 Good morning or good afternoon, depending on where you are in the world. This is Mark Follier with OpenStack Foundation, and we're excited to have another webinar. And more interestingly, we're excited to have a new release of OpenStack. I'm sure most of you know a new version comes out every six months with the hard work of many, many people. I want to talk to some of those folks today that are joining me on the webinar and talk about what's in the release and what it means to users and also answer any questions that the audience may have as we go along. So feel free anywhere along the way here to pop questions into the BrightTalk webinar system and we'll try to get to as many of those as we can as we go along. So Juno is the tenth release, hard to believe, and just over four years the community has been able to produce ten releases. And to start with, I'm going to talk a little bit about some of the highlights and then we're going to try to get quickly to asking Kyle and Andrew who are joining me on the webinar some questions about it and get their point of view. So Kyle Mestri is a Senior Director and Chief Technologist of Open Source Networking at HP. He's also the Project Technical Lead for the Neutron Project OpenStack Networking. So you can certainly direct your toughest networking questions at Kyle and I'll quickly deflect them to him. And we've also got Andrew Metri who's an architect at Comcast. They're one of the largest OpenStack clouds out there that they are running their business on and we always love to hear from users and not just those of us who are involved in OpenStack every day working on building it but also really hear from those people that are using it and find out what they like and don't like and how they see the new release, how they think about evaluating each new version and what goes into that. So we wanted to start with a quick poll and the question was really how familiar are you are you with OpenStack and there's a voting system that's part of this webinar platform and I have just started the poll so you can actually click inside the webinar app. Wow we're getting tons of votes already. So really just wanted to know if you're just kind of hearing about OpenStack, if you're actively evaluating it or if you're already running it, maybe some other scenarios or other categories but as we go along they'll help us kind of get a sense for who all is in the audience and your familiarity with OpenStack. Now I know that probably a few people are here just learning about it for the first time so at a very high level OpenStack is a cloud platform that helps you manage your compute storage and networking throughout your data center, really helps automate that and includes a dashboard both for administrators who are operating the clouds, you know folks like Andrew at Comcast and the end users, the developers who want to provision those resources also can do so through a self-service dashboard. So at a high level you know that's what OpenStack is all about and you know businesses really use it to move a lot faster to give their developers the resources they need to be creative, to create value, new features, get them to market faster than sort of the pre-cloud era of opening a ticket and waiting weeks to get a new machine online just so you can start hacking on an app idea you have, that can happen in minutes now with all the enterprises adopting cloud and OpenStack is there to help that transition and you know with the 10th release one of the things that continues to be front and center for users is how wide of support the platform has for different technologies. So by that we mean things like drivers, there were almost a hundred drivers that were tested throughout the development process of creating Juno that are included in the Juno release and there are many other drivers available from all of the top tier vendors from compute storage and networking and of course we see people using it in many different industries. Some of the themes that we see Juno really hitting on are enterprise maturity so you know because we have a lot of plug-ins and driver support that means that it works in very diverse environments. Some of the improvements in the object storage around storage policies really help with enterprises that maybe have different classes of data and they want to store it in different regions or at different levels of redundancy depending on how critical it is that they have X number of copies. A lot of improvements in the identity side and make it easier to plug into LDAP and also to federate identity across different environments. You know NFB is something we'll be talking about a little bit later which is a new buzzword if you haven't heard it that comes from the telco world of network functions virtualization and then last but not least the biggest new overall capability that was added in with this integrated release was the data processing capability and this is really for things like Hadoop that many users were already running on OpenStack. They're now able to do that within the OpenStack dashboard and really help create different node pools, different cluster templates and quickly spin up worker nodes and control nodes and help run jobs and do all that from inside of the OpenStack dashboard. So for people that were interested in running things like Hadoop on OpenStack that is now a lot easier thanks to this new Juno capability. One of the things that is really key to how we as a community approach developing OpenStack is that we always want to have as much user involvement as possible. This is not like traditional software that might be developed in a very closed manner by one company and then kind of trotted out on stage at the Moscone Center and sort of force-fed to the people in attendance. This is much more about getting users involved. We've had ops summits. I was really lucky to attend an ops summit in San Antonio back in August where we had users from Comcast, Timer, Cable, GoDaddy, Yahoo! There's a whole list here you can see but these are all people that not only are running OpenStack but they really care about the future of OpenStack and they're helping drive the direction of it. And that's a big part of how we got to Juno and how we do things in OpenStack to make sure the software really meets the needs of the end users. People are always interested in the stats. Over 1,400 people contributed to Juno which was even more than Ice House. I think at some point you kind of lose track of how many features when you get over 300 to kind of too many to list. But we also certainly saw a lot more activity in the bug fixing area and I think that's a testament to how many people are running in production and helping identify different issues that were brought forward and fixed in Juno. And there were also some really cool statistics that I didn't put on the slide but Jeremy Stanley who is one of the members of the OpenStack infrastructure team was letting me know just how massive the testing infrastructure has become. We have over almost 100 third-party testing systems plugged in that help verify all those drivers I was talking about and we actually saw just in the six month time almost 20,000 changes merged and that includes not just OpenStack but a whole host of things that the cloud computing industry cares about that the infrastructure team is kind enough to help manage and facilitate as part of the infrastructure and actually over 18 terabytes of log data was generated through the process of all that testing. So pretty amazing scale for a distributed team that's really not driven by one company. It's really a community-led effort. And so the number of contributors has become so massive that you really can't put them on one slide well though we tried but just want to thank all those contributors that made this possible and I really hope that we're able to, they're all able to come to our next summit in Paris in a couple of weeks. So the summits are a great place for all the developers and users to get together. So with that intro out of the way I wanted to throw some questions out to my guests on the call today. So I guess we'll start with Andrew Mitri from Comcast. What do you see are some of the most significant changes in Juno coming from the Ice House world? So I think we see a few different things that we're really excited about. First is being a large service provider. Having IPv6 support baked into Neutron is something that we were excited about and we were excited to kind of participate and lead the sub-team on that effort and see it come to fruition there. Also some of the new services that are coming up with NFV and also the big data that Sahara, like you mentioned, we've actually had a lot of our internal teams ping us about, oh, when are you releasing Sahara? When can we leverage this? When can we do that? So that's been generating a lot of interest where they're actually, you know, it's probably one of the first times that our internal teams are tracking OpenStack and saying, hey, I need this feature. When can you get it deployed? So we're getting a lot of internal interest there as well. Also, yeah, and lastly is the, you know, just the live upgrades and the better stability and enterprise integration with Keystone and all those services as we mature and scale out the platform. We're excited to see those come. Okay, great. Kyle, I know you've been living and breathing the networking Neutron project. So I'm sure that that's near and dear to your heart, but overall I guess we'd love to hear your perspective on kind of what makes Juno such a big relief for OpenStack and the user base. Yeah, thanks, Mark. I think one of the main features on the Neutron side for Juno that the team has worked really hard on was the distributed virtual router support, which is otherwise known as DVR. And this was important because as a team, the Neutron team really tried to focus in Juno on parity with Nova Network, which is something we've been working on, you know, stability and parity with Nova Network. And we really achieved that during Juno now and DVR was one of the key features for that. And at a high level, what DVR lets you do is it lets you remove the routing functionality to each computer host, and it also moves the DNAT functionality down there as well. So this was a pretty huge feature and we'd really love to see users provide a lot of feedback for this in Juno. The team is going to be working to expand support for this during the TLO development timeframe as well. And one of the other things that I'd like to mention is one of the big changes that we saw was we really tried to involve operators a lot more during the Juno cycle. This was facilitated partially by the ops meetup which you had mentioned before. And specifically, when we were looking at migrating Nova Networking users to Neutron, we as developers and engineers spent a lot of time thinking about that. We came up with some prototypes. We worked through some examples. But it was really key to us to kind of get input from the users and the operators. And so the ops meetup was really key from that perspective. So it was a very useful experience for us on the engineering side. Yeah, that's great. I think I'm glad you brought the ops meetup again because I know that Comcast had quite a few folks involved there. And Andrew, were you able to attend the ops meetup or what was the kind of impact for your team? I was there along with some of my team. And actually, we find the ops meetup beneficial in a couple different ways, not just in interacting with the developers and giving feedback there and kind of having them back and forth there, but also interacting with other operators and kind of learning best practices and how they make things work at a larger scale and what challenges they have, how they solve those challenges. So the community around it is actually really valuable for us. It's definitely one of the events we really look forward to every cycle. And I think it helped us build a relationship more directly with those developers so that, hey, now we can put a name, a face of that name, and reach out over IRC or whatnot, or reply to a thread and understand what the context was there. That's great. And for anybody out there who's not aware, you know, we do, we've started more regularly having these ops meetups for users in between summits, but we're also doing them at the summit. So there will be some breakout sessions dedicated to operator discussions and collaboration at this upcoming summit in Paris. So if you're an operator and you're coming to the Paris Summit in November, you know, be sure to check out those sessions. So let's go to the next slide. And I think I want to make sure I pause and encourage people to put questions in the, you know, in the question box, and we'll try to get to those as we go along. So how has OpenStack evolved within your organization? You know, this probably applies a little more to Andrew, although I'm sure Kyle recently joining HP is probably aware that they run quite a bit of OpenStack as well. But Andrew, do you have any insight you can share on kind of how that's evolved over time? Yeah, sure. So, you know, we started looking at OpenStack back in 2012 and did our first, you know, small kind of proof of concept production deployment at the end of 2012. Now, I think we demoed X1 on OpenStack in the spring of 2013 and was doing production traffic with our set-top boxes there. And the past year since then, almost a year and a half now, we've deployed more than five times the amount of infrastructure. We've kind of become the de facto in-house cloud provider within Comcast and we've changed, we've had kind of a paradigm shift where our products used to, you know, deploy infrastructure just for their product, kind of all in one vertical. And now the default option is let's go deploy on this multi-tenant platform that we've rolled out within Comcast. We can get our products out to market faster. A lot of wins because we can now, we don't have to wait for infrastructure. We can spin up compute and storage on demand. And so we can do things like support events that need infrastructure just for a short amount of time like the Olympics or the World Cup, spin up a lot of infrastructure and then spin it back down and use that for something else. So it's really helped us evolve the cloud paradigm within Comcast. Wow, so the Olympics or the World Cup, those are, I've heard of those. Those are pretty important pieces of infrastructure hanging on in the background trying to deliver that to a lot of wild fans all over the world. So that's pretty awesome that OpenStack can have some part in that. So we did have a question from the audience and I think this would be a good one for Kyle. So this user or audience member asked that a lot of users seem to be using Nova Network, which for those of you who aren't familiar, there's really two ways to configure networking in OpenStack. There's Nova Network, which is the sort of legacy networking capability built into the Nova project. And then there's the neutron, the more forward looking kind of direction. And so the question was, there are a lot of users using Nova Network that aren't planning to start using neutron. Do you see this changing with Juno? What needs to be done to facilitate this transition? This is definitely a frequently asked question. I hear from users and we heard it at the Op Summit. So maybe you can shed some light on that for us, Kyle. Yeah, absolutely. So I think you're right. This is obviously a pretty hot topic. And both teams, the Nova team and the Neutron team spent significant time over the Ice House and Juno timeframes on this. And in fact, we entered the Juno development cycle with a charter from the technical committee to close this parity gap. And the team did a spectacular job in closing that. So the one area where we really didn't close it was on the migration plans. And as I touched on earlier, we really wanted to get some more input from users on that because that was the one thing you were lacking. So I expect during the key low time frame, we'll try to move forward on migration strategy there. You know, my general consensus was that it was hard for us to get a lot of data around who's using Nova Network versus who's using Neutron. And Matt made the migration story a little bit better. So we would certainly love to hear from additional users in this area for people that are using one or the other or that want to migrate from one to the other, that sort of thing as well. So I think that that's one thing that we'd really like to help to facilitate this migration maybe from the new Nova Network to Neutron. If we could hear from more people, it would help us to be able to better plan what the migration story will look like, I think. Yeah, that's a great point, Kyle. And this is a perfect time for me to plug our user survey. So we continuously, the foundation and working with a lot of different people in the community and the user committee have created a user survey where we're constantly collecting insight and input from different users. So if you go to openstack.org slash user survey, you can go in there and it takes about 10 minutes to tell us what you're doing with OpenStack. We get data points like, you know, how many people are using Nova Network versus how many people are using Neutron and what do you think about Neutron or what's the hold up for you to make the migration. So that's actually our single bit biggest kind of tool that we have in the toolbox to get that insight. So please go fill that survey out. We've gotten a tremendous number of responses in the last few weeks because we tend to go over the results at the summits. But, you know, we're always welcome people to fill that out whenever possible. Now, we have another question come in, which is what are the plans for bringing in Docker support on Mainline of OpenStack? And I can speak to that a little bit, I think. You know, I think I just actually caught up with Michael Still yesterday, who's the PTL for Nova. And he would certainly be the best person to provide kind of the most recent details. But the way he explained to me was that, you know, Docker support is very strong at this point through the Stackforge hosted drivers. And so there is some interest, I think, in bringing it back into being part of, you know, shipping as part of the release, you know, potentially in a future release. It's not the case for Juno. But the good news is that to a certain extent, it's kind of semantics as to where the code lives. You know, having the driver shift with the software certainly means it's passed a lot of QA tests. But on the other hand, if it's shipped through Stackforge, meaning it's not actually shipped with the release, that doesn't mean that it hasn't been tested. So some people, I think, kind of conflate those two things. So, you know, in this particular case, you know, the feedback I've gotten from Michael is that there is a lot of activity around Docker. We certainly know there's tons of interest. And there are multiple ways it can be used with OpenStack, both in a sort of pseudo hypervisor type of a role through that as a driver in that sense. As well as on top of OpenStack guest VMs, you may want to use something like Heat, the orchestration framework to orchestrate the deployment of Docker containers on top of OpenStack Cloud. So there are a lot of different layers. You can use Docker with OpenStack. And there's certainly a tremendous amount of interest. I think there'll be a lot of discussions about that at the summit in a couple of weeks. And that was actually why I was catching up with Michael, was to make sure we had everything in place for those discussions to happen. So hopefully that addresses that question. So here's a question that is definitely for Kyle. With the new DVR on Neutron, how North-South traffic is going to be managed on OpenStack? Is there going to be a need of an additional edge router for traffic outgoing or incoming than virtual environment? Yeah, definitely. So I will say that there's actually a great Wiki page that the DVR team put up. I don't know if we can share links here, Mark, or not, but in the background maybe if someone could let me know, because I can share that it has a lot more detailed information. But I mean, in summary, the intent of the DVR is to do that. So with regards to the North-South traffic, I think that the person maybe needs like the added traffic for floating IPs. And so we can do the DNAT on the host, but the SMAT still is required to be done on a network node. And the main reason for that is the SMAT requires burning IP addresses for compute nodes. Now, there is some work going on to figure out how to possibly do this in an efficient way in Kilo, but it didn't make it into Juno. I will also say, though, that we did land the blueprints around making those network nodes highly available. In other words, making them so you essentially have HA with them there. So that actually helps in that particular case for the SMAT nodes for DVR. Okay, great. And just so you know, Kyle, there's a box that says live audience messaging below in the webinar page, and you can drop links in there. So if you've got a Wiki link that would be useful for the audience, you know, definitely put it in there. So I think we had... What's that? Oh, I was just saying thank you. Yeah, perfect. Oh, cool. So next question was, is designate fully integrated into Juno for DNS as a service? So designate is an official project that's in the incubation phase. So it will be in, I assume, a future integrated release, but it's not at that state yet for Juno. So it's available, the software's there, a lot of people are using it, but it's not at that phase yet where it was classified as being part of the integrated release. So hopefully, you know, we'll see something along those lines next year. Next question was, this is a great one. We definitely hear a lot from users, which is about upgrades. We'll talk more about upgrades from Icehouse to Juno. So I could take a stab at that, but I'll maybe throw that one to Kyle and see if you want to take that one. Well, you know, like it might be worth having Andrew talk a bit about that as well from a user perspective, if you're comfortable with that. Sure. I mean, I could talk a little bit. So we've, you know, right now, we're going from Havana to Icehouse just about done with that process. And it's been a lot less painful than previous upgrades. So we're happy that a lot of feedback is making it through the operator to meet up and the feedback we've been given there. We haven't made the step yet from Icehouse to Juno, so I can't speak as specifically to that. But it does look from everything that we've been monitoring that should be an even more seamless transition than going from Havana to Icehouse. So we're excited about doing it. We try to maintain a regular release cadence with the goal of eventually getting to be a month off of an official or stable release. Right now we run about five to six months behind. Yeah, definitely. I mean, one thing I can say on the neutron side with regards to upgrades in Juno, we really did a lot of work to clean up our database migration story, which definitely will impact upgrades going forward from Juno. I mean, previous to Juno, we had an issue where we had plug-in specific migrations which caused different issues around the item potency of where you landed. So we had a team working on that, and they did a great job during the summer, and they cleaned all that up. So now in Juno going forward, we didn't no longer have plug-in specific migrations, so I think that's definitely going to be a win for people using the Trump. Yeah, I'm glad you brought up the database piece of that. I know that in the past, from the Nova side with Compute, there was really from Havana to Icehouse was the first time it was possible to really do an upgrade without impacting end-user workloads in the Compute realm. But there was more painful and took more time than it really needed to because of the database migration updates. And so I think that was one of the things that improved from an upgrade perspective on the Nova side from Icehouse to Juno, which is what you were alluding to in Neutron as well, which is all that work to make the database updates faster, helps in a lot of different realms. So let's see. Another question was, how's the IPv6 integration going on? And I know that this was a big part of the Neutron networking story for Juno. So maybe, Kyle, you want to give an update on that? Sorry, Mark, this is around the IPv6 stuff. I apologize. Yeah, the IPv6 improvements. Yeah, definitely. I mean, the goal in Juno was to essentially close the gap around IPv6 on tenant networks. And like Andrew said earlier, Comcast was heavily involved in this. They had someone leading the sub-team. And this was another example of great work by the team, and we were able to close all the gaps so that we now on Neutron support full IPv6 for tenant networks and everything. And I'm sure Andrew can go into much more detail around the specifics of that as well, given his team kind of led that way. Yeah, so to be clear, we're actually running V6 via some backports today on Havana in production. And all our set-top box traffic and email traffic, all that requires V6 today. And so we do run it today, and we're excited that it is fully supported upstream in Juno. So it worked well for us. Okay, thanks. So there was a question about installation and configuration. You know, there's so many different options out there, Crowbar, DevStack, Packstack, and all these other different options. And, you know, maybe this would be an interesting one. Andrew, for you to comment on, I'm sure you haven't had the time to evaluate every possible installation tool, but, you know, maybe if you can talk about how you guys approach, you know, picking the right installer or talk a little bit about how you guys manage that, that might be useful for some of the people in the audience. Yeah, so we've been running OpenStack for a while. We started off at the time leveraging Puppet because it seemed for us the most mature way of getting OpenStack installed, and that's kind of where we're leveraging today. I think we are very interested in the evolution of OOO, and we're keeping an eye on it, but today we've been kind of focused around just leveraging Puppet and it's working for us there. Okay, thank you. So I'm just kind of going through some of the questions here, and I know we had a few more in the deck. So maybe this is a good time to ask Kyle to talk a little bit about NFV. For those of you who don't know about NFV, it's network functions virtualization. We discovered there were three letters that hadn't been made into an acronym yet, so we felt like that was a good opportunity to do that. So this is really about the telco world and trying to virtualize all these functions that happen at the back end of these massive networks and route all of our cell phone calls and text messages and MMS, and trying to make that more of a software-defined world for those guys and taking advantage of that. And OpenStack plays a role in that and Neutron as well, so Kyle, can you shed some light on what NFV is about and what role OpenStack is going to play in that new market? Absolutely. So I think what you said, NFV is really about virtualizing all of these different network functions and being able to kind of orchestrate them so that you can dynamically spin them up and spin them down. And if it sounds familiar, I think it's a lot of what, you know, the existing server virtualization world was doing. This is applied to kind of network appliances or functionality that was done there. And OpenStack is really a natural place for this because we have the orchestration capabilities which are necessary for this, and we have the capability to manage these images, these network functions as their virtual images. We have the networking capability to string them together, to spin them up dynamically, to place them where they're needed, close to the workloads, or wherever you want to put them. So I think that's why it's kind of a natural place. During the Juno timeframe, there was an NFV sub-team which was sitting kind of a, it wasn't a sub-team of any particular project. It was a broader team that was focused on all of the different pieces that are involved here. And I know during Juno they spent a lot of time on Nova Blueprints doing things that were important for NFV on the Nova side. And, you know, we're already engaged with them on the Neutron front because I suspect that during the Keyboard timeframe, we're going to really focus on what Neutron needs for NFV and, you know, to enable that from the Neutron side. Okay, thanks, Kyle. So, Kyle, there was a detailed number of questions in the webinar from one of the audience member about some specifics in the Neutron world. And hopefully you can pull down the menu because I don't know if I could read all the details without, you know, missing a detail, but... Yeah, I definitely, I see that question and they were all around the specific V6 apps. Some of these, I mean, Andrew, if you could pull it up, too, and if you know some of these, I'm not sure I know 100%. The Horizon support, I'm not 100% sure I'm not. So this is asking about Horizon dashboard support for V6 provisioning, DHCP V6 stateless provisioning. It's possible that that may have lagged a bit. So I'll have to go... I mean, our end users are able to add V6 addresses into the security groups and view them in Horizon today. Okay, so that's definitely really good as well. Actually, Andrew, you actually know the answers to some of these as well. Yeah, I know that one and two, I think, from a security group in Horizon dashboard, I'm pretty sure that's working right now in the upstream. Yeah. So number three, the DNS mask enhancement service server, that could not make it, I believe, right? That's because I think we need the RAVVD support. Yeah. Yes, that one did not make it. But now the functionality is handled by RAVVD, right? Correct. Yeah. And scalability numbers, that I'm not 100% sure on either. I'd have to look that up. So some of these things that are more specific like this, maybe take them offline or allow people to reach out that. Yeah, I mean, I don't have the exact number, but we are running a pretty large quantity of V6 on the infrastructure. And we can get stats later. And we do do everything dual stack. We're only doing dual stack right now. So dual stack does work. Good. Well, that was from a very large cloud provider that's deploying IPv6. So thank you to whoever that anonymous person is, because we love big time cloud providers. And we'll be hearing from a lot of those at our open stack summit. And I wanted to hear from Andrew and Kyle kind of what you're looking forward to in the Paris summit. Is there specific sessions or topics that you think are going to be really critical, really kind of looking forward to Kilo, I guess, which is what we plan in the summits. Maybe I'll start with you, Andrew. Do you have some specific things you think you're looking to get out of that summit and to make sure, you know, make it into the Kilo release? So, you know, I think as open stacks have been maturing, you know, we've been happy with, you know, how things are being able to support larger scale and more stability. And so I think we're starting to look more at these new features that have just landed in our certain mature like Grove and Sahara, and maybe even some of these auxiliary features like Manila and MagnetoDB and whatnot. So our customers are kind of asking for that full suite of services now. And so we're going to have a lot of interest about where things are going to be landing in Kilo across all those, you know, kind of maybe up to stack services a little bit. And we are definitely going to be tracking the NFV discussion as well. Oh, good. Okay, Kyle, what are your thoughts going into the summit planning on Kilo? Yeah, definitely. I think NFV is definitely going to be going to be a hot area. You know, like I was saying earlier, I expected to be a lot of focus around that. On the neutron side, another big area, which is, again, something we've talked about is this whole known network to neutron migration story and how we're going to weigh that out now that we've gotten more input and certainly hope to get more input at the summit from different users and operators around that as well. These are kind of the types of operational things which are more challenging to work on from a developer perspective because we really need more input from operators and users. But they're really important to those operators and users. Another area on the neutron side where we, like I was saying earlier, was this DVR functionality. We're hoping to expand that a bit more in the Kilo timeframe. One of the areas, like right now it requires OVF, the virtual switch on the edge. There's been some talk of looking at supporting Linux bridge for that functionality as well. So these are some of the areas that I hope to be and the team really can discuss in pairs. Okay, thanks. So one of the questions from the audience was about security and just wanting to kind of get a general sense for how, I guess, we as a community and through the OpenSec releases are thinking about security. You know, maybe Andrew, since you're running in production and are thinking a lot about security, running these mission-critical apps. This wasn't a very specific question, but I think if you just have some general thoughts around how you approach security as an OpenSec user, that might be useful for the audience. Sure. So, I mean, security has been part of the paradigm shift as we move to cloud for us. So, you know, we heavily leverage security groups. We love what we can do in security groups and that we can enable the user to manage a lot of that type of stuff. What we're doing now is spending a lot more time looking at additional tooling for both auditing and reporting around the security. But the fact that we now are living in this standardized infrastructure world with APIs available and the data available gives us an extensible platform to kind of manage all that and create a unified picture of what our security looks like that was actually pretty difficult for us to do before. And I think that, you know, enables our users so previously kind of our security model was ticket-driven and now we've enabled this via, hey, we can manage security via an API or a web portal and that kind of, well, does place more trust in the end-user that flexibility is also a huge win for us. And then the additional visibility, I think more than makes up for the fact that we're replacing more trust there. Okay, great. I think we just have a few minutes left. If there are any other questions, please drop them into the question box. We are definitely excited about the next phase of OpenStack. I think we'll see more and more users talking at summits and the ops meetups and really happy that Andrew is able to join us today. You know, one of the questions that we often get are kind of, you know, what are your business drivers for using OpenStack? You know, what is the actual underlying result you're trying to get from a business perspective? And I don't know, Andrew, if I can put you on the spot there. But I think that's just something we hear all the time and would be curious to hear your thoughts on that. Sure. I think there were a few things that really kind of drove us to look at OpenStack and at Private Cloud. One is we wanted to place infrastructure close to our customers. So we actually placed data centers out deep in our network so that we can provide, you know, lower latency and high bandwidth apps a place to live close to our customers. And that's something that, you know, we weren't able to do on Public Cloud. And then when we started to evaluate platforms, finding an open source platform that we can collaborate on, that we can do things that are critical to our business and can contribute back and be part of that community was a huge win for us. Also kind of the shift to the Cloud paradigm where, you know, we want our apps to think more in the cattle versus the pet mentality and also have infrastructure that they can deploy via self-service or via API. Those were kind of all drivers for us in trying to enable this platform and this environment for our developers to be able to push things out quickly and also have something that they can build upon and deliver, you know, innovative, reliable apps. Great, thank you. Well, we've just got a few minutes left, so I wanted to ask Kyle if you had any final thoughts on OpenStack Juno before we wrap up the webinar today. I guess my final thoughts would be this was my first cycle as a PTL. So the experience has actually been incredibly positive and very exciting actually. The team of people that kind of worked behind the scenes to make these releases happen is really impressive. It's a team of people that are global across the whole world and it's amazing what happens behind the scenes to make these releases happen. And, you know, I thought I'd just give one shout out to our release manager, you know, engineering director at the OpenStack Foundation, Tiri. I mean, he does a great job making sure things are wrangled together and, you know, he definitely deserves a shout out for all the great work he does to make these releases happen. Yeah, thank you for saying that. I think, you know, Tiri Perez is the release manager if you don't know him. He's an amazing amount of work. I was really tempted to ask him to join the webinar, but, you know, he has already done herculean efforts to get, you know, the biggest, most innovative release we've ever had out the door. So I thought I'd spare him, you know, at least one little chunk of time. But thanks again to everybody that worked on OpenStack and including the infrastructure team, which is a big part of what makes OpenStack happen. And for everybody's knowledge here as we wrap up, this webinar will be recorded and put online shortly if you want to go back and look at it or share it. And I put the Twitter handles in the live audience messaging box down below if you want to follow our speakers today. So thanks again, and I hope everybody has a good time with OpenStack Juno. Thank you, Mark. Yeah, thank you.