 So first of all we wanted to thank you for taking the time to join us welcome We really appreciate you're going here. There has to be 15 different things or 20 different things We couldn't do so your presence is very important to us And hopefully if we don't inform at the least we'll entertain right? So before I get started what what we wanted to cover today was four things the first one is we wanted to give a brief Overview of the session. I'll take you guys through it My colleagues over here will then go through the challenges of what we're facing now earlier on we've had we've heard several times in today's Keynote sessions that now the the majority of our of our open-stack community were now in production Later on I'll go through some statistics So now that we're sort of in the in the middle to the tail end of the engineering problems of the cloud The biggest problem that I feel we actually face is a migration of business users into this cloud There say that we'll also talk about the solutions. What strategies tool sets and services and processes have we seen? So so as a bit of a background I work for a company called Ericsson and transformation is something that we do very well from Transforming networks transforming next-generation OSS and basis systems Transforming IT systems transforming data centers and hopefully the combination of what we've been doing from the traditional world to the new Virtual and cloud world we can give you a flavor off today and finally a forward view What is it that we together working as a community can put our heads together now? We we're working with different partners and trying to solve this migration beast this migration challenge and Then I'll give a run-through through it But before I do that let me start off with some introduction. So my name is Earl Villanueva I lead a services unit within Ericsson that just does cloud platforms We're actually in the in the open-stack top 20 list right now To my at the right most there's the gentleman here with the name of Fred Karris who's somebody who's done things from not not just large Open-stack migrations and deployments. He's even done things that he probably shouldn't have told me with spy satellites and what not Rodrigo Morales who's a lead of our cloud center of excellence. He's also an excellent excellent open-stack architect So hopefully we'll be able to share with you the examples of the pains that we've gone through the learnings that we've had Now speaking of learnings it I thought I'd start off with this one little picture that it seems that now the more and more that we get into it The more and more that we get into open-stack seems to be things seem to be getting harder not easier Okay, we've deployed open-stack now. I need to figure about the life cycle management. How do I update it? Oh, okay? Now now if I'm deploying a new a new next generation network I'm not just talking about deploying a new VEPC, but I'm also figuring about what about my underlying cloud underneath Because of course you can't the next generation network needs to sit on top on open-stack cloud for example Now if we if we look at the the data There's a big debate that I hear time and time again with our clients should we go with a public cloud route? Should we go with a private cloud route? Right now obviously and we've heard this in the keynote I think it was Boris who mentioned that earlier that right now in terms of sheer dollar spend It's the likes of Amazon who are kind of opening up the market the the azures of the world But if so if the answer is should I go should I go public should I go private? Okay, sheer numbers. Maybe it's plug public, but wait a minute based on a survey 71% of everybody is actually doing a hybrid strategy So people are and what we found this in a lot of our important clients. We there's one group Pursuing an AWS strategy for example another group pursuing an open-stack strategy So 71% of the people are using a hybrid cloud. Okay, it gives me it's that that's not a good answer Well, that means you have to do both So then we started to look into the numbers from a market data perspective and say if I had the hundred dollars Where should I be spending my hundred dollars? So first of all if I look at it from 2015 to 2020 everybody has a 2020 vision There's going to be a time-stand spend spend in open-stack So what this means to a lot of us is that we're going to see a lot more executive interest management interest And we've heard this we've heard this from the likes of Volkswagen AT&T Congratulations to AT&T and winning the the super user award. We're seeing this trend in proof right now We've heard it with our own with our own two years now in terms of growth rate The spend in private open-stack Private clouds such as open-stack is actually going to outpace that of Publix cloud now There's a reason why accent why why AWS for example is is announcing their partnerships with Accenture's or Ericsson this this is something that we announced in the Mobile World Congress because if you look at the sheer growth rate 36% Versus 19% okay Maybe what's happening is that the momentum that each of us are building in the community is starting to really pick up now If I go into the numbers even a little deeper Let's see where the quality of the growth is at so from from one of the reports that we've looked at where they compared Okay, they segmented the the the the both public and private cloud consumption based on the number of VMs Just from the so it so it's practically a 50% growth if looked from a 22% benchmark from 22% to 31% of large Private enterprise clouds were private clouds were on we're on Enterprise clouds were on the on a private cloud of course open-stack is one of the the fastest growing Private cloud for private cloud types out there So if I look at what's happening now if I look at the growth spurt and if I look at where the quality of the growth Is large enterprise clouds? I think it's an affirmation of what we're doing and what we're hearing and what we're striving to do Now we've heard many times the whole pets versus cattle discussion So a lot of our efforts right now is in engineering the cloud and that and too often the problem this okay Now that I've engineered the cloud mr. Business users. I have a fantastic CICD environment. I have a fantastic DevOps capability Let's start moving on the first thing there is that how many people are actually ready to move into the cloud So it's not just how many how many? Cloud native applications are actually out there. How much work do you need to do to make something closer to the Tosca type of model? In a pets and cattle's paradigm that a lot of us drive towards it kind of takes care of itself But the reality is if you do a wild migration Thousands die now you might like I said you might have the most advanced DevOps capability, but at the end of the day Many business stakeholders will say okay. You have all this capability, but I need a migration window. I need a certain batch schedule. I need a u80 type of test So more and more it starts to go into good old waterfall sometimes, right? So rather than thinking of it as a oh now that we're in cloud its pets and cattle. I can just migrate people over What Fred where you're gonna? I started brainstorming last year said we were saying hey we need a better Metaphor to this migration then we realize that you know what in the US that people move an average of 11.4 times in their life They're more or less going to understand what we mean when we say if you think about migration into an open-stack cloud The best the best thing that each one of us can can relate to is when we ourselves move right and when you move You're not just putting your things in boxes and moving it. You're actually moving your whole life You're moving where your kids go to school. You're moving your mailing address You have to go and change all your bank accounts or credit card statements So imagine you're just one person and it's moving as a massive change that you have to go through What more with with a tenant where I I have millions of Transactions that I need to worry about hundreds and thousands to millions to tens of millions of customers Why do I move? What's a TCO for me? What preparation should I do? How do I know that I move everything down to the network settings and my security certificates? How do I test how do I unpack and make sure and once I'm there? How can I take advantage of my new home or take advantage of all of the features that open-stack has so with that I'd like to pass over the floor to my colleague Fred to take us through through this more perfect, thank you Earl and Yeah, I the more I talk to Earl and to Rodrigo and other colleagues that have been involved with some of these large migrations in the past and current migrations to the cloud the more I became convinced that the The sort of the overall metaphor of our move is is absolutely correct. So As Earl was describing, it's not just your Lot your boxes that you're moving when you move your house you're actually moving your life and the same thing Occurs when we're moving Open-stack large application systems of the tenants in the cloud environment These are operating businesses they've got real customers and it's their business life that you're actually moving so The sort of approach is for a typical migration and just to put it in parallel with this is You know just as if you were moving you would basically decide to start looking at what? Hey, where do I want to move? I want to be in this school district. I Want this size of house my family is expanding those type of things So there's an entire move sort of strategy and then discovery of of all of your your needs That has to happen and then ultimately hey my long-term plan is to get into this city this neighborhood this cloud by the way and then After that we basically start wanting to do the detailed look just like the detailed planning which For the migration itself comes down to some things that I'll talk about here in a minute But in the house analogy, obviously we start inventorying how much furniture do I have and those type of things? Then there's the execution. That's actually the day of the move And I think if there's any point that I'd like to make sure we get across today is It's not just a technology problem any more than a move is simply a truck problem There's much more to it than that After we move we want to make sure Do our verification everything in this operating business this life is ready to go and then finally I'll talk about Optimization as well Okay, one other thing before we start going through the process. I wanted to mention about the Successful migration approach is really very holistic And it has to incorporate elements of people and process I really liked what one of the speakers said this morning in the keynote addresses about how it's a cultural shift I mean those are people and process issues not just technology issues and One of the things that I found is extremely important as we approach these things is to get cross functional Participation and as you'll see there's a lot of different moving parts as we go through these migrations It's just not simply the cloud technology and the tenant VMs or workloads The second thing is we're looking to do I think an unprecedented way of migrations with this very sizable expansion of clouds and we hope they're open stack clouds in the enterprise environment We have to industrialize harden these processes one so they're not susceptible to error But also so we can improve them We'll talk a little bit about that as well and then finally there is technology involved These are large-scale migrations anything that can both automate it and make sure that it's foolproof is something that we need to concentrate on Okay, first a deep dive down into the these elements So on the strategy the first thing that we found is the need to go ahead and Work with the enterprise that's starting a large-scale enterprise cloud buildup and you understand what their strategy is for for rollout one thing particularly important if if the enterprise time scale can allow and we know we're doing it rapidly but There's there's different options right you could move your your your most big applications first or go for early adopters first We'd certainly recommend doing that Approach if you can simply because you want an early adopter that he's gonna go through frankly a real rigorous discovery exercise Migration strategy with tenants understanding their business issues Where they're gonna be located what their requirements are vis-a-vis their customers because these are operating businesses And then we have to look at that cloud discovery We'll talk in a second about how we have some automated tool sets and then more are being developed to do this type of work Looking at the actual application architectures out In these environments because here's an interesting thing we found in the enterprises. We worked with Even where applications are currently resident on clouds They are not 100% cloud ready there. They tend to be hybrid 80 20 90 10 On the cloud but also for instance external networking That has to be engineered into these moves as well. So that's a very important Function and finally there is the overall right roadmap. How are we going to get you over? Mr. Tenet you operate mr. Mrs. Tenet you operating business because we want to make sure that we get you over You know your active active passive setups in the right way and and so forth So those are considerations that we take into account during this phase Now I just mentioned about the discovery process The business Perspective is clearly understanding exactly where the business Is in terms of their operations so that we can now schedule them and move them But there are some technology tools and ways we go about the rest of the discovery and for that I'd like my colleague Rodrigo to talk. Thank you Fred So as Fred was mentioning earlier is not really a technical challenge to my great attendant But the technical part is very important to write so Other than all the process that you need to do in order to handle all the Tenet requirements you need to make sure that all the the virtual Infrastructure that the tenant has in the legacy cloud is going to work in the the new clouds the destination cloud for the tenants so if we look at open stack and everything that that tenant has for a Virtual infrastructure so starters networking virtual routers Virtual machines the sequence that you need to create those virtual machines in the destination cloud everything Needs to be accounted for so as part of this discovery process that Fred was mentioning You need to account for all those things because sometimes the legacy cloud is completely different from from the new cloud You may be changing the architecture complete the architecture Completely in the the new cloud and you need to account for all of this you may changing the Networking for a software defined a network that is completely different So as part of the discovery process and you need to go through over all those those challenges Technical challenges and make sure that the tenant who is going to to work And perform Accordingly in the destination cloud Thank you. So next step of course detailed migration assessments and preparation The first thing that I want to stress here is the need for labs I mean, I think most people in in agile development open-stack work Appreciate the needs for high fidelity labs. It's that and more when we get into migration for two reasons one as Rodrigo was just saying the tenants are actually moving even if they're from a prior addition of an open-stack cloud To a new version and maybe a new architecture Getting that vetted out and certified is Incredibly important to do that before the migration rather than where we're in the active phase of actually migrating workloads over and the second reason Is we need to have a way of certifying our migration tool? I mean a migration tool and we'll talk about that in a minute to Rodrigo go over some of the technical aspects is really a big transform function going from cloud a to cloud b and Making all of the very very sophisticated Adjustments not just copying workloads and pasting them over but all of the different interactions that happened between The VM and application and networking for the tenant setup in the source cloud to the destination cloud There's also what we found in in moving some very large enterprise Applications as I mentioned before No tenant that we've dealt with in some of our large adopters of open-stack our 100% cloud certified there they've got special requirements could be external networking could be certain storage type of of things that requirements that need to be Incorporated into the new new cloud made sure and because it's not there when we move them I can guarantee you from personal experience the migration team is the one that's at fault So we want to make sure that that's taken care of and then finally it's really phasing out and understanding What it's going to take to move these these workloads We've moved now some large enterprise applications that have a storage volumes measuring in the many terabytes hundreds and hundreds of virtual machines and very large storage So planning that out is that going to be able to be accomplished in one day? Five days those type of things so that we understand what the time impact is going to be and realize that we're operating in Production zones we'll talk about that in a few minutes So that's a very important thing as well to take into account how we approach this So with that I'm going to turn it over to Rod Rigo and let him talk about two of the technical areas again so again mentioning the the lab environments right to test the the migration process right what sometimes We saw happening is that the some of our customers they they will start with the Very early open stack deployment They they struggle a lot trying to make their applications to work in that environment they spend a lot of time then they change the architecture they go with something new and They don't test the migration process in in a lab that really reproduces the the legacy cloud and The the future the new clouds that were where the tenants are migrating so Well, it's it's key to have an environment a lab environment to actually simulate those situations and and then run some tests in an environment that actually reproduces what you have in production, so It's very hard to build a lab reproducing a legacy cloud, but it's it's needed and and also the The new new cloud deployment the new cloud where the tenants are migrating to It has to be made available to to the tenants to certify their their applications there because the architecture may be completely Different and they have to run their applications there in a lab and make sure that that they're going to work after the migration Okay talking about migration engine, right There there is part as Fred mentioned it earlier There is part of the the the migration process that there is a only processes But then you you need some sort of tool to do migration especially in large dealing with large tenants with thousands of virtual machines with a huge storage Space that you need to migrate sometimes through a new storage technology So this is just a very high-level architecture of a migration engine where we have The the the component that is responsible for scanning the the legacy cloud and and these Scanning components should be able to actually log in as a tenant Right and try to to find to discover all the virtual infrastructure that the tenant may have Including all the virtual machines the networking everything that is involved For that tenant to work in a legacy cloud and then later the migration processing part of the This engine should be responsible to actually copy things from the legacy to the new cloud And that is a key process especially when you're dealing with large amounts of data The networking has to support this migration process If you're parallelizing things it has to be very well controlled because there must be certain things Things created in sequence in the right sequence You need to first lay down the virtual infrastructure with networking Storage everything and then you start creating the virtual machines So there are some solutions, but I don't think we still very very mature in that that space When talking about open stack and open source tools to do this this migration. This is the where all the fun starts when you actually Start this process and you look at the virtual machines migrating from one place to another, right? And then after that you certify the migration and make sure that the tenant is in the new cloud We're great, and I said before it's really not all a technology problem But there is actually one of the technology problems I'm I'm I know of at least three different independent efforts from three different companies to build these Migration engines Erickson by the way my company being one of them, and I think I'm safe to say no one has has Got it perfected yet. We're all on that road to discovery and part of this session Is actually called action to even see how we can get that more Industrialized for the future because the wave is coming now. It's execution This is moving day. This is the day that the moving van shows up As of right now We don't really have a good way within open stack for cloud migrations to do a live Migration where we can actually keep the VMs and workloads up So just like on your moving day when you move from say an apartment to your first house or from one house to another You probably took the day off and monitored the whole thing and this is exactly what happens as well in the scenario of a migration Currently so there is a very sophisticated Procedure that has to be built because there's a lot of people involved a lot of technology involved and in any enterprise that you're dealing with The migration team is going to conform to whatever the operational procedures are for working in that IT operational environment because you're usually working in a Cloud that is actually in production and may be taking one of 10 tenants that's on that cloud and moving them over while not impacting the other tenants and then the same is true For the destination cloud so all of this is happening against a backdrop of keeping Not interfering with the operation of other tenants within those clouds But for the tenant themselves there is the freeze they're taking their time off and there is a change window usually put into a Thing into effect and if other impacts are there as well We've dealt with scenarios where one tenant was actually coupled to another tenant application feeding data Obviously all those cutovers have to happen in in sequence and then once we've moved everything over and this is where the Automated process becomes so so needed with this engine The complexity of these clouds and the tenant interactions is is so great that frankly while you could Theoretically have people do it by script. It would be impractical from an accuracy Perspective with with hundreds or thousands of VMs and it certainly wouldn't be very efficient And then finally we obviously use do a quick look sort of a verification test But that's followed by our post move Verification and acceptance and this is a process that literally can go from several days to several weeks As a tenant gets over onto a new cloud depending on their level of complexity that They will want to run through a rigorous set of tests before accepting the live customer traffic and then There's usually a soak in period of burn in to make sure that hey There's just no interaction before we start engaging live customer traffic internal or external To the enterprise So that's sort of these these sequence for the post move verification acceptance As a migration team that has to be worked very rigorously with the cloud tenants before you get into this process so it's all understood how you go through and Certify the end result again, I'm fine. Oh, there's there's no And to to this process without what I call optimization once you finish your house move, right? You get everything in you've sort of unpacked your boxes, but now okay I'm gonna put the pictures here or there or arrange the furniture this way I think I like it this way we all do that, right? This same thing happens as the final sequence of a migration. So first of all There is a need to redistribute workloads now in the new cloud to get the right level of Affinity settings and availability and so forth that the application Desires and had in their prior cloud There is also Frankly some hand-holding there's new services available with each release of open stack and the tenants want to consume that some They know about some they don't My my experience is there's no one better to hand hold the tenant through this process than the migration team So it's really sort of an onboarding and migration team to get them exposed to The new features the the new areas of the cloud environment that they are now in and then finally Many of the large tenants we work with are in a cloud network They exist in different geographic locations You know have high availability active active passive type setups And so there's actually the then bringing those online in the proper sequence and getting them restored I'll say to the exact level of high availability. They were before we went through this migration So that's the the overall scenario. I hope you saw the analogy to the house move Because I think it's very very accurate But I think this is the type of process from start to finish that we propose needs to now be industrialized with appropriate technology backing in order to Make our process is hardened to the point. We can go through the upcoming wave of migrations in the enterprise private cloud environment Finally I have to mention about governance So as you can see from the the model here that you know, we tend to think in the open stack community You know about this wonderful software. It is great software, but there's frankly several other degrees of freedom here Organizations that are involved one is the data center infrastructure and then finally the cloud tenant and of course all of this runs on top of the enterprise sort of cloud governance who owns the cloud and We need to basically bring in to play Coordinated governance and program management across this thing so no Migration story would be complete without saying across the the the entire process at the very beginning bring in Your your key stakeholders from the tenants from the cloud Organization development and so forth to guide this entire process that I just described Okay, girl. Thank you Fred and Rodrigo now on the tenants By the way before I get into these last two slides before we open up for the for the open forum in the Q&A I just wanted to plug that some of my colleagues have a very very interesting session tomorrow called Open open sack tenant perspectives So what we did was say we actually ran a survey We also did in-depth face-to-face or over the phone interviews with different open sack tenants across our different different clients to get what they need and and it's part Of our migration practice that we have now From my own perspective from my personal perspective We have a lot of work or potential work around legacy cloud to new cloud from public cloud to open sack cloud to some other Internal cloud including open sack clouds to public cloud and and and and through throughout all of this What what I see is that there's a big groundswell of support now having having said that We went through the approach we went through and I think but like it's it's by now It's clear to everybody that migration isn't a trivial matter It isn't just a matter of shrinking and growing VMs and moving workloads across But it's actually a fairly comprehensive systems integration approach that we need to we need to think about of course over time What I'll get into hopefully we solve this together over time But I just wanted to point out that if you think about what's the most impossible and crazy cloud Migration that you can possibly ever think of to actually have a colleague in the audience by the name of Palash I won't single him out because he's too modest But we're he's actually running a program in one of my clients before we're we're taking a mainframe system We're maybe the last major code update was in 1994 PL one and cobalt were modernizing it and with the end goal of Well, maybe I shouldn't say how long exactly we planned it It's not it wasn't a short program with the end goal of moving it into a cloud So if if you can consider moving a mainframe system Last majorly built in the early to mid 90s and move them into an open stack cloud, you know It's it's only our imagination that limits us. So from a visionary viewpoint What we hope that together as a community we can work towards is that migration happens with zero downtime There's a lot. There's a very important temptation that we have to be careful about You're there you're presenting to your to your your your business owners and they're saying hey Well, we want to do a live migration. We can always try to do it There are other ways to mask it if it's a true cloud scaling app it a cloud native app You you can shift components off it so the end end user Result is different but at the end of the day for the virtual machines that you're touching right now It's very very hard to have this no zero downtime The other thing that we hope to get is some migration happens at the push of a button We talked about the migration tool sets that we've been developing on our own with Ericsson With Ericsson and other partners or from partners that we leverage and in each case in each case It's a it's it's either something built from scratch Or it's something where we use this like a tool like a cloud management platform And we have a lot of scripting on top of it Right now with there's some degrees of automation that we can do but I think we're not yet There were a migration that happens at the push of the button and finally when we talk about migration It's not just a linear Journey from legacy to cloud but where we see the market heading is that migration will actually be a multi-directional movement across From your legacy to open stack open stack to feature open stack open stack No, maybe a public cloud or maybe a public cloud back to open stack So my last slide is really a call to action a Call to action to the open stack community Industrialized migration is needed. We're working with some of our partners in different projects that that hopefully will will at one point be accepted as part of the The core industrial migration is needed for the because because the rate of production Adoption is just simply increasing the rate of open stack adoption It's increasing and the only way to do this is for the community working on well-defined projects or well-defined Workstreams and from an Ericsson perspective and and I know my partners feel the same way that that we welcome welcome you joining this effort so with that We talked about migration being a beast so hopefully working together. We contain this migration beast So with that we'd like to open the floor now to questions and answers and Fred and Rodrigo here can help me out Yep, and there's so one microphone there and one microphone there. So feel free to ask away. Hi My name's on a kid cloud engineering at box Here's why the moving day analogy doesn't really resonate with me In our production environment. We have hundreds of services Managed by dozens of development teams and when we think about a migration It's usually a phased approach where we have services moving over to the cloud Then interacting with other services that were part of the legacy environment as one larger ecosystem That being one thing the other thing is when you think about a moving-day analogy What do you do with the millions of dollars of infrastructure that make up your legacy environment? Okay, that's a way. Did you want to think I'll let me attempt to do that. So When we're talking about the migration of active tenants at least in our experience my experience it They have to be they can't be moved in a phased approach They're existing on one cloud and they need to then be instantiated and come back into operation on the second cloud So it's a very complex operational problem if if we didn't have that issue We could live in this sort of bimodal environment for quite some time so My perspective is coming from once again I don't know if this is where you were coming from or not But my perspective is in an operational environment We simply don't have the time to go ahead and and sort of do this long more drawn out And by the way, I would admit Logical phased type approach the time and the operational and production needs just don't support that So there are cases where we can group different applications that are fairly self-contained and this becomes a unit of a face But what I'd also like to address you because you raised the second part of your question is really important about what to do with the Infrastructure so first of course, we know that there's this big wave to move into an open compute type of infrastructure Within Ericsson. We also have our own solutions there just like many many other of your partners And in this moving to a new opens I want to have a Google or Facebook like data center It's a it's a different question But one of the strategies we have to try to preserve the existing Investment you've made in let's say your your storage devices your blades that we try to do on an In zone or an in data center implementation. So for example, we spare out some servers This is where we deploy the new version of the open stack and earlier on I was making an illusion where we start to shrink VMs and shift VMs across so as I move workloads over I'm growing the new cloud within my data center And I'm shrinking the old cloud so that's the one of one strategy that we tried But thanks for your question. So in that sense, it is a phased approach then for certain applications You will be doing it same day But what you're essentially doing is from your entire ecosystem you're shrinking some and growing Oh, okay, definitely, right that actually by the way makes a lot of sense And that is exactly what Earl was referring to actually a migration We've just been part of for the last nine months is doing exactly that in the same set of Zone or set of racks in a data center Literally shrinking the one cloud as we're migrating over so when we're finished instead of just deprecating the equipment It actually now becomes the new right clock. So by no means are we suggesting a big bang approach Anyway, there's a gentleman who had the question. Oh, maybe if you don't mind, sir We'll just go over to this side and we'll switch back to you Maybe do quickly dovetail on the previous question Proportion of these migrations. Do you see that happen within the local data center as opposed to public clouds? huh so in terms in right now the Right now there's a lot of and this one. I don't have hard data There's a lot of business interest into AWS for example, but where I see a lot of Harder traction happening because it's like a quicker hit is a movement from let's say a VMware VMware into open stock I think for AWS because there's a very AWS Requires has a very specific governance model very specific way that you have to operate and manage So a lot of IT shops they see a lot more control with if they keep it in-house But I don't have hard data on this but that's a very good question. Yes, sir First thank you for the presentation one interesting thing you said was if I heard you correctly these days It's not a good idea to script everything and it's much better to have a live person do it Can you give me an example or two of things that can't be scripted these days at least to that a reliable degree? Well things that it can be really automated is is for example the creation of virtual machines in the right sequence respecting the affinity and entire affinity rules and and the creation of virtual infrastructure like virtual routers and and Whatever you need in the destination cloud Some some of the tenants that we we migrated they they have like hundreds or even thousands of virtual machines So you you can't do those things manually, right? You cannot create everything in the destination cloud using people so you have to automate that also the Transferring all this data. You need to have something Automatically doing that right you you cannot have people clicking those things and And also the test part after you migrate you need to run some automated tests That is a very important part of the the automation right there, right? You cannot have people executing test cases and manually depending on the size of the tenant If you have a small tenant with three or four virtual machines that may be possible But most of the time you need to automate and let me I think though You asked for a specific example of hey something that couldn't be automated So in our experience like I said what we found in the enterprise environment is hey, we they make Our clients our tenants may be ninety ninety five percent I'll say cloud cloudified and and and and all that but there's still a lot of external networking We've also so that that's one example that can't be automated or is hard to automate Another actually is is is as you're moving from a source to destination cloud. There are Potential conflicts in IP addresses if you just move tenants over and and their virtual machines and the new system The new cloud system starts assigning their IP addresses There is the potential for conflicts there and so those are best handled in a manual way We haven't found a way to automate those yet Thank you for your presentation You present the strategies for migration or migrating business critical application to open stack I haven't shared anything's about the security boundaries on how to mitigate the risk and what's the strategy to migrating These critical application to open stack. I think that alone is it's a whole presentation, right? looking at the security point of view but some of the the things we were mentioning they have to do with security for example the Certificates that you may need and and again mentioning the networking if things are changing, right? And and you need to to to make sure that all all your networking or all your security part firewalls If they if you're changing IP addresses Are you going to reflect on the new cloud all the the security aspects that you had in the legacy one? so It's a very broad Subject, but but there are several aspects that you need to take into consideration Those are great points rod and maybe the other part of your question is from our risk management perspective This one sometimes it can also one asserts the usual. I need the appropriate testing. I need the right rollback I need somebody from the business owner to actually do the verification like what we talked about but The risk here and and there's no I can't think of any efficient ways that sometimes the There the next level of resiliency will actually depend on the tenant itself like for example The behavior of its load balancing capabilities. We we've had seen we've had instances for okay. I might have Three machines as I say a broad example. I'll shift one first Okay, then how does my load balancer behave so I know okay. He's now over here, but he's down Okay, he's now up now now I can load balance across three now. It's time for a machine number two So once you start to go into discussions like that It's almost as if you're doing it tenant by tenant. There's no one-size-fits approach. That's so far that we've seen Yeah, I I'll bring up one other thing too and and that is as I mentioned early on one of the key activities up front Is a deep dive architectural assessment? So when we're lifting and shifting over naturally all of sort of the the perimeter cloud security that gets instantiated But the architecture that the tenant with their firewalls and things like that that they've actually set up within the boundaries of their zone That needs to be very rigorously looked at from how you set yourself up in the source cloud And how is that going to work as we move you is it still going to provide you the exact same level of security or not? So that's actually something we do and frankly it's done by people like Rodrigo that have the technical ability to do it Yeah, some other aspect is do remember that we mentioned the lab right a high fidelity lab So you need to make sure that your new cloud All the architecture that you have there it adheres to the CSO Organization and they will actually look at it at your lab and they will certify you According to all the rules that they have so thank you. Thank you Well, thank you guys you've been a wonderful audience We think this is a very important subject and so feel free to reach out to me Or my colleagues here, but you use me as a point of contact I would like to hear from you and hopefully there's an area that we can generate momentum in the open-stack Community going forward because it's going to be needed. It is needed and I think we're all up to the challenge Thank you