 Hello and welcome everyone. This is a brand new episode of Open Infra live. I hope you are all having a great day. Well as for myself I'm still waking up sipping on my mango flavored coffee and yes you heard me right I said mango and coffee together in a really short sentence and it's actually amazing but this is not the only interesting thing that you will learn today. My name is Eliko Vance. I will be your host for today's episode and despite of all my efforts the best angle still had the light on top of my head so I feel like I'm your angel for this episode so I both do my best to act that way. Before we jump into today's content I would like to just welcome those of you who are new to the show. Open Infra live is an interactive series that is brought to you by the Open Infrastructure Foundation. We are coming to you every Thursday at 1400 UTC mark your calendars live and we are bringing you production case studies, open source demos, conversations with industry experts and just updates from the global open infrastructure community. So I mentioned that this is an interactive show which means that it is very important that we have you to participate if you're watching us live so please use the chat of the platform and tool that you chose to follow us on today and type in your questions, comments so we can have you on the show as well. And if this would not be enough for you and you like to have the spotlight then please collect your ideas and submit your story to ideas.openinfra.dev so we can have you on for one of our upcoming shows. And now that I got the housekeeping items out of the way I have to tell you that I'm really excited for today's episode. I myself started to contribute to OpenStack eight years ago. This number is getting so high I can't believe. But this is such an amazing and welcoming community that this eight years just felt like eight days. And I'm so happy to be part of this community that I've been also participating in a lot of onboarding activities. We have an upstream training that I will be talking about just a little bit towards the end of this episode. And whenever I do the intro I always start with reminding people that you know OpenStack is not just a collection of suffering components but it is also an open source community. And today is the day when I get to turn this sentence upside down and remind all of you that this is not just an open source community but it is also an open source cloud platform. And then you may ask what is cloud exactly? So for today's episode we have Ben Silverman with us who will share his thoughts about cloud and he also has a lot of experience with OpenStack, how to deploy it, how the architecture looks like and just really what it is all about and how you can build a cloud platform out of all the services that OpenStack contains as a big box of Legos or as a big package. So he will help you understand all that. Before I give him the word I will just again remind you to put your questions and comments in the chat and I'm also turning to those of you in the audience who might already have experience with OpenStack to also please help us answer questions in the chat and make the chat even more lively for this episode because we have a lot of amazing content to go through so we would like to ensure that we have all the questions answered one way or the other. So with that I'm turning to you Ben to please tell me what is cloud and why do I need one or well as many as I can. Thank you Odika and thank you everyone else for joining the session. It's been about two years since I've given a summit presentation or any sort of presentation on behalf of the Open Infrastructure Foundation. Doesn't mean I've been absent I have just been kind of redirected so I couldn't be more happy to come back and deliver a presentation for them. You know some of you may know me from from YouTube you're one of the I don't know 70 something thousand people who who watch my videos on YouTube from the you know the annals of the Open Infra YouTube archive but if you if you haven't I'll let you know the session's been focused on on those of you who aren't really clear on you know on on on OpenStack what is it how did we get there how did you just sprout up in the desert or you know or how any any how anyone even arrived at OpenStack where it came from or how it works there's you know there's really no prerequisites for this session other than some basic IT knowledge and the desire to learn. So this has always been my least favorite slide I don't like talking about myself so I'll make this short and sweet. I started my OpenStack journey way back in 2013 as I was wrapping up a data center migration project for American Express. I was looking for my next big assignment and one of our senior executives asked me if he if I'd like to help him architect an OpenStack cloud and he was good he said it was going to be the next target platform for all the company's applications. Well of course if a senior executive comes up to you I said yes and they walked away think about 10 steps and then I wonder what is OpenStack what did I just agree to no problem I thought I just Google it and I'll watch some YouTube videos and I'll have the thing up at the end of the week not a problem so off I went and you know I thought here's you know type it into Google and I typed in Google and it came up with some very strange results I thought well maybe it's just the way I'm asking the question so I went over to you know went over to YouTube and I saw the the lack of of instructional videos there there are a lot of videos so you know I started asking myself things like you know you know where do I even get OpenStack do I go down the staples and get a CD with OpenStack on it because that's the way I thought I did and I figured I'd learn from YouTube you know what OpenStack was how to install from the CD you know so I get from staples and then how to customize it for the financial industry requirements you know at the time you know as well as working for America Express they had all sorts of policies and procedures and and and requirements security requirements but you know everybody had guidance on that right so there had to have been guidance out there on OpenStack well you know that wasn't very easy I went into YouTube and I remember seeing a lot of videos when people saying you know especially the videos from Rackspace saying hey we've decided to join with NASA's Nebula project and we're creating this new platform called OpenStack and it's going to have these following components and and I thought what what do you mean you're creating it but you've got it it's already created right it's on the CD down the staples and I spent the next couple months watching every YouTube video that was out there that mentioned OpenStack that even talked about OpenStack and I realized that I wasn't very much further down the line after a couple months you know the if you're done reading you know these these words about me let me show you a picture what I look like after I was reading you know watching all those videos right there right there that's what I contracted that's what the doctor said I had OpenStack overload syndrome that's right I was diagnosed with the OpenStack overload syndrome it was horrible my eyes were stuck like this for weeks I watched every video on YouTube about OpenStack and I learned you know lots of things about how people were excited you know to be involved with OpenStack lots of discussions on how it would be created gap analysis how it theoretically be set up and lots of technical information but very little on how to how to install it so guess what I didn't learn I didn't learn how to install it where to get it or which state which staples had the CD luckily when my eyes returned to normal my company decided that they'd send me and a team to go to the next OpenStack summit during that summit I attended a lot of sessions most of which I didn't understand nor did they help me a whole lot but what they did do is put me in touch with a lot of people who were in the same boat as I was and they may have been in the boat a year before or two years before when when OpenStack first started in 2010 and they helped me they helped me a lot they helped me they point me to some some resources they pointed me to some installers that were being created by hand and I even met some vendors that were creating installers for OpenStack at the time and helped us move that project forward but you know it was at that moment was at that summit but I vowed to never let anyone else suffer from from OpenStack Overload Syndrome and once I learned OpenStack I began doing these beginner sessions about the why, where and how of OpenStack so today's agenda is going to be based basically really is what is cloud we have to start from the beginning how did how did OpenStack even get here without some other things happening OpenStack would even exist then we'll talk about what is OpenStack without you know without definition it's just sounds like two words that they threw together how does it work how to get involved and then we'll have a little session on Q&A what is cloud so in order to start our journey we need to understand why cloud computing is even a thing today let's go back to something like the early 2000s yes I realized this is probably before some of you were even born maybe you're just a baby but I was working in a data center where we simply had the job of racking stacking and imaging servers with the latest OSs and helping the applications teams load lamp stacks as web 2.0 kicked in the developers took more expanded role in middleware and as dynamic web applications started become available if you're old enough you might be able to remember when web applications started to actually do something within browser at you know Apple applications or interactions with social media think WordPress Facebook Twitter and blog sites what powered these applications was simply an infrastructure with an arbitrary vertical scaling algorithm of hardware each time a trigger a triggered event happened like the load on a server hit 70 percent this would trigger a series of manual tasks and workflows and that's what we kick off a more capacity this is great for system administrators who kept us in jobs middleware engineers data center folk but really bad for efficiency and scaling so what was so bad you asked well these servers were ordered by the dozens and they were placing data centers and racked and stacked and powered on just waiting to be loaded with images and software and middleware when they reached a certain trigger capacity new servers were ordered and later they arrived in a cycle where pizza soft it was about a 55 day cycle all the time today could you even imagine if an infrastructure team tells would tell a developer development team they have to wait 55 days for new space new server when they came up with an idea I can't I can't even imagine I don't even know what would happen it'd be chaos it would be there would be cats you know holy hands with dogs it would be insane so you know when we think about this scaling and this efficiency you also have to think about what happens at Christmas right when orders push the servers at these data centers to 99 capacity on Black Friday right they were sitting at 40% and then in one day they all go to 99% there's not 55 days between Black Friday and Christmas so how would you scale up well the the answer is you would have to scale up way before so when those dozens of servers might have become hundreds of servers and then after the Christmas holiday you couldn't ship them back to Dell or HP couldn't put them on eBay to sell them and just temporarily you recoup some of your money those servers held on and they became expansive expansion expansion hosts and those expansion hosts they would get old you know time you know time is money and new technology comes out so in six months when they were adopted and put in the processors were old sockets were old everything was old so there had to be a better way so in 2006 Amazon had run into the same scaling and efficiency problems they realized some of their you know some of its hardware was mostly idle off peak we're talking 10-20% off peak well while their other systems went overloaded at peak times but they had to make sure that availability was there like everyone else for their retail customer for their book buyers yes that was back when Amazon just sold books with that in mind Amazon set out to create a way to give development teams infrastructure and development teams and infrastructure admins a way to deploy infrastructure in a self-service way this is how AWS was started it allowed Amazon to abstract all those wasted resources those idle resources that are happening in in off times yet scale up on demand when they needed to it allowed them later to resell their own resources during off peak which increased efficiency exponentially and then the rest is history additionally another company had a different idea on and again it wasn't called cloud at the time it was just ideas on how to become more efficient and scale scale efficiently company called Salesforce and maybe you've heard of them I think they're still around had a different idea on how to make themselves more efficient they decided to enable their software to be run via web browser and have customers use their infrastructure to operate by bringing customers to them they controlled the scale were able to fully utilize the infrastructure and didn't have to deal with customer software upgrades or any of the on-premise legacy problems of customers running software applications because they managed and ran it themselves of course it's hard to keep a good thing a secret these two companies inspire a lot of new ideas about what was to become cloud computing so we talked about the lengthy acquisition cycle of new resources for app and dev teams in the early web 2.0 days days and days of approval workflows hardware provisioning cabling network connection requests and we talked to I talked about the 55 day provisioning cycle and and I saw that in my experience 55 days was not uncommon I didn't even mention what happens if they ran out of blue cables or green cables or red cables or any of the cables that they used to connect into the racks that had to be of a certain type or color those if those were back ordered or if they had some sort of problem with those or even if there was a logistical issue it could push that provisioning period to 100 days I mean we're seeing today with customers with certain vendors things coming out of China even in the modern day where we have you know we have the cloud concept we're seeing delays for memory we're seeing delays for chassis we're seeing all sorts of delays and when we have to deliver bad news about a 10 day delay they're they're up in arms but it wasn't that many years ago that you know a 10 day delay was really nothing it was absorbed into the the 55 day 75 day cycle of getting hardware today 10 day delay is almost unthinkable it's it's it's almost an emergency so we can see the difference between the two different times so you know we talked about this lengthy acquisition cycle during the you know getting resources for the app dev teams in web 2.0 days and days of approval workflows and then we experienced those as well we would have to you know back then we were we're showing age again we were using Lotus Notes and we had notes forms and we would have to submit them to leadership and they'd have to be signed off on and of course Lotus Notes wasn't wasn't the greatest software in the world to begin with sorry guys but it wasn't but in order to go from manager to manager to manager to manager to sign off on each piece of the infrastructure we had these huge workflows approval workflows that were would take for forever so we had the hardware provisioning the cable in the network connection requests there was never enough ports or your your request was had an SLA of 15 days and you know it seemed like oh they would always use 15 days so and again these these are common things within companies but it was all to just get infrastructure for whatever reason let's say if for reasons that for for this discussion it's to get space to develop a new app maybe something cutting edge something that's fresh some blockchain or some i iot app that that is really hot on something that's going on you know in the community or in the in it you know we're we're talking 55 60 days what if your idea was so fresh that if in those 55 60 days you lose your competitive advantage waiting for infrastructure to be delivered to you if you don't get if you don't get to the market first with an idea anymore your last and i don't be ricky bobby right if you're not first your last in a lot of these cases if you're not first your last so you know during this time people were saying there's got to be a better way it's got to be a better way and there were tools that were coming out for other vendors of other software to allow them to do self-service but it was never cloud enabled it never api enabled it was never programmatic it was never infrastructure as code it was always just ways to shave off workflows or do workflows a little faster i think when we when we had enabled some things like that it took us from 55 days to 51 days which i consider negligible so it's not when we use the speed of the process is the is the oh it's what's the outcome of the goal not the output the goal is to increase efficiency and to deliver infrastructure so that we can be first to market that's the outcome not the output you know how many vms you can create in an hour or how many vms you can create end to end with the fewest amount of workflows it is the output goal so in this slide you know you can see here's the value of cloud computing is in the outcomes it enables not unlike the outcome of an elliptical trainer the value is the outcome building cardiovascular health or losing weight versus how many times your legs move back and forth right we don't get on an elliptical to see how many times our legs can move back and forth right we don't get on there to make ourselves better at ellipticals um maybe maybe some of you do i don't know i i don't um i don't get on there to you know become better at elliptical using right it's that's the that's the output right my arms are moving back and forth but what what is the outcome that i'm looking for greater cardiovascular health weight loss muscle toning and the same thing we we look at with cloud computing um you know we we we see the definition here that it has to be scalable have shared services service-based reachable via the internet these wrongness definitions great those are each have output goals but it's the outcomes that cloud would need to achieve in order to be successful so we talked about the elliptical trainer but now imagine elliptical uh with an arm and a leg based api so instead of spending all that time getting dressed going to the gym waiting in line for equipment waiting for the sweaty guy to get off and then wipe it down for you and then spray with those chemicals that don't really kill anything you could just plug your body into the api the elliptical api dial up a thousand uh repetitions and you just get skinnier and stronger just standing there right because you have programmatic access to the elliptical you don't have to flail your arms and legs around you just plug in to the api or set of api's and and you know instantly you're smaller um fortunately that doesn't exist um and if any of you have any ideas for that or feel free to write to me um i i would love to get in on that project but this is sort of like the magic that cloud brings to programmable infrastructure so in 2020 moving forward um idg reported that 81 percent of companies they surveyed planned on increasing cloud adoption and 19 percent of companies they surveyed were planning on completing a total cloud transformation in which they were completely uh moving to cloud infrastructure for their company open stack is part of that story but in order to see where fit we have to talk about the two different types of clouds so we went from a hardware based and workflow based infrastructure provisioning process to a programmatic process where it's driven by by api's or driven by a web interface to create infrastructure but that's not the whole that's not the whole story when we talk about cloud it's not just like we talk about ice cream right when we say go get me some ice cream what's the first thing you ask well what's the flavor what flavor do you want now of course it's you know you want gelato you want you know yogurt there's a lot more dynamic to it but we we ask the flavors well cloud is not just cloud cloud has different modalities and different types what you know what we talk about here with the types is first private cloud and private cloud is where uh open stack comes in to the picture it's it's dedicated in a many many times it's dedicated to a single user can be hosted in a can be a hosted private cloud in a customer customer or vendor's data center or as a remotely managed cloud but it's private to the the enterprise it's private to the company it can be hosted but private not to say that again that that it can't be one of the other modes but this is where it primarily lives public cloud is a shared resource um it with a pay as you go metered billing mode usually so these are your aws's your azures your gcps your rack space now now that i said that that open stack is mainly a private cloud or start off as mainly a private cloud platform um rack space has turned it into a public cloud platform platform so they use open stack underneath their for a public cloud platform that being said that brings them into these other two modalities of multi cloud which is using multiple clouds like aws and azure or open stack and aws or open stack and azure to spread an application across multiple clouds so there's no cross pollination or across api traffic between the two clouds but the the apps live in those clouds and for reasons of follow the sun or ha or disaster recovery the app is on multiple clouds and that's multi cloud hybrid cloud and again you may hear different definitions of this as you go through i used to have in my presentations definitions from all the different vendors and you could see really how disparate they were but these are kind of the generally understood definitions hybrid cloud is a mix of of private cloud and public cloud or it could be two public clouds mixed together however the workloads are spread across the different clouds so you may have your web and app tier on one cloud and your database tier on another cloud so you may have your web and app in in amazon but it reaches back to your private cloud let's say open stack cloud on premise to get its its source of record to get its data so those are the four main types of of modalities of cloud within those clouds there's service models so these are the service models are what you hear about when you talk about provisioning something or the waste the way it operates the five current modalities that everyone's usually talking about our infrastructure as a service container as a service platform as a service function platform or function as a service and then software as a service and if you look at the slide that i just put up you'll see that there is a private portion of it private cloud portion at the very top that shows examples of private cloud parts of these different service models for example ias well for short email for infrastructure service examples of that will be open stack vSphere azure stack i can't think of any others there and then the public version of that would be in the box with the dotted line below that's listed hosted and that would be aws ec2 google compute azure vm's things like that same thing with the container platform as a service you have kubernetes you have open stack zoom you have open stack magnum as a management service and then in the public cloud you have gke ecs and acs again if we look through that look down through all of these you'll see examples of all of them you can even have a sass platform private sass platform where it's hosted on your private cloud with open stack and a byo is to bring your own so there's really i mean there's a there's an unlimited amount of sass offerings that are out there but they're usually custom to the to the particular cloud but you can see that the open stack logo is there and has a big bracket across all of them open stack services all five of those whereas you know some of your some of your other options here come from multiple different vendors but open stack can handle all of these on its own or with third party vendors so open stack really is a total solution for not only modalities but for the models here as a service model it is a complete solution if if you were confused by the as a service models i like to use this slide i think it's it's as funny as entertaining but it's pizza as a service and it really illustrates um but what we mean by as a service so for example we would start it at on premise right on premise is akin to the the old data center model where um you know you have configurations functions scaling runtime os virtualization and hardware uh here we have a conversation friends beer pizza fire oven and then your electric and gas you know as all the things that you would need for your your your application stack you know all the way down into the into the hardware um so for example like on premise would be you know like a homemade pizza made through scratch you do everything yourself no no change you know no changes so far right this is the old way we did everything we bought the machines we we plugged them in we powered them up we cable them we put the os on them we put the middleware stack on them we loaded our applications and we did everything ourselves there was no provider that helped us out with that whether it was a private cloud provider or whether it was um you know a public cloud um and then we get to infrastructure as a service well things start to change right so now we're given infrastructure so it's akin to like you share your share your kitchen with other people the utilities and oven are provided but you make you cook the pizza yourself kind of like ec2 kind of amazon's ec2 or um you know even even open stack vm's um you know so the kitchen would be akin to your or infrastructure your virtualization your hardware your infrastructure then you get to containers as a service you bring the pizzas but someone else uses their uses the facilities and cooks the pizza for you so you bring your app you bring your your maybe even bring your os um and then you hand it to somebody and you say hey here's my here's all the ingredients go ahead and make it for me and some some chef sets it all up and gives you your containers or your pizza platform as a service you order the pizza for takeout pizza rea makes and cooks the pizza using their facilities and then you just come pick it up um and again all you have to bring is the conversation the friends and the beer right so the pizza is already made for you but you still have some things that you have to do um and again which would be application uh you know different functions uh scaling these are all things that you still need to worry about and platform as a service function as a service again your responsibilities decrease and the provider's responsibilities increase so function as a service you go to a pizza rea with some friends you order and then eat the pizza made by the restaurant you order drinks from the bar and they're made for you you know and then you have those you have those drinks so now you're not supplying the beer it's it's a you know craft beer where you order mixed drinks and they make it for you now all you do is you have to bring friends and bring the conversation that's it um and again that would be akin to you know bringing your functions and you know whatever whatever they're going to do whatever the functions are going to do other configuration of the functions and then software as a service you just you just uh you go to someone's house for party they provide you with the pizza they you know they invite others to come talk you know and hang out with you um and the only thing you have to do is talk you just show up and talk so with software as a service you just go to the site connect to it run the software and then when you're done you log out you're done you don't have to clean up you don't have to make pizzas you don't have to serve drinks you have to do anything everything is done for you just like the um the sales force for if you think maybe even gmail gmail is a good software as a service example right you don't have a mail server you don't have anything to worry about but just going and reading your mail maybe that helps you maybe that's confused you even more before we move forward I would just jump in a little um I would like to say that I really love uh all this background I have personally a much better understanding of why I need cloud and what kind of solutions are out there and and how those different terms are relating to each other and what they mean I also want pizza now so I need to go and look for a breakfast pizza option somewhere um I also love the uh really lively chat uh that we have on the different platforms people are talking about deployment tools and options hardware options things like smart nick support in open stack um CICD solutions and um there was also a question about you know where can you find question uh answers to your questions so I would just like to remind everyone that one of the best places to reach the community is the open stack discuss mailing list that you can subscribe to and if you have any questions that we did not answer today then you can always reach out to the community and find them there and now we are switching gears and looking into what open stack itself is uh how its architecture looks like and how it will deliver pizza to my front door back to you Ben. Thank you so what is open stack then right so we've talked about cloud we talked about the different modes the modalities what is open stack um and really just to be very simple you may already know this because you're here on this call but open stack is open source software for building clouds it's it is it's programmable infrastructure that allows users to have one platform for virtual machines containers and bare metal I'm going to say that again because that's very important right because you're not going to get that anywhere else for virtual machines containers and bare metal it's software that's been written by tens of thousands of individuals from hundreds of companies it's not a single product it is a set of projects that make up foundational services like compute networking storage but also covered areas like telemetry containers security and billing open stack is not virtualization it uses virtualization open stack doesn't manipulate the underlying os it uses tools native to the os to achieve that open stack uses all these elements to build clouds from modular architecture that you choose not someone chooses it for you open stack builds clouds using apis to abstract physical resources you combine them into useful programmable infrastructure so open stack free open source software I'm not going to spend a lot of time on this slide I would encourage you to go out to the open stack.org site and and really look at these four open principles but one of the things that makes it great and I want to mention it is the choice the choice you have it's open source you can get involved open design people you know their people can put in their own opinions they can open development and it's an open community that helps everyone that they share knowledge and they have open discussions about the future of the platform and that is what makes it great so the you know based on the the open stack user survey it's been run since 2013 people people have placed their feedback as to what is most important why use why do they use open stack right we talked about operational efficiency of cloud and now that we're talking about open stack as part of that domain 94% of users said cost was their first business driver people do not like to be locked into vendor managed or vendor vendor platforms they they find that pay for enterprise features and then not getting exactly what they want and then having to pay for more add-ons and add-ons and add-ons and add-ons it's just not operationally efficient and that's what drove so many people and continues to drive people to open stack to accelerate innovation we talked about that being first if you're not first you're last in this market if you can't get your infrastructure ready for your developers to develop and push out into production you're not going to win and again avoid or avoid vendor locked in we don't want to be locked in you know we don't want to be locked into someone else's idea of what a platform should be and and you know again some of these then I said it and I say it again and I'm not going to pick on one vendor but certain vendors you send them requests for enhancement and I'll tell you where it goes it goes right in the garbage they don't even look at it but with open stack you can put in a feature request you can design a feature and it will be discussed and you will be part of it and others will be part of it and and if it's a if it's a good idea and the the community recognizes it it will be implemented so really who uses open stack I don't know maybe some of these companies actually all of these companies I mean there's gotta be there's gotta be a company out here you've heard of or seen before 15 million cores so those the people who are who say open stack is dead although they've been saying open stack is dead since 2010 if open stack is if open stack died as many times as they say it's died then it holds a record for dying and coming back to life because it just keeps growing it keeps expanding it keeps reaching out the new people and again this is proof positive these these these logos themselves show that it's not going anywhere so what really runs on it right so it's out there people have it right well maybe it's just one sector no it's not they've got lots of different sectors again if you go to the open stack or in site and go to the software section or the use cases you'll see that there's tons of them out there just not just telco not just banking not just big data it's it's really really diverse we're 10 years 10 years in it's hard to believe you know it's really it's kind of technically technically we're 11 years in but you know they have some of these milestones here that just blow me away as I said I've been involved since 2013 20 2012 2013 and I've seen things happen with open stack that I never dreamed that would happen and I just it's still still going strong still still have milestones to reach who you know who is um who is who is who is open stack right what is open stack open stack is people people are teams the open infrastructure foundation helps people form teams to solve problems and create solutions right that's what that's what they're there for I look at this picture and I can recognize a ton of people in this picture and it's amazing I can because I've met a lot of people you know with with open stack and this is just a small smattering of of the people and you know 111,000 people it's just it's just mine it's mind-boggling of how many people have touched the code have looked over the code have found bugs uh to corrected bugs it's it's really it's really a an amazing network open stack releases open stack releases are have always been on time sometimes the times change but when they say they're going to release it they release it and it's about every six months um back to back in 2019 in october it was it was train usury victoria wallaby and the next one is is a release going to be released in october and that's zina um I I always laugh I'm not going to do it today but uh I always when I do it when I do this live um on you know in a summit I'll always challenge myself to to go through every release see how fast I can go through it it's a little early today so I'm not going to do it but um these the names are are interesting uh they allow you to kind of they're unique you're not going to forget them although um newton I always call null because that's what I voted for and I really thought that was going to win until I had some discussions with some programmers and they said you realize what would happen if we named the release null and you know there were lookups against the release and it was coming up with null I said yeah yeah you're probably right so I'm glad that one didn't win but anyway um it's uh that's the release cadence you don't have to upgrade every time uh you you upgrade basically when you feel that that the feature set um is is appropriate for you to upgrade but it's backwards compatible to at least n minus two so they release minus two um and then the apis are backwards compatible so you're you're not going to for the for the most part they're backwards compatible there are some things that get deprecated uh in in new releases but they're well documented all right so we've made it to the important part how does it work right so we've got this cloud operating system or this cloud platform um so how does it work remember when I talked to about before it's a common misunderstanding uh that open stack is a single product open stack is a group is a open stack is a group of open source projects and which projects you need will depend on how you want to use open stack while there's a large list of projects you can plug into your open stack cloud not all of them are necessary um you know for all use cases and all projects have reached maturity uh with the service that they provide these projects are still included uh under the open stack umbrella all of the ones that that have reached maturity and because they work towards the open stack mission and some users want you know some some users want access to um the the projects that used to be what they call an incubation or projects that may have lost uh support those are all still there uh they they're there to be picked up again to have a team work on them but there there are projects for almost everything you could possibly think of when you explore open stack projects you'll you'll find several tags that'll help you understand where this project is and what projects are right for you on the foundation site uh open stack.org you'll find a list of sample configurations these sample configurations are combinations of services for different use cases uh kind of informed by the case studies of real-world users in the sample configurations i encourage you to look at a sample called compute starter kit the the services uh they built for this use case um are the foundation uh supported by the optional services around them and they will help you start a cloud uh the cloud that you need not a cloud that you know has everything right you don't you don't want to bolt on a bunch of stuff that you don't need um so you can look at that look at the compute starter services there's the compute starter kit and you will see that includes glance nova keystone and neutron we'll talk about those in a little bit but not all the use cases require all four services but you need the services relevant to your use case and that is a good starting point that will for 98 percent of the use cases that we see that is a good starting point so remember that when you go out to the site this is the open stack landscape these are the the projects that are active right now and uh i won't i won't harp on this we i mean i could spend all day going through each one of these projects and talking about what they are and what they do but um this this diagram is available for you there's explanations on open stack.org to tell you about each one of these projects we'll go over to the main ones today but they just know that there's a whole world out there um that does all sorts of things from um building business logic to performance testing and benchmarking and those are these are all things that you can plug plug into your open stack cloud so as i said i'm going to go through the uh the most common services or the most common projects um and this is the list of them that we'll be going through today i'm gonna i'm gonna have i'm gonna say the names a bunch of times so that if you're not familiar with uh the actual names because that was one of the hardest things for me right somebody said hey you know when when you know nobo reaches out to to glance and what nobo who glance what are you talking about like you usually you just say it and what it is and image service compute you know no no no we we use the the names right the cutesy names and i thought that was that's pretty dumb but then once you use them for a while you start to kind of get a feel for it right with how how they interact and then it starts to make more sense but i'm going to keep saying the names so that you're more familiar with them we'll go over them right here the dashboard the gooey that you interact with is is horizon um that uh is it drives all these drives all the api requests into these other services um the identity services keystone compute is nova block storage is cinder networking is neutron image service is glance and the object storage is swift so now this is that's the first time you've heard them that's the first time i've said them we'll move on so nova compute nova compute you know i don't really have a good pneumonic pneumonic way to remember nova compute it used to be called nebula that's really irrelevant today but nova in spanish is no go that's why the chevy nova never sold many cars in in the lat latino world because would you buy a car that said no go uh they actually renamed it to the diablo which happens to be a uh open stack release i don't know where i'm going with this but this is really going badly um so nova you're just gonna have to remember nova is compute um i remember there was a public certain public broadcasting show called nova um and it was like a big deal for me so i always thought well nova's a big deal nova's compute it actually implements services and associated libraries to provide uh on-demand self-service access to the compute uh resources including bare metal virtual machines and containers it also provides a way to um provision ec2 uh resources uh utilize utilizes the rest api and is driven by messaging um and uses the placement service as well there we go that's the next one cinder okay here's one that i can actually give you a give you a pneumonic device for right cinder think cinder block um i'm sorry if english isn't your first language um that may not translate well for you but cinder blocks are large concrete blocks that a lot of the buildings are made of um and i always thought you know storage cinder block storage facility made of cinder blocks cinder block so the open stack block storage uh service block storage cinder block provides persistent block storage for compute instances so it provides persistent block storage it also uh can provide you secondary storage well i much like amazon's ebs or azure storage disk uh it can be used to create volume snapshots for bootable volumes that can be detached and and reattached uh through new instances uh use or uses the backup volume um a lot of the backup software that's third party actually uses cinder neutron okay here's another one so jimmy neutron jimmy neutron was a uh cartoon character with uh when i was a young when i was a young kid used to fly around in a spaceship and go bouncing around all over so it reminded me of network packets uh or think of neutrinos something like that it's the network component it provides networking capability for open stack it helps ensure the components uh can can talk to each other um and it's it's software defined networking for for open stack so before for net neutron there was quantum uh and nova networks there are ways to to and they implement kind of a flat network nova network but anything you read with that is kind of this it's not really relevant anymore um and uh so just think neutron neutrinos bouncing around network um and it just basically standardizes and abstracts the networking from the physical layer and allows you to to concentrate on building um programmable infrastructure glance glance picture in your mind kind of like glanced over there at that there's a picture in my mind um image image picture image image service um glance holds the images for your your your your compute so um again rest a table it's uh used for discovering and registering and retrieving vms it stores images and metadata about the um about about your images your am i as if you will if you're if you're at aws um and it supports so many different formats uh including am i uh q cal 2 vmware vmdk uh raw you name it swift swift is the object storage um you know it's not very swift i think that's what i you know come to learn about it over time it's eventual consistency so it's you know it's it's not instant consistency it kind of takes its time to get through especially if you have uh swift spread across uh large distances your swift nodes it kind of moves slowly um so it's uh i don't know that i don't know how us to help you remember that and again this is one of the early services and i don't know why they call it swift if anybody knows that um that would be great to know um but it's it's a highly available distributed even you know eventually consistent object blob store um think of it's it's it's it's fully s3 compatible so if you want to think of object storage um this is what it is it's object storage it's it's um you know you know everybody i think a lot of people are very familiar with object storage nowadays so uh it's the same technology that Dropbox used for many years um you know for you know many many enterprise storage clouds of the Fortune 500 so again object storage is swift swift swift swift keystone identity right keystone key lock safe keystone identity key lock safe keystone um and you know it's a it's the it's i would say one of the most important um but it's not the most important project in in the open stack architecture it processes api requests provides identity tokens allows you to log in allows the services to talk to one another in a secure manner um and provides your service registry that registers all the services that are available in the cloud okay so for you aws users out there that are looking into open stack here is a slide for you i'm not going to go over it um but it's not an all inclusive list is this list could probably reset a stone for aws could go on for pages and pages and pages there's not 100 feature parity here but if if you're trying to memorize it based on what you know about aws it'll be it'll be a good help so here's the open stack simple conceptual architecture take a good look this is what it looked like to me when i first started but really this is a tokyo subway map but this is what it looked like when they started giving me diagrams when i first started open stack they look just like this i don't bring them up anymore the real ones anymore because um they're just more confusing than anything else but here how is that one better so this is the simple conceptual architecture and really this this is not complete but it really does a great job of saying this provides for that this provides for that this connects to this and and does what it this is what this does in relation to that um it's it's very very useful and we'll go into this a little bit more detail in a second this is the most important slide if you understand this then you will understand how open stack works at the very top you see your stateless services your rest api the request comes in the top and then it goes down all the way to the bottom and along the way um it sends it sends um messages out to the message queue rabid mq uh for for crud operations and for and have an information so that the next service or the next part of the service can pick it up um and for example like you know if you if you were talking sender you would have a request for a volume it would come in it would put it out on the message queue the scheduler placement service would pick it up make a placement decision on what storage that you have available whether it's a storage frame or saff or whatever it is say yes there's space available you put it back onto the message queue then it would go down into the service functionality it'd be pulled off the message queue and then at the same time it's being recorded in the stateful database the maria db and um you know placement decisions are being recorded in there and then there's other metadata that's recorded in the the database and through the service function then into the driver for whatever storage that you have and then onto the provider so if you have some sort of net app or something it'll talk right to the provider but all of this is taken care of for you by open stack you don't talk to the you don't actually talk to the providers open stack will do that for you with this service by abstracting all of the things at the bottom so you can just talk to the apis at the top and an example of this would be this would be a complete vm creation process right users login to the horizon dashboard and an initiate a vm create the identity down here um authorizes says okay to the dashboard then the the nova api initiates um provisioning and states the state remember what i said it saves the state to the db as it's going along the nova placement service finds the appropriate host of where it's supposed to go neutron then gets involved and configures networking then cinder gets involved and provides the block storage for the vm the image is location in glance is looked up through look through glance the image is retrieved out of the object store which is hosted by swift and then all this information comes together at the end and is rendered by the hypervisor whatever hypervisor you're using whether it's kvm or zen or ec2 whatever it is whatever the hypervisor you have because you have choices with hypervisor as well and now number 10 the vm is accessible by the users and that's how open stack works everything works just like this rest api goes through puts things on the message bus records them in a database moves through the different steps and then does the abstraction and creates the resources for you it's not magic if you understand this part of it even in a general sense you've got it down and i'm going to turn it back over to ildik oh i know we're running really late uh well just a tad bit late so i will be quick but would like to thank you for this amazing overview and now for me things are turning back to normal where i can remind people again that open stack is not just software but it's also a community so um you could already see that in the in the chat of the different platforms where where we had a really lively chat with questions coming in people some from the community who i know answered them and there was a really lively discussion i believe we don't even have any question that did not get answered at all um so this is what really the community is all about and you all who are watching this episode can be part of it and you may ask how so there are a couple of resources like on the next slide we do have an upstream training it's called upstream institute and it is actually teaching you how to become part of the community it talks about open stack as a community how it works how it's structured what kind of processes we have and how you can participate we have materials you can see the links on the slide so we do have a contributor guide that you can look up independently from the training and it will guide you through all these processes and tools that the community is using and we also have the training materials that will give you a little bit of guidance in terms of maybe what order you want to go through the contributor guide if you don't want to read it as a book we are having some online versions of the training due to the current circumstances so look out on the website we will have the new dates up once the next occasion is all settled down we will probably have one upcoming this year around mid november and once you're kind of have all your courage then you can go on the next slide a couple more links in terms of I already mentioned the open stack discuss mailing list that is a really good place to get in touch with the community like it is a good tool for asynchronous communication because the community spread all around the globe so if you're jumping on IRC maybe the person who you're looking for is sales sleeping because it's night where they are but you can always drop your question or idea to the open stack discuss mailing list I mentioned our foundation at the beginning of this presentation so you can also join the foundation it's probably a good time to also say thank you to all of all of our members because without their support we could not bring opening for a live to you and if you would like to join them and support the foundation then you can do that at opening for adult staff slash join and user survey for all of you who are either considering to run open stack or maybe you deployed it already and it's running in production we would love to learn about your plans and actions so please fill out the user survey to help us understand how open stack is used all around the globe what kind of use cases you're using it for because from Ben's presentation you can see like how cloud is applicable to our to our everyday lives and in what ways you can use the software that turns your bruh metal servers into a lively flexible manageable programmable infrastructure and there are so many ways to use it and I believe there was a question about the future of open stack so the user survey was the one way to bring you the answer just based on the learning how it is used around the globe and just in a bit I will tell you what we are coming to you with next week which also highlights some amazing use cases so again thank you Ben for for this amazing presentation and I would also like to thank everyone in our audience because the chat was really really lively today that was amazing to see and now just because we are over time let me remind people what an amazing episode we have for you for next week it is actually part of a mini series within the series it's called a large scale open stack that is brought to you by the large scale large scale SIG or special interest group within the open stack community and they are always talking about the ways of how open stack is run on a large scale all around the globe the challenges and solutions of that large-scale infrastructure how you deploy and manage that and with all the use cases next week's one will be super exciting because we will be talking about super computers so next week we will have Steak Telfer from StackHBC sort of Adam from Swiss National Supercomputer Center Happy Sidol from Center for High Performance Computing in South Africa Steve Quinnette from Monash e-research center and Jonathan Mills from NASA Goddard Space Flight Center and you can see that I have a little space shuttle in my ears so it will be a really exciting episode now that you know the basics of how open stack works you can go deploy and explore the software and next week we will tell you how you can build a super computer in your garage out of it and if you got inspired by this and you have a story to share please don't forget to submit your proposal to ideas.openinfra.dev and mark your calendars for next Thursday at 1400 UTC and be back here for the OpenStack Powering Supercomputers episode and with that thank you all for being here with us today and I hope that you will have a lovely rest of your day. Thank you everyone.