 Hi everybody this is not on yes it is awesome all right all right I need to give a mic to you guys sorry okay can everybody hear me okay yeah so we have the coveted right before lunch running 15 minutes behind spot but I like to think that we are the most anticipated session of this morning so everybody's willing to wait you can't hear me how about this one better can you hear this all right so my name is Sarah Waldner I'm a product manager at GitLab I work on that yes to my mouth there we go troubleshooting I work on our monitoring product so as Priyanka introduced we're gonna take a look at multi-cloud from the provider's perspective but I'm gonna give everyone a chance to introduce themselves first my name is Vicky Glacius I'm a senior staff solutions architect with Google I'm Jason McGee I run call platform for IBM Bill Shetty director of developer advocacy for multi-cloud at VMware and hi I'm Bob Quillen VP Debrell for Oracle Cloud so we spent majority of the morning hearing about the multi-cloud verse the challenges where we're headed the obvious value proposition this is a critical part as enterprises grow a choir different companies how are they going to approach it so the discussion today is going to take a peek into this from the provider's perspective is this a threat do they see it as an opportunity I bet everyone says opportunity but we'll see and their advice when giving how end users should craft their strategy so with that first question you guys is threat or opportunity I can start off with opportunity for all the folks in the room the users you know this is this is a chance for the cloud providers to you know fight for your business and I think that's really helpful for you folks it is also I think an opportunity for Google and the likes of companies that are investing in Kubernetes to make that you know universal substrate for multi-cloud operations yeah so it's definitely an opportunity and like we just heard I mean I look at it from a perspective that it is your opportunity to arbitrate between these clouds for the first time ever right when back in the data center days right and building out your data centers if you ever guys ever did that and when the heydays when vendors came out right you pitted them against each other to effectively figure out how to get the best price and feature set right we are now at a stage in the multi-cloud kind of verse I guess as we're calling it where we now have abstractions along with Kubernetes and others that you can now have an ability to pit one against another to get those price points and feature set so it's a great opportunity for you guys I guess my twist on it is it's just inevitable like I mean the history of IT is not as much as any vendor would love to say like we're going to take over the whole world and everyone's going to run on her stuff it never plays out that way and the cloud error is no reason for that to be different and you guys are going to use multiple providers the different writers are good at some things and not good at other things like we have strengths and weaknesses and multi-cloud allows you to play and leverage those strengths and avoid the weaknesses and kind of solve the problems that you really have right so I just think it's an inevitability that we get here yeah I I'd agree it's a it's a reality today especially as you kind of look at the maturity model we talked about this morning everyone's on that spectrum somewhere and I think every customer I've ever talked to has got multiple clouds running if you run in you bring the SAS providers are using there's multiple clouds are in every enterprise every enterprise has always had partners and a stack of solutions they use for different use cases and I think as the the technology can use to mature and all this infrastructure continues to grow to enable more multi-clouds you'd be have to workflow portability and app portability and all those nice sort of up the continuum of maturity model I think you'll get there but right now it's true everyone is using multiple clouds knows next next chance is where it's going from there so on that topic workload portability data portability as as a product person I'd see that there would be opportunity for providers to potentially partner together to create a more streamlined experience for the end user right now everybody's kind of picking and choosing and building their own strategy have any of you considered that well I'll jump in you know we just announced a integration with Microsoft Azure's Oracle cloud and Azure integration and you know while we're waiting for the technology around cloud native to kind of mature and create all these inroads and on ramps and bridges to cross multiple cloud I think these pairwise integrations will start where we have you know guaranteed interconnect between data centers we've got a unified identity model with single sign-on you have standardized applications that are you know built for with partners we have a shared support model so you know while we wait for the the market to come together like you will see these pairwise and freeway kind of integrations happen until full standardization happens but I think that's a start yes so you know VMware we are looking at this in a very interesting way which is we realize that our customers are going to be on a spectrum right you can either completely on prem and some will stay there or some will go completely multi-cloud public cloud someplace in the middle it's where you're going to land right or maybe on one of those two vectors at the end of the day our strategy here is to give you guys choice where to land on that bag on that spectrum in order to do that we're doing multiple things right one of the things that you've probably seen the announcements we definitely from the infrastructure level if you're on-prem and you're trying to move out giving you the opportunity to actually have consistent operations with vcr on Amazon or Azure or Google right and then if you wanted that portability or if you wanted an ability to take your applications out to start migrating them and have that as we're at kubecon we have no products like tanzu and some other stuff and finally I think that the biggest thing that we've invested in and with integrations is really to give you a consistent set of operations around multi-cloud right whether or not you pick one or another there's multiple abstractions and we are looking at how to actually help you manage that both on public and private so the the end goal is not just the integrations but it's also choice for you guys to be able to get what you want from an operational perspective I mean to go back to your original question on kind of how we are working together I mean I think the fact that there's how many people are here this week well over 10,000 30,000 wow I was gonna say I was a lot bigger than I was expecting you know the fact that there's so many people here is I think a sign that we have all been actually collaborating together in open source for five plus years at least in the cognitive space I mean many more than that in other domains and that you know we've all actually been working together to build this platform I'm actually pretty excited about it my personal history I spent kind of the first half of my career doing Java and app servers and kind of working in that era of computing and back then you know we all work together to write like this giant document this big spec and then we all went off and implemented the spec and then we tested whether it all works the same way which of course I never did today we we kind of saw these problems in a much different way right we actually built code together and and the tension I think you sometimes see between vendors whether they're software vendors or cloud vendors is okay there's an edge above that hasn't been done yet in open that we're all trying to glue together or wire together make easy but below that there's this constant evolution that we're we're making the projects better we're solving problems in a more consistent way and that's I think good for everybody that we're you know we I think for the first time that I've seen in 20 plus years in IT we have one platform in Kubernetes that literally everybody agrees is the runtime platform we're going to go build on yeah and I think that the nice thing is that it's a collaboration between users like people who are just part of the community and want to start using it you're placing your bets along with all of the providers and then eventually we're converging I think three years ago Kubernetes was not you know the ubiquitous thing we've all landed there now and we're moving forward on that decision so it's kind of interesting to have all the users and providers kind of betting together and eventually a winner comes up so we are at cubecon how do you think open source what role do you think open source plays in developing or evolving multi-cloud strategies for different enterprises I just said what I thought right there it's all of us being in the room together and choosing together I mean there's the obvious technical alignment piece of it which I think most people will probably get one of the other dimensions I think is interesting about open source is skills you know like I talked to lots of enterprises and as a as a technology company you know that the people that I have on my teams I have a lot of clients who would love to be able to hire talent like that on their teams right because you know they it's hard to get these skills and I think open source and these common platforms are building also allow companies to build skill that works across all of us that works across all these environments and it's therefore much easier for them to kind of build a workforce that can go fast and stuff so like sometimes I think from a user perspective that's an unseen benefit of us all working in open sources we're all building the same skill base in the industry yeah you said it best I think in the earlier question right I mean I think open source is effectively good for us right it gives us the ability to all contribute and get the consistency that we need the API is the abstraction levels and without it I don't think we would be able to achieve where we want to get to in multi-cloud which is you know as some of the first questions really to get to that level of homogeneity right cross multiple clouds right you want to be able to pick the only way to do this is to allow us here like a cube con is a perfect example of just contributing and getting into it right and and starting to contribute and bring up potential issues and ideas to build on right that you're seeing it's a constantly evolving process right there's no end point here necessarily yeah it is no doubt the catalyst for the multi-cloud future that we're going to there's no doubt that without open source we're not going to get there together and we could continue to build the proprietary services that are you know vertically integrated versus kind of horizontally integrated and so I think there's a it's work to be done and to fill in a lot of the gaps to bring people forward on the culture side and kind of get them through the training process and to connect and bring over all these existing enterprise applications was hard work to migrate especially a running application and we're seeing more and more tools and techniques and you know some presentations even today around how to enable what are the what are the gotchas where the lessons learned how you do these migrations to a multi-cloud world so can I have maybe one counterpoint the one thing I think open source is sometimes not as good at is baking kind of operational knowledge into the code like open source projects tend to be and I know it's a blanket statement but like they tend to be really good at kind of feature capability and kind of iterating our capability and then we all leave like how do I actually stand this up and run it as something else and and I think that's a really interesting area where there's a ton of activity going on and I know as it's certainly as a public cloud provider and I see this in in Anthos as well and the work we're doing in multi cluster management at IBM like taking the operational experience you know like at IBM I run the our Kubernetes service we have like 20 plus thousand kube clusters that we're running like taking the knowledge of running 20,000 clusters and baking it into the environment is super valuable and like where a place where open source it takes a lot longer for that to kind of work its way in did you want to add to that no leave that one so a number of times this morning before the panel we touched on acquisition as often a driver or kind of the catalyst behind a particular enterprise or company becoming multi cloud and I'm curious for the audience are there proactive steps that enterprise can take tools or different tool chains that they would want to invest in that perhaps are better suited to multiple clouds than others I think one one thing is that all abstractions are bad some are useful so eventually you have to choose some abstraction that you're gonna have for your multi cloud days you know we see things like control plane and other places like just choosing Kubernetes as the abstraction you're gonna have starting to figure out what that abstraction is gonna be a lot of us in this room have properly chosen Kubernetes maybe thinking about a layer up on your roadmap of how you're gonna choose your abstraction if and when you land in a multi-cloud scenario so that you're kind of prepared for those situations yeah that's one aspect absolutely and I think as you as you're buying other companies and you're trying to merge these different environments and cultures at the same time together right it's it's a little bit of an obvious it's a mess but I think well one of the best pieces of advice is understand what your end goal is right and how you want to use let's say multi-cloud maybe they're in a different cloud the important aspect here is is as you get into that and you have multiple semantics that you're now starting to deal with maybe on Amazon and somebody you bought is let's say a sewer at the end of the day right the first thing you want to do is to really ensure as an organization that you have the expertise and the skill sets to kind of manage that multiple sets of semantics right because it is not just about having silos you're gonna want to have to be able to work together and so having a unified I would say an operational team that manages a multi-cloud environment just how to be Kubernetes only right this is about understanding like our back and I am models across multiple clouds which are a mess right and they're all have their different components getting that knowledge and skill set started is important by far you know in the beginning right as you do that and then secondly is understanding I think in my viewpoint is getting the observability tools that are required and not just obviously the Kubernetes level but also at the cloud level and married the two and understand what you're gonna use you may use Kubernetes you may use you know lambda you you know understand all the variants and finally I you know understand your process right how you are going to manage that new addition right you leave it alone and and it's a slow process but understanding the process and how that may be melds or merges in is gonna be important and as you understand your process you can constantly iterate so it's it takes a bit I agree with both of those comments I I guess what I would add is the thing I try to work on a lot with my teams nice talk to client slot about is just velocity like one of the ways to prepare for acquisitions new things coming in something being different is to really focus your team on like how you deliver with speed and like react to change quickly don't get caught up in the like we're gonna pick the perfect platform and define like all the attributes of it make all these strategic decisions and then we'll just put everything in it because by the time you're done with that process like half your decisions are probably wrong and so you really have to like figure out how to get high developer velocity because then you're just more able to deal with the change and it's interesting I think one of the things that's in the whole kind of Kubernetes microservices space I think we have this like built-in assumption that you're good at DevOps like you're good at delivery and integration and the reality of the market as many people are not very good at it like we just kind of assumed that they were good at it and then all these cool things we built will be great you have to like put effort I think into that like put effort into figuring out how to do rapid delivery and then you'll just kind of be able to adjust courses you go I'll just add one one more thing that Kubernetes is not gonna solve your DevOps problem yeah and I want to just reiterate what you said I mean you think about the velocity and change that's occurring on the cloud itself right now right the number of features that Amazon releases the number of capabilities and deltas that are occurring whatever process let's forget about tools I totally agree with them understand your process know where your SLAs are what you want to achieve and wrap yourself around that first before you start picking stuff because you're gonna pick you may change it in a year what you pick as a tool but you want to know what you want to achieve yeah I would just add on setting setting standards and kind of architectural models as early as possible not allowing things to fester I think even internally inside Oracle you imagine thousands of thousands development teams tens of thousand developers you know many different clouds being used over the last ten years and platforms etc we have consolidated a lot of that in some standard models and and achieve a lot more velocity by having at least some top-down push of standards and moving to those you know it was kind of a late move and in the jungle developed and took a long time to prune that back so I think the sooner you set those and enforce them top-down you know the more able I think you are to move quickly without having you know these pockets of abstractions happening all around you so awesome thank you gentlemen so I think that we have about five minutes left and I wanted to give all of you the opportunity to ask for any last pieces of advice or the burning questions you've had before we go to lunch you can set for us take it and run comment about the cost of being multi-cloud I mean I come from an academic environment but we spent the last five years being multi-cloud due to the academic environments and it's cost us at least two years sorry what it's cost you what at least two years two years a lost time time okay can you comment on the amount of time you think it takes to be multi-cloud and is it ultimately worth it I can say five years ago it would have been much much more difficult than it is today I think like I said before we're we're getting useful abstractions now that can if you started today it would not be a five-year journey with two years lost I think it would be a much better ratio and I think it is like you got to just start choosing an abstraction and then going down the road being agile and how you kind of leverage that and make sure that there is some plugability there in case you have to make it you know a change later down the road but a lot of the stuff that that we mentioned before is different today than it was five years ago or even two years ago yeah I think that's true I also think I don't know many people who sit down ago okay today I'm gonna figure out how to be multi-cloud like it kind of happens which maybe isn't a great thing but that's kind of how it happens like people just wind up being multi-cloud and in a lot of cases I find it it starts in kind of isolated pockets like this team this business line whatever picked one thing and somebody else picked something else like as a cloud provider I love reading articles or it's like so-and-so customer is a Google Cloud customer or Microsoft Azure customer Mike every one of us can publish that article about the same person because some part of that company is using us so it kind of just happens and it's really kind of working backwards on like right now how do I drive some consistency across them I think one of the hidden costs of multi-cloud is you have to be a little bit more intentional about what you run where and what the interactions are between things like I was having a meeting with a European insurance company and they were telling me about this one workload that they have and it runs an IBM cloud on Azure and AWS and in kind of one cycle of the data it bounces back and forth between all three providers three times each like in out in out in out and I mean one I can't imagine what the performance is of that but to like think about egress charges and all the kind of actual cloud costs that goes into that so like once you're good at multi-cloud you can be too good at it and you actually have to start to put some structure and to put some sanity and and I think data is often the center of gravity that kind of leads you to these apps go here and those apps go there yeah I was gonna say it gets better and it's getting better that's why we're here today I don't think the tooling was there two years ago I mean even for a single cloud was hard for multi-cloud I think we're still some different platforms away some more maturity away but people are addressing the problems now and I think it's we'll see it improve much more of the last next two years I think yeah no these are all great comments I think the only thing other piece of advice I'd give you is as you it just happens right I mean it will happen one of the things to always remember is to understand what the services are that you are going to allow your and users to use and have a handle on some of that right in some cases it may run away from you and you may not want that right and so having a little bit more of title control around the costs and utilizations right and the type and components that you're going to allow them to use well I'm not not gonna say it's gonna completely solve your problems but it's definitely gonna start mitigating some of that right but it and it's always changing right so even it to spend two years and it there's probably new things that are gonna come about and it there's always a process here it's gonna continually change so yeah that's it sounds like you were when everybody else was just starting lift and shift is when you were going multi-cloud it's like trail blazers often traverse uncharted territories so hopefully you learn more than other people though other questions oh yeah all right so you mentioned that Kubernetes kind of allowed this whole multi-cloud thing to become smoother what do you think is still the biggest missing piece in going forward with that storage and networking those are the hardest pieces okay so it's definitely storage right or databases right let's just talk about state information the other one is okay networking I'll leave that there absolutely but I think that I am integration from the multi-cloud kind of models back into Kubernetes are backs and having some sort of consistency is the biggest problem right now it goes back to policy management right and just having that level of consistency it goes to control right that's not there I don't want to get it is is there any like plans in the work or software in the work that you guys have to kind of address that yeah well I think we've seen a few people trying to dress it no VMware and Google are and you know from a we talked about vendor you know vendor relationships you know Oracle's working VMware on this kind of stuff we work with Azure on this kind of stuff to get work on the identity issues to work on the partner the partner the ecosystem around that the support around that so I think there's this work that's being done at the open-source level the standardization level you're seeing products emerge that create a nice unified platform a unified control plane we've heard some discussions of that today those are all positive developments those are early but that's kind of what's needed to kind of lift the market up to the next level I think so all right all right guys thank you so much I think it's time for lunch yes