April 22, 2024

Exploring Alternative Hypervisors with Apache CloudStack | Episode #82

Join us in Episode 82 of Great Things with Great Tech, featuring Rohit Yadav, VP of Engineering at Apache CloudStack. Rohit and I talk about the evolution of Apache CloudStack from its early days at cloud.com to the Citrix acquisition, to donation to the Apache Cloud Foundation to now, becoming a key player in providing turnkey, scalable cloud solutions across various ind…

Join us in Episode 82 of Great Things with Great Tech, featuring Rohit Yadav, VP of Engineering at Apache CloudStack. Rohit and I talk about the evolution of Apache CloudStack from its early days at cloud.com to the Citrix acquisition, to donation to the Apache Cloud Foundation to now, becoming a key player in providing turnkey, scalable cloud solutions across various industries

Not a hypervisor, but a management platform with hooks into many virtualization stacks... #Apache CloudStack orchestrating #IaaS environments across cloud infrastructures including public, private, and hybrid setups, containers and more!

Join us in Episode 82 of Great Things with Great Tech, featuring Rohit Yadav, VP of Engineering at Apache CloudStack. Rohit and I talk about the evolution of Apache CloudStack from its early days at cloud.com to the Citrix acquisition, to donation to the Apache Cloud Foundation to now, becoming a key player in providing turnkey, scalable cloud solutions across various industries. Apache CloudStack has matured into a robust cloud orchestration platform, enabling enterprises and service providers to efficiently manage large-scale cloud environments. With its comprehensive support for multiple hypervisors, seamless Kubernetes integration, and low total cost of ownership, CloudStack continues to be a pivotal force in the cloud computing sector.

Technology and Topics Mentioned: Apache CloudStack VMware, KVM, XenServer Kubernetes Integration CloudStack UI Enhancements Terraform and Ansible Multi-tenancy and High Availability Open Source Licensing: GPL and Apache Virtualization Technologies Storage Solutions: NFS, Ceph, MinIO Network Virtualization and MPLS Support VMware and Broadcom Developments Cloud Computing Platforms: OpenStack, OpenNebula

☑️ Web: https://https://cloudstack.apache.org/
☑️ Crunchbase:https://www.crunchbase.com/organization/apache-cloudstack

☑️ Support the Channel: https://ko-fi.com/gtwgt

☑️ Technology and Topics Mentioned: Truebit, Blockchain, Offchain Verification, Web 3.0, Decentralized Applications, Game Theory, AI, Supply Chain, Compliance.

☑️ Be on #GTwGT: Contact via Twitter @GTwGTPodcast or visit https://www.gtwgt.com ☑️ Subscribe to YouTube: https://www.youtube.com/@GTwGTPodcast?sub_confirmation=1

☑️ Subscribe to Spotify: https://open.spotify.com/show/5Y1Fgl4DgGpFd5Z4dHulVX • Web: https://gtwgt.com • Twitter: https://twitter.com/GTwGTPodcast • Apple Podcasts: https://podcasts.apple.com/us/podcast/id1519439787?mt=2&ls=1

☑️ Music: https://www.bensound.com

Transcript

RAW TRANSCRIPT

back in 2010 I was working on a project to find our next generation of cloud control plan for our infrastructure as a service platform one of those was cloud.com or cloudstack and while I didn't choose it back then if I had my time again and we were looking at this today Apache cloudstack would be top of the list let's take a listen this is episode 82 of great things Great Tech hello hello and welcome to episode 82 of great things with great Tech the podcast highlighting company doing great things with great technology my name is
Anthony speri and in this episode we're talking to a technology stack not so much a company one that's versatile and scalable in the open source Cloud orchestration world that supports multiple infrastructure as a service platforms underneath it it can handle various types of cloud environments including public private hybrid and Edge Technologies key features include deep hypervisor support which is very important in this day and age platforms like KVM VMware Zen server xcp all highly scalable and then a
multi-tenant virtual networking model that platform is a pachy cloud stack and joining me today is RO Yad VP of engineering at shap Blue the former VP of Apache cloudstack and a member of the ASF and on the project management committee that's a lot to get through welcome to the show Rober uh thank you Anthony pleasure to to uh to to see you and meet you have seen your podcasts in the in the past so always uh you know good to see you and lovely to be here with you good stuff well we're we're on it now with we're
together talking about Apache Cloud stack so before we get into that world and it's it is a fascinating world and a world where I've had some you know historical interactions as well with it which we'll touch on later if you like great things with great Tech and would like to feature in future episodes please visit jtwj t.
com catch all episodes on all major podcasting platforms Google Apple Spotify please subscribe to that hosted and distributed by Spotify podcast finally head to YouTube at GTW JT podcast hit the like And subscribe button to get all past episodes and keep up to date with what's coming so with that out of the way roer let's talk about um your journey um with becoming you know so involved into the Apache cloudstack and ASF in general so you give us a bit of an origin story in terms of how you got to be yeah so um well um I mean the short
answer is I started with Cloud stack project in 2012 I was hard out of the uni by CX I started in the Hyderabad office then worked uh in the Santa office for a while but the backstory of how I got you know in that place in the first place is in 20110 actually I uh joined CERN as a volunteer summer student uh was mentored by by Dr Ben seagull who is the mentor of Sir Timber nurley really exciting times for me back then to learn about the technology in general and they were doing something with bank to sort of generate interest
among the crowd about LC and they're doing something with virtualization so that's where my journey with virtualization started and when I got the job opportunity I jumped right into Citrix because back in that day they were doing Zen server which is a hypervisor I got placed into that team to work on Zen server and then uh the news kind of uh hit that you know they acquired a new startup called cloud.
com a bit of history about cloud.com they started in 2008 uh and some of the founders had the privilege and the pleasure of working with chir and Alex Swang and uh Shang leang will Chan and Shannon Williams and and the whole gang uh and um yeah I started there I was just a rookie started to learn about clot stack you know learning from the the leaders in the AUG as a sponge and these guys actually started in a small garage and and the the reason I know know about this is because I got the opportunity to work in the Santa Clara
office as a tiger team to work on a sort of next version of cloud stack where I would hear about these stories in during lunch hours from you know from straight from the founders Alex Wang and ship in particular were software Architects and that's whom with whom I kind of worked very closely for those two to three months uh so they started in a very small uh I wouldn't say garage but a small office so four of them yeah and their idea back then was sort of a clone of ec2 sort of AWS what AWS back then was
started as S3 and ec2 they started with a virtualization layer which they thought back then that they could sell it to hosting companies because their idea of go to market back then was that you know AWS is going to get into the hosting space so maybe we build a tool which will be used for hosting providers and then they they actually got Telos excited you that was probably one of those areas where you know Telos started paying more attention to Cloud as a as a thing you know uh as a utility almost like electricity and whatnot um and then
they got reband um I mean initially they started at VM offs later on they bought the cloud.com domain I think from juniper or some of the company and that must have been a great coup right like I remember that when I we talk about my little interaction with them at that time but yeah cloud.
com like holy that's like that's like the mother of all domains right absolutely and then that was a really strategic move a good sort of marketing move uh by Shannon and the rest of the team back then and they first uh thought you know they they thought open source as a as a good sort of way to uh improve adoption so they licensed their product uh their initial versions under GPL and then Citrix acquired them so Citrix acquired cloud.
com they moved into the cloud platforms group that's what sort of Citrix rebranded that that whole uh software stack as C Cloud platform Under the Umbrella of which they would sell net scalers of the world and Zen server as the hypervisor and uh Cloud platform of course as the virtualization layer which will work with all these components and they pivoted again and they sort of changed the licensing to Apache because well in in our industry you can never go wrong with AA license it gives people flexibility to poke the code and make changes and you don't have
to share it like you would if you were to stay with the GPL license uh so that was 2012 when the uh the the code base was donated to the Apache software foundation and sort of stra straight up donated like donated completely interesting I mean of course yeah I think they they maintained some secret sauce uh which wasn't that big so from my understanding when I joined around July or August 2012 um much of the work was done in community but Citrix had this sort of you know transitioning issue like what is what are we doing you know this GPL
now we are in the ASF and we have a a private folk of uh Cloud stack which was a tri Cloud platform so they were trying to do things in the community get a version there and then put in their secret sauce yeah yeah I remember this because obviously you know so from my point of view um I remember I was tasked um at one of my around this time 20091 to basically find a cloud management platform for infrastructure service because before that we were using VMware we were just doing hosted uh virtual virtual server it wasn't really you know
scalable it didn't have all the multi-tenant aspects to it so we were looking at you know what do we use and back in 29 2009 2010 days it was really only a couple things out there you had open stack I think Zen had an or Orchestra which was I think actually which has got links to xcp um and and VES and whatnot so it's all kind of inter wound isn't it to a certain extent um yeah but I remember being very keen on on um cloudstack because it was cloud.
com I remember it being VM Ops uh and I remember that it did everything pretty pretty well like in terms of deploying VMS working with the networking the multi-tenancy all that kind of thing but the reason why I didn't end up using it was for a very simple fact it was because it didn't didn't do the console at the time it couldn't bring up a uh a VM console sounds like table absolute table spakes today right but if you think about ec2 and AWS they couldn't really they still almost don't do that you know they they do it they do it to a certain extent now
they do it through the SSH and and whatnot but you couldn't bring up that actual VM console and that was a big part of why we didn't choose it and why we went down the VMware Cloud director path and that's a whole different you you know trajectory and kettle of fish for me but I do remember this a lot of the controversy that was around when Citrix acquired acquired cloud.
com what they did when they split it up um and we worked fairly closely with the team um in terms of from a competitive point of view a company called ninefold um in Australia which was an offshoot of mcari Telecom and they were they were T this as the the Next Generation awesome hip hosting platform that was going to you know basically blow ec2 out of the water and dominate the market and they were very very big on the whole Cloud sack thing I mean were you were you kind of aware of of that of ninefold back in the day yeah I think I I did heard uh the
name of the company and and there are many many such you know people springing up and experimenting with with the cloud as a domain yeah and I think it's not to harp on on the negatives but that didn't end up really well because there was a bit of a bug that they had I know if forget it the bug was that um they lost some storage as is usually the way and then there was a bit of a bug in the in the cloud stack stack which meant that any VMS that were turned off couldn't be turned on and VMS that were on couldn't
be turned off and basically for I remember about two or three weeks they were in this state of of that locked state which means that customers obviously couldn't turn VMS on they didn't want to basically lose data so it wasn't great for the company fast forward that company ninefold um ended up folding because of a few different challenges um but from what I saw from that the cloud stack you know got better because of that problem right and now the way that I've seen the trajectory go over the past few years it's certainly a
very trusted platform so it's safe to say that the early days there was just always going to be problems always going to be bugs it's a nature of what we do in this world right but I'm really interested to hear how it's evolved since then and obviously now it's very mature and has got lots of good names as well that actually leverage it you mentioned Telos but I know quite a few Cloud providers that actually leverage it at the same time and obviously shap blue who I've had on the show I had
Giles on the show a couple of years ago now Forefront of basically making sure that cloudstack is in a lot of service provider community so it's just an interesting history so it's it's really cool so talk a little bit about um you know what happened what was the years after that so 2016 right um you know and and also the significance of it becoming an Apache software Foundation license model yeah absolutely so what the license and the you know uh or almost like transition or you can say migration
of the code base to the ASF uh dead was it it invited a lot of people you know earlier it used to be CX who would you know make all the design decisions implementation releases and what whatnot so between almost 2012 and 2015 2016 we had a period where Citrix was the dominant sort of vendor you know they were dominant in terms of the development they will sponsor there'll be they are the sponsors of the conferences meetups and all those kind of activities you would expect and a lot of customers would of course use end up
CX Cloud platform they interested in that but this phase uh invited a lot of companies like shablo like sh phis the Le web and there are other hosting companies you know companies both from the Enterprises space but also from the cloud services space right so uh we saw a lot of that uptick there the ASF gave that sort of governance model which I think is is a testament to to be seen in a lot of the top level projects like Apache Cloud stack but also a web server right the Apache web server Hadoop and so on and so forth that governance model
gave this sort of structure where anybody can come uh you know collaborate and then was a Paths of you know starting from a contributor to becoming a committer which means getting code access to be able to change the code uh then of course to be able to do releases and and be part of the voting and governance model and join the project management committee which is tasked and responsible for the oversight on the project related matters like releases like doing meetups and conferences and so on and so forth so that can you
explain that a little bit can you explain that that and how that's so fundamentally different to how a normal company would approach that you know software development releas and support of the community right I think there are pros and cons to both the approaches uh what I've seen is because I was at Citrix I could see you know both side of the world and a lot of my colleagues including me would sometimes have a dilemma of that now we are doing a produ fork release for customers and we also have a community so where where is our
you know where should be our Focus yes uh so that that was a thing we struggled with citric struggled with a lot and uh but from a from the other things that you wanted me to explain and talk about uh the governance model allows anybody to come in and start as a contributor right so ASF uh I mean when you donate the project to the ASF we start with what we call a podling PMC a podling project management committee it's like an incubator almost where okay you transition your code base and you learn the processes the and you define the
bylaws so we started by defining the bylaws of the project like what are the laws under which this project must govern and the diff the key difference is instead of one entity I Citrix making all the decisions these decisions are then made by the PMC the project management committee where the PMC members are from different organizations so uh they can be from they were largely from Citrix during the incubation stage but then we saw a lot of people from different organization interprises Telos service providers those kind of
organization who had a uh uh you know skin in the game because they were users of the product to begin with uh and and that sort of governance model allowed that everyone has a say once they have proven their meritocracy their sort of commitment to the project by becoming committer and then eventually becoming a PMC member so that way one party Citrix wouldn't influence all the decisions you know those key decisions can be release related feature related but can also be fundamental to how you know where the me
tips are being arranged or conferences are being arranged so there's more sort of Tricks over the period of time had less say in how the project would function and that's what we saw between 2012 and 16 that's where Citrix was more involved and afterwards the cloud.com team kind of split they they went on to I mean some of them went on to do Rancher some of them you know other places pedig yeah good pedigree in fact uh uh um and that's when a lot of the other people people um could you know get in
their way of doing things and because it's more democratic that way so that's where we saw sha blue sh phis Le web and others kind of you know becoming more uh able to to support the project in a more sort of community oriented fashion which means not just the releases but also meetups uh uh uh the features the road map where they want you know what features they want in those kind of things that's cool yeah so so clad stack itself so it's self service portal web UI it's got CLI it's got arrest based
API seamless support for lots of hypervisors kubernetes container as a service um it's got hooks into infrastructures code terraform and anable it's it's it's such a like sort of utilitarian sort of platform so if you were to explain what cloudstack is what is it right so the the simplest definition is it's an open Source infrastructure of the service cloud computing uh platform and uh to kind of go a bit detail is uh it starts with the what the project stands from an ASF point of view so the
Apache software the Apache cloudstack Project's mission is to develop maintain and release uh infrastructural service software uh that's that's one of the missions you know when you donate something you have to Define what's the mission of this project the PMC members and what have you so that's our mission now how does that translate in a day-to-day basis it's a platform it's like a virtualization technology layer that supports like you mentioned multiple hypervisors so we support uh a
broad we have Broad and deep support for various hypervisors uh the three top of which are the KVM VM very 6i with v Center and of course xcp and Zen server which later on was also called Citrix hypervisor along with that we have a a broad support for storage Technologies so some of them are sort of uh you know default something like SEF and NFS and local storage but also managed storages like net app and store pool and lens store um Pure Storage and HP primary You Name It We we got support for that and now we also support object storage with
minio we support uh various other Integrations and features that you can like you mentioned uh we got multiple support for cubes engine one is the cloud stack community service that's our own inbuilt feature you get but we also support capsy which is cluster API provider for cloud stack which is then used by by a commercial offering like eksa so Amazon's eksa supports Cloud stack through capsy then like you mentioned we support terraform and anible multicloud uh uh libraries and Technologies like jcloud and what have you so we have very
large ecosystem of of software and um what are the benefits so I've defined you know explained what cloud stack is and what the components are but who are the users and where who do we benefit so we do largely benefit anybody who wants to build a private is cloud and they can then have their own use cases you can build a public Cloud uh we have a lot of regional and tier2 providers who using Cloud stack like leas web and PC extreme in the past and so on and so forth who you know build their own public Cloud
infrastructure based on GL stack and we have a lot of Enterprises we have a lot of Telos service providers manag service providers who use cloud Stacks that's basically large Lely what the user basis is like and then there are a lot of people lot of companies um AT&T and NT data is good examples who use cloud stack almost as a virtualization layer so for them it's abstracting their infrastructure their data center for them and then they use which is V director yeah which is what V cly director is an abstraction tool yeah
would do exactly uh and some of the very large companies I mean I can't name them because I do work with them uh at at work and they get upset if I call their names out but these are two of the world's largest companies by Revenue so two largest companies in the world they all users of cloud I think easy enough to look that up then people know who we're talking about potentially yeah absolutely in fact in fact if people come I mean anybody go to the YouTube and come to meet this and conferences they can see who's coming to give a talk
and you can make sense who are involved in that Community but that spes to the point that I mentioned earlier about you know the the early days and the roughness of of of what you go through in the early days of a project and of a platform and you know you could go either one of two ways and the journey of cloudstack is that yeah it went through cloud.
com it went through um you know the Citrix days the aache days to where it is now and to your point now it's being picked up and used broadly by top top top Enterprises and I think that's a huge tick and I think some of the advantages that I can tell and I know from from the class perspective is that yeah the multi-layered approach to to the hypervisor just gives Choice um that's huge and I remember that being a big thing back in the day and if I was looking to build or start an infrastructure company today if I was crazy enough to do that you know I'd
want the choice of multiple hypervisors today because look what's happening today in the world right so this is where cloudstack fundamentally has has pretty big advantage and I'm not too sure that the broader community knows this you know what I mean in terms of they cuz you can lump open stack open shift Cloud stack open nebula um all of K you know all of these get lumped together but they are very different right I mean so yeah there's a good question I get often and that we talk about and that I know is around what's
the difference between open stack and Cloud stack right so what's what's differentiating between gloud stack and other options uh I mean I would always say that open stack and open nebula they have their own strengths and weaknesses so I'm I mean I'm still an engineer first so my uh advice to all these kind of debates is go and try you know this is open source it's only going to cost you some time if you up for that or or you can look at videos on you know go and try everything and see what works
for you so even though I've been involved and I'm deeply passionate about CLA stack as a project I would say it's the best thing you should do but don't take my word for it go and try it open NOA open stack Cloud stack all open source you just have to try now to answer a question what is the differentiator the biggest thing is cloud stack is a turnkey product right it's not like a jigsaw puzzle you need to put together it's like a monolith which is important some people's yeah which is important and it's it's
strength right because you install and it's just there you you want to upgrade there's just one group of services components you need to upgrade it's not like a jigaw puzzle where you have to run around you need a very large team you need months and months of Maintenance uh or or planning to prepare for an upgrade and those kind of things so apart from the fact you know the things you mentioned multic Cloud multi-hypervisor support multiple storage support uh monolith it's a good product I mean uh if you're a very large
user who wants a very tight integration and flexibility maybe this is not the stack for you right maybe open stack and open neula might be but if you're looking for something that just works it's just boring because is um bear in mind we are in 2024 we're not in 2010 and 2012 now so this is not the layer to be in right you want this to be as much boring as possible because you want to you know play with AI or or or containers and something else you don't want to worry about your infrastructure
as a service that sort of virtualization layer so those are the strengths of cloud stack um it's got the apach license so you have an open- Source software that you can take you can modify you can redistribute if you want you don't have to share with others um it's vendor agnostics so compared to some of the other Stacks you mentioned it's highly the aparture governance model makes it super when diagnostic right thing you know look at things what have happened in the VMware ecosystem yes which I'm sure you'll probably have
some questions about um that's something that's not the first time it has happened and I don't think that's the last time it will happen so you have to also look at apart from the choice of software platform open source gave you that advant right it's free as a product you don't have to pay for it it helps you lower cost of ownership but even among open source software you have to see what is the operational expenses of you know running your infrastructure with that yeah the total cost of ownership
right so uh we did some studies with uh you know people in the open source ecosystem and we found that after 3 years of of of um you know capex uh third year or fourth year onwards the operational expenses with lot stag are dramatically huge and people um uh uh get lot of you know uh uh Edge over the competitors if they use cloud stack versus other stack so that's like a data point for you and we are also seeing that as part so you you've given an example of Cloud9 and maybe some other companies uh one of the key things I me
I've seen is you have to have an open source strategy but you also have to see whether that's sustainable for you like choosing one stack versus another how does that ban out over a period of time when uh let's say the software is free but then you're paying somehow right you have a small team of people or do you have a very large team of people who manage and operate that upgrade that maintain that I think those are the key advantages with Cloud stack so total you lower the total cost of ownership but
also you have something which is just really boring it just works yeah and I think that's a good thing to have right and and I think from the point of view of the seamless integration the turnkey solution you're you're kind of getting that from an open source perspective you know for free as such but obviously nothing's ever just free so you got to work that out but then under the surface you still got VMware or you can do the KVM or whatnot but everything is controlled the multi tency the high
availability the networking the multi- tency it's all controlled by the cloud stack layer and by the way quite a pretty interface these days as well like the interface is really really like I think it's it's ahead of the game in terms of what it looks like so there's obviously been a bit of a an emphasis on the team to make sure that the UI is is good the control plan's pretty because that makes a big difference not only from an operational point of view but also what it means for the end us and
the tenants you know if you're an MSP you want to give your tenant an i offering that's looking good and easy to managed through a portal and I think cloud STX got that um it's it's and it is interesting because underneath the surface you're still paying for ESX you're still paying for you know xcp so there's there's all sorts of different levels of um conversations that people need to understand or have because it's a multi-layered approach to the costing whereas I think some of the other other
solutions they more sell themselves as an allinone you know you get the hyperv VIS you get the control plan you get everything put into it so I think that's something that people need to sort of consider when they're looking at cloudstack that that top layer platform that overlays onto everything else absolutely and and that's why and um I I I say that you know a lot of users see Cloud stack as almost an abstraction layer so you don't want to know or care what's underneath and you also want that flexibility to take one
component out and replace with another so a lot of uh users and in fact you know at work customers they like this approach that CL stack supports VMware and also KVM uh within the same availability Zone almost and in the newer version of cloud we have features where within the same deployment if you have multiple hypervisors you can migrate VM from VMware to K I was going to ask that question I was going to ask that next question yeah yeah but but that's to say you know that those feature sets and that capability is Possible only because
cloudstack has understanding and support for multiple hypervisors so implementing a feature like how do I migrate a v from VMware to km is even easier because your control plane has a unified uh you know pane of glass or abstraction layer for upper layers so if if people are let's say they have they want to change in their strategy they want to move away from esxi or the VM world to something else this uh model the whole architecture allows you to do something like that yeah and and talk about the importance of kubernetes and containers
in in the success of cloud stack has that been something that you've been able to Leverage to build on for especially Telos and organizations that are looking to you know integrate kubernetes and containerization more because I think from what I've read you you can do kubernetes or you can do container containerized platforms just individually right right so I think couple of years ago we started with the agnostic container as a service uh support which we call CK I mean CCS Cloud stack container service but then
back in the day um you know there was the form misos was there cuties and there was a fight right and cities emerged as a winner out of that we pivoted our CCF implementation with cks so we have a native support for CLA cuity service and since then what we have seen over the last couple of years people are some of the users are using clo stack simply because we support we have a community support in fact we have multiple support for different engine through capc or eksa or our own cks uh from functionality and then um people
are also using anable and terraform so a terraform user who's using a different Cloud wants to they look for alternative they stumble upon hey we have Cloud stack so it's the interesting thing I'm seeing with the Telos pH the Enterprise and the service providers phas where people are decided to use CLA stack uh uh not because it's a fantastic is platform which by the way it is but because it supports it's an excellent support for these cubes engines where they want to you know they want to see
their sort of go to market or their uses usage going off that's that's important because kubernetes kubernetes as a service from my perspective hasn't really been cracked yet outside of what the public clouds are offering it's either you're using you're doing open shift on Prem um or you're leveraging an azour an AWS or a Google you know kubernetes Service as such or Microsoft whatever it might be um so to have something that you can take and cons consum as a platform consume as a set of apis that you can
plug into across multiple different kubernetes platforms and containerization implementations the same way that you were abstracting the VMware and the VM hypervisor between VW Zen KVM whatever it might be if cloudstack can offer that same level of portability and just a single pain of glass to go I hate saying single pain of glass but I just said it because effectively that's what it would be like that's going to be where you're finding success there yeah absolutely in fact um there's a lot of uh opportunities once you can
run uh containers with confidence and at scale on an is layer like Cloud stack then you can do a lot of things your app development goes on top of it you can do database as a service you can uh do GPU as a service or container so we have a lot of those use cases um you know driving a lot of the adoption is that okay so in terms of that so GPU is that separate to can you guys offer that sort of service now are you looking at offering the GPU as a service from the AI perspective for private models like what's what's very quickly what's there
what's what's in that world at the moment right so so so two things really uh I mean one is a point of sort of data for you which is uh one of our top search queries on our Cloud stack documentation is is cuini right it's like in the top three so there you have it and then lot of these service providers in fact in the last Cloud stack collaboration conference two of them actually gave a talk about how they're using Ai and uh GPU workloads with Cloud stack and there's an upcoming feature which will come in future but
people have figured out how they can orchestrate Cloud stack where they can do GPU pass through either through a VM or through a cks cluster where their containers can leverage that so once your cuetes cluster or your virtual machines can you know use these gpus then it's like a layer above us like What's happen happens in the VM or the container is not our business per se but we happy that CL stack you know has the feature set the capability through API orchestration at scale people can do that um one public example I can give
you is there's a talk by AT&T where they've shown how they do giops orchestration and you know containers at scale with Cloud yeah it's all very exciting world I mean I think about it 10 15 15 years ago now it's getting on but this was this is almost what we wanted back then but we've got it today you know it's taken it's a journey it's been a long journey to get to a point where everyone is cons the promise of what these platforms offered at the start yeah you know at the start it was
really just VMS right yeah bit bit abstracting storage abstracting networking pretty basic but it's taken the best part of that 15 years to come to a position today where it's just really exciting that you can abstract everything and everything that's new and building on the building blocks of virtualization of VMS but now we're adding the GPU component we're adding you know kubernetes and clustering componentry that then we're able to build on top of that we're able to do you know large
language models and gener of AI and and and GPU intensive tasks just through this sort of platform and consuming it in the way that we knew we were going to be consuming it 15 years ago but we couldn't back then because the technology wasn't all there it's super exciting yeah yeah yeah absolutely so so finish off let's talk about um the VMware broadcom scenario because obviously this year it's been quite dominated by that since the since it closed and obviously I've had you know companies that you guys have
support like I've had KVM on here you mentioned a number of different storage platforms like minio and what not like I've had them on the show but specifically when it comes to the broadcom scenario how have you seen that change of landscape and how has cloudstack been able to number one take advantage of it and on top of that how's it going to continue to take advantage of that moving forward yeah so um I mean first of all I would say there's huge interest in Cloud stri right the number of people coming to conference and
meetups asking questions I mean if you look at the Project's sort of History I did uh State of the Union talk in the last conference where I saw in the last almost 10 years the number of developers have become constant so we are in a what I would say a plat of productivity where there there's almost like 10 new users every week you know there's more number of GitHub issues being logged in people who are you know who can actually fix at that sort of speed and run rate um so there's a lot of interest from different
people who are trying out Cloud stag you know you see uh questions on the mailing list about installing Cloud stag so you know and you know questions around cuties and whatnot so that means people are interested in Cloud stack and I would say big chunk of them are relooking at their own sort of uh you know is strategy whether they want to stick continue with VMware they want to look at something else and a majority of them who uh couldn't afford the let's say the new price hikes by by broadcom they are looking into open source
Alternatives so they are stumbling upon Cloud stack and the way we are supporting that sort of user base is looking into what are the features we can do to you know support them in their infrastructure so I would broadly speak of two or three features one of them is what we call uh VMware injection so people who want to continue to use vmw a6i but they have lost other tool Ling functionality we have feature in GL stack now where you can like uh you know uh connect and import like absor their existing infrastructure so you install
stack and you just inest their current in VM infrastructure and that gives them that sort of UI that tooling and other capability that they didn't had either earlier or you know people who lost uh being able to use vcloud or something else you know they can now use cloud stag with their existing VMware infrastructure the second feature is of course migration where they can like stop their VMS uh on the VMware infrastructure and with a click off button they can import those VMS from VMware esxi to a cavm based host uh so
that seems to be a very popular choice with people people who are um reconsidering VM exide the whole ecosystem in the first place they are adopting VMware sorry they are adopting KVM and maybe other even storage Technologies like uh SEF and and look storage NFS and so on and so forth so they're looking into that and the third feature is um you know a bunch of things around the software defined networking so we are looking into we have an upcoming feature where uh clack can support NSX but also we are looking into people who will lose that
those kind of functionality how do they migrate to Cloud stack so having ability to have yeah having ability to do bgp and uh you know dynamic in static routing through cloudstack Virtual Router and to be able to have you know other feature and functionality so we're looking into all those things so with the current uh release that's 419 these early two features I mentioned they're already out there people are testing and using them we have got Improvement requests and we have working really hard
to you know improve those two functionality a to be able to ingest VMware in class Stag and also to be able to allow migration of infrastructure from VMware to km and moving forward we'll look into investing more time in basically listening to the community what they want uh some of them may sponsor uh those features they may contribute bug fixes or even implement the features themselves so we looking for all those kind of collaboration and moving forward the way I see it is uh there'll be still a big chunk of VM
users who are going to migrate to a lot of these stacks and I'm optimistic that many of them will end up choosing Cloud stack some of them I work with at work uh uh I can't say in detail but yeah there are a lot of large companies and users who are using these and leveraging these features yeah that's it sounds great like I think the future of pachy cloudstack is is is set in stone because of the everything you've talked about rer I think it just sets you guys up for the ability to win the market which I
think is scramming out for just that Simplicity because there is a lot of complication out there right and I think what cloudstack has over the others number one the ability to do multiple hypervisors I think it's a massive tick given where we are today and offer that abstraction layer thanks Roy hit for being on the show just as a final reminder if you do love great things with great Tech would like to feature in future episodes please click on the link on the show notes or go to jtwj t.
com and register your interest please tell your friends tell your family about the show it's a good show hopefully you know if you're not a tech geek you get something out of it if you are a tech geek in the industry you get a lot out of this we tired all together this has been episode 82 of great things with great Tech with Apache cloud