adaptive new hardware obviously the architecture of the hypervisor itself The dumb you don't dump their dumb you structure and the ability to scale with as and primarily the commentary Was around the number of cores as the number of cores increase the core count on a given piece of hardware Having the ability to adapt to that to that new hardware And we'll see this in the future as well as We see, you know, does that arm get deployed on on arm processors Seeing the densities there increase on a given within a given piece of hardware But I'll move quickly past that and start with the beginning of three-terrain app logic Three-terrain was essentially a small cloud Startup that was looking for a fabric Three-terrain was one of the very early entrants into the cloud space actually predating the launch of Amazon And at the time there was no cloud there was no concept of scaling something to cloud levels and Really we had to build our own we looked at a number of different Technologies that were on the market at the time looking at Zen and its capabilities And I simply threw this up as you know, just a rough. I mean this specifically addresses, you know, PV the PV architecture But we're really looking for You know something it was open source and I'll speak to this on the next slide a Solid architecture for you know security was obviously a primary concern at the beginning as we we're looking to design this product And the concerns around security in the cloud space Scalability obviously was another Another concern and looking at the future direction and viability and kind of tying our wagon to a technology specifically a hypervisor What we were looking for Obviously HVM the concept of HVM was has been long-standing and in the space back from the days of of IBM mainframes versus per virtualization and It's adoption in the x86 space with Zen specifically And obviously looking at At the architecture there so specifically on the open source. I think we really wanted to be able to Build our own cloud fabric and we wanted to be able to have some some element or some some amount of control And the ability to patch or optimize And it really have a control of our own destiny in the respect that we were kind of uncharted territory at the time Obviously we primarily just use Zen itself rather than using any of the other tools that came along after that Some of the things that Citrix obviously infused into Into Zen with Zen server and and all those packages around it We we were using Zen native Zen just the hypervisor and instrumenting it ourselves and Because of that we wanted to have control if for whatever reason we needed to integrate in some weird unique way We'd have the freedom to as it turns out. We really didn't need to I think back to look back and Talk to our development team Senior engineers that were around at that time We really only submitted five patches and they were bugs that we found along the way So it was a testament to not only the original research effort at Cambridge, but also the community that was contributing at the time and If any of you recall and Intel and IBM were huge contributors at the time to The hypervisor itself so there was a great deal at the time a great deal of Contribution growth in the hypervisor itself We've obviously found founder. I might probably have to look up the patches to figure out exactly what what specifically we addressed in the back in the day, but But again, that was one of the tenants that we that we were looking for in in a platform Solid architecture obviously security was a concern from day one The way we architected our platform we isolated as much as we could obviously it started at the hypervisor layer Whenever we spoke to customers, that's you know One of their primary concerns obviously, you know moving workloads into the into the cloud that was an obvious concern in discussing the hypervisor architecture and identifying the isolation between guest OSes and and host on a Dom you environments and identifying that there was a a very rigid well strict or structured delineation between Guest OSes and then one of the other things looking at with scalability one of the the The anecdotal things that that came out of that time we were doing dimension testing and I Called the 1024 bug actually I should call it the 1025 bug Essentially, we were pushing the envelope to to understand how many virtual machines we can actually launch at once and We hit a bug that was the 1025th failed So obviously it was a you know power of two bug and it was relatively easy to find but I think the the response back for me to the community or one of the mailing lists was like Why would you ever want to do that? Right why would you ever want to want to launch that many the VMs and I really kind of mark this as the you know the point where we kind of stepped into the cloud era and Stepped into you know the the realm of cloud scale And the fact that you know, this was a relatively easy to bug to fix because it was obviously a dimension I don't know if it's been on a rate boundary or something like something obscure You know we ever thought to provide space for that But this is the sheer fact that someone took a step back and and realize oh we just pushed this thing to a limit and Took a step back for even further and said What could you do with that and that's really kind of the point where we saw the ability to to kind of push the envelope on You know launching virtual machines into Into that realm and really looking at and I we can only imagine that Amazon saw this as well you know and looking at and what I call the cloud scale and Many argue as well if you know We always get into some some people get in this argument of Zen versus VMware Or what about all the other hypervisors and they're depending on how they've you know come to cloud Many still many think that the VMware is is a cloud hypervisor. Many may argue with that All right for its inability to scale at certain to a certain degree But again, we we chose Zen for that for that one of the primary reasons we we saw that Capability in place and then looking at future direction of viability The pair virtualization capability of Zen really saw obviously HVM had had Fair performance but the the ability that to provide pair virtualization capabilities within the guest OS primarily Linux was obviously that the the The beginning of our Of our support with within our platform is providing that performance on Linux hosts We obviously Beyond that provide Windows based pair virtualization drivers for Windows OSes And I think sensor provides their own But again that the ability to even provide that level of Call it punch through or optimization the ability to actually address the hardware Rather than going through an emulation So jumping into the present. This is an important slide really what You know obviously and I don't find the Zen is still the best and and most robust Hypervisor that we have available today and again, this is what I was referring to as cloud scale Looking at Amazon and what they've done with Zen in its history But the other thing that I kind of pulled this up. I pulled this is actually a screenshot of one of our reports This is you know simply looking at our we pull a great deal of analytics out of our customers Great environments 2.4.10 is a very early release of our code. We actually Had a customer in Asia that was one of our first customers and this grid of five servers is still running You know 1474 days later, so and that's one of our you know long oldest running grids out there We had some other older ones that actually recently been shut down But it's a testament to the stability of the hypervisor and the environment that we've you know been able to leverage The strength of hypervisor at our core as a product I mean we we do a great deal of work on top of that and and really we You know we promote that as our value add but without this and and again You know showing this to the our customers that say you know We have a solid foundation and that sound and a foundation is found in Zen and Then the other thing that we did recently And this is again not getting into the bits and bytes, but moving from a Zen Linux to pvops For those not familiar historically Zen wasn't part of the Linux kernel It was essentially a patch to the primary curl There was a move from from the old way of doing things to the new way of doing things which is pvops capability that's native to the Linux kernel introduced in Zen 4.x and Recently the Linux 3.x kernels One of the steps that we took Was obviously to move to that we saw benefit in a number of different. We primarily were running Linux 2 6 18 and I forget the exact version of Zen we were running It may have been three doubt something anyhow In any case we took a step forward because we wanted to gain that functionality all those patches that came into Zen as well as the Kernel base and launched this and so this I think two releases ago for our product And it was a little bumpy in the beginning because it was obviously brand new Zen and new kernel As you can imagine we tested it tested it extensively and we still ran into little mixing crannies it but again we've and We've been able to go from you know an old architecture an old layout for Zen and and and the Linux kernel and jump into a The future direction of of Zen and and maintain the stability. I think we've had some some great stability on this On these versions as well, so quickly looking into the future Continue on innovation. I think that's one of the things that Why we're continuing to drive and leverage Zen in our product I threw up improvements with the stub domains and better security performance and scalability Simply because I think there's something that we haven't leveraged enough in our own product and looking at the mini os concept as far as being able to launch and And Similar to what mirage OS is doing The ability to control and better To move Driver IOs or the putting things into a driver stub or putting I owe into a dedicated domain Is something that we want to leverage more in the future the reason I put more Mirage OS up here We spoke about it this just this past week As they're emerging as you know the ability to to deploy smaller VMs with that without the heavy weight of Of a full OS and really pushing if you kind of imagine the dimensions of Zen as a hypervisor and its ability to do different different things Zen itself is being stretched in different ways And again, we we're looking at that in that ability To leverage that in our own product Obviously, we've leveraged a lot of the other capabilities I mean and again Zen Zen has a wealth of other capabilities as well that I probably even I haven't even really touched upon that Though we that we leveraged some some things live migration the things that you typically see from from a from a hypervisor Has evolved over time that it's you know, it's a full blunt a full fledged Comparable hypervisor to anything else out there in the marketplace And that's why we continue to use it and we'll use it in the future It's such upon the Zen arm. It's very interesting for for us. We the Zen arm aspect of the Zen hypervisor Is very interesting to me personally simply because I think that looking at and I meant I Reference the moonshot servers coming from HP and the idea of a higher density Lower cost lower heat dissipation lower power consumption chip As the future of cloud simply because looking at some of the deployments that we're seeing out as far as Platform as the service and really eliminating as much of the weight of a virtual machine from the OS level and being able to deploy the smallest footprint of a task and Being able to boil those cloud tasks if you want down I call them workers And the ability to get to that micro level and the ability to deploy across a smaller compute element and What we're what I've seen in the HP is taking a very high density approach to Packing as many of those compute elements into a rack unit With the lower power consumption with the lower heat That I think is the and the ability to provide that amount of density. I Really see that as the future of cloud Personally and I think Zen is right at the forefront of that and it's phenomenal that that the Zen is bringing all of its capabilities or As many as many of its capabilities as possible on arm to that space simply because I think that it is an emerging Technology, I think will be leveraged very well In the cloud space The other one that that I didn't even think of and that Lars pointed out to me was in its use in the automotive industry and you know the ability to virtualize on top of an an arm system on a chip for the automotive integration industry Is something that I think will Prove a rather interesting present as well But again all of this all of this Is a testament to its its design its architecture and as we sit here today the infusion back by the community into into into Zen Will only improve the process. I think that And again if you get a chance to to grab I keep referring to Lars's presentation because it was an excellent one Seeing the history is important as well seeing you know the bumpy road as we've as we've had in the past and moving into the future with The effort of the Zen project Board as well as the community We're really cement its future. I Think that's what my next slide here one of the other reasons why I think that works. We continue to be Committed to Zen and will continue to be committed to Zen in the future is The work that the Zen project Group is doing The motivated advisory board. I think we all have a very vested interest in seeing Zen Continue its successes The fact that we're having both user and developers on it I mean the fact that you know most of the projects focus on developers and the fact that we're here sitting here today and Focusing on the user specifically is a very important thing as well. We believe And again, you know, obviously the Zen project will continue to foster and support innovation in the space Touch upon the open stack We're looking at open stack obviously as everybody is To understand what its impact will be in the cloud industry as it emerges as a as a I think many are saying it's going to be a de facto cloud Fabric orchestration and all some people call it a the grand unified Standard of the grand unifying standard It's good to see Zen involved in that. It's very good to see Obviously it's been involved with it, but it seems like more and more You know from from presentations to speeches to Videos up on YouTube of people getting Zen server up and running As the underlying hypervisor layer underneath open stack Simply because we have obviously vested customer space as well as engineering effort with Zen seeing that Getting adopted in the open stack space is as important for us as we look at You know our future plans as a product and interfacing and integrating with open stack So we'll be able to bring all of our Our capabilities and our be able to onboard our workloads and our virtual machines and our applications onto an open stack fabric That's running with that Zen server org Citrix did a I think is a phenomenal effort with releasing Zen server Into the open source formally officially Through Zen server.org. I think it's going to foster a great deal of adoption and integrations with you know that with It's you know with this release that support through Zen server.org. I think that More than ever before I think it was it was available But I think the tool sets and everything all the other pieces and elements around Zen server itself that haven't been released prior to to Zen server or having those available. I think that I've already seen some This it's it's interesting Zen server and an open stack mirror using a Zen server under open stack And Seth which is a distributed object store. It was actually posted to the Zen the Zen project web web page that someone very quickly was able to You know in a completely turnkey fashion very quickly get an infrastructure up and running with Zen server and open stack And the fact that that's you know, that's even available and readily turnkey like Getting from zero to a fully blown open stack environment Is phenomenal and I think that there was there was already Contributions for better installation of Zen server and you know, so that's I think that will Continue to prove invaluable for the community As well as we look at you know, some of the capabilities of Zen server as well From our own product perspective And then I'll jump here to the end here ran through here really quickly And I really kind of gave a top top level overview of why we we still strongly believe that Zen is a viable hypervisor will absolutely continue to contribute to to the Zen project as well as You know doing everything we can to to contribute back into that as well as I use it in the future Obviously look at open stack as well But if there are any questions