 I think it's time to start. So welcome to my talk. If you are just here because of the title fail, it's quite short. It is about failure in bigger companies and open stack projects, mostly because of the companies are big. So I would like to know if someone on you is currently evaluating or planning to evaluating open stack in a larger company. So if you could raise your hand, that would be nice. Oh, it's a couple of. Oh, perfect. Cool. Anyone else already has some failed projects in big companies in the past, like we tried open stack and didn't succeed? A few. Okay. Maybe you will find the reasons again in this talk. Before we start, I would like to introduce myself a little bit because not many people may know me. My name is Sven Michels, German. Company is called Zektor, which is based in Germany in Frankfurt. So that is something like the internet city in Germany because many of the big telcos come together in Frankfurt. We are doing IT business for more than 15 years and open stack for more than three years right now. So you may have seen me in past summits somewhere in all over the world. Oh, okay. I'm counting on my widespread knowledge because I'm doing a lot of things. So just to count a couple of things. I'm familiar with project management, training, I'm familiar with OSS training, DevOps and stuff like that. And I'm also into network storage and stuff. So it's more like you can ask me mainly everything, which is good because in open stack, all these come together and in past companies you had always silos like these guys are doing network, these guys are doing ops, these guys are doing security. And the problem was always like they need to talk each other and they don't like to talk to each other because it means work. So this comes together and I was used as a kind of proxy between teams in the past and now in open stack. This is quite easy because all comes together and all has to work together. And I'm happy to be one of the first open stack mentors. So in case you plan to be a mentee, you can contact me. Otherwise, if you're interested in the open stack mentor program, you can talk to me also. Yeah. So let's start with the talk. Oh, no, forget something because that one is something I added last minute after I got the details of the survey. I'm not sure if any one of you have participated in the survey. One, two, three, oh, three. Okay. This is one of the reasons I would like to talk about that one because there are a lot of people contributing to that survey, but sadly only 3% from Germany, which is from my side a little bit depressing because I know there's a lot of open stack ongoing in Germany, not that much as here probably or in other regions, but still more than 3% I would suggest. So if you are able to participate to those surveys, do it because it's important for the project for open stack and for all the people involved in that because your feedback is used in future deployments and stuff like that. And there's another thing I would like to add because I have not this failure on my list anymore because I thought it was already gone. But there's a interesting number in that sheet and that is the third from the below. It's saving money, which is one of the top business drivers why people start open stack projects. And the survey showed that if it's important and two-thirds of the people said it is important, then it's top one on this list. So whenever an open stack project starts, the top business driver for doing open stack is cost reduction. It is nice and open stack probably will save costs, but putting that in front of everything else is probably something like doing it wrong because saving money is a wonderful outcome when you have an open stack project running and your company is adopting to it because it will save money in long short. But for looking at like in half a year or a year, you will probably end up in noticing that it will only waste money because you need to educate your people, you need to maybe buy new hardware and stuff like that. And if you just focus on that one, you will probably have a bad time. So just try to stay away from that as top one for your reasons to go open stack because open stack can do much much more than just saving money for you. Okay, so why is it this talk about failure? Because I think everyone in the business know that failure is part of your way to success. You need to make mistakes to learn from that, you need to see where your borders are, which are your problems and you probably don't know about them before. So failure is important, but I would suggest not to repeat history. In this kind of history is not as like 30 years ago, it's more like three years ago, but what I have seen in companies is that they repeat the errors other ones did two years ago, three years ago over and over again. Because they think, oh, we are different. We can do it the same way they did because they didn't succeed. That means not we will not succeed. But most of the time, the failures are so generic that all will suffer from them. So you could just stay away from that and you would probably succeed much quicker than when you stop, buy and hit all of those. So let's start with the first one on the list. And the first one is the one which I was called for like nearly 90% of the time. Companies have something like a virtualization platform or even not virtualized. It's more like bare metal anymore and they would like to move on and they say, okay, we just replace the platform. So what they do is we have something, we replace that one by open stack and then we will have our success. You can try that. I will not stop you from that because then you need me again. But yeah, it probably won't work. It's not probably it will not work because open stack is just more than just a drop in replacement for your platforms or maybe for your processes. That's another one coming up later. It is something which is quite complex and you need a lot of homework to do to understand how open stack works and how you need to adopt to be able to use open stack in a way where it can really powers you because most of the time when you just migrate from one platform to open stack without changing anything, you will come to a point where you're like needing upgrades. Everyone has heard about upgrades in open stack. Have someone did a successful upgrade in open stack on a running platform? Three. So many of them will come to the point where they try to upgrade open stack and maybe it doesn't work but they don't have a plan B for that because yeah well if you do open stack in a from my perspective correct way you will use this as a cloud service and then you should rely on that the cloud can fail at some stages. And if your application is able to handle that one you will have a good time. If you don't then you probably will end up in a disaster. So if you start to plan open stack in your company you need to adopt so you need to change your processes you need even to change your people a little because they need to work different than they have before and if you stay like you are today and you just move on to open stack because that's oh everyone moves to open stack so we go then you will probably end up in losing stuff you already have achieved in the past because it simply won't work on open stack like it was before and if you don't adopt you will end up in having nothing. So this is something people think that often as easy because hey it's just another platform move on and we will find ways to do it but if you have ever used open stack for the first time if you remember that one it's not by just clicking a couple of links and it works and there's a lot of documentation to read on so plan ahead and start changing before moving to open stack. Number two also something I heard often it's like why couldn't open stack be more like I don't call names now but you know plenty of virtualization platforms from different vendors and they have something nice over here and you would like to keep that one and open stack doesn't have that one and then you start asking people from the open stack foundation or the developers why is it different why isn't it like I used it like five years ago it was so easy with the other product and yeah I don't like to adopt that one because that was a good one and yeah but the reason you go to open stack is probably because open stack provides you with something your old product or your other product doesn't support so there's reasons for that and most of the reasons are good reasons so in case of just bugging people making open stack more like something rare or some other platforms try to find the good points on open stack and adopt to them because those are the things that makes open stack powerful like you have all the apis I'm not aware about any other platform that have that much apis and possibilities from the beginning so this is something which is a a good point on adopting because people are often complaining like oh there's no UI for doing so and so because that's a complex task to do yeah but you have an API and you can work on that API and you can develop your own dashboard or connectors or whatever and that will enable you to do much powerful things than you would be when you rely on just what the vendor provides to you number three we're getting quick so this is something I was also hearing a lot from companies because they start open stack projects and they started behind the curtain because they don't want to be to yeah let's say people are interested in companies doing open stack and they won't see what they do with open stack and they would see the move to open stack so they stay behind the curtain with open stack and start their own project with a couple of people involved and everything works nice and then you come to problems and when it comes to problems the best way to do it is to ask people because whatever you think you're doing with open stack I would bet that most of the stuff has already be done by other people there's nothing special and in this time people start to hiding and trying to solve them by themselves and they waste a lot of time and people by fixing stuff other ones already have solved and this is something which i learned the hardware on open stack summits because when you come first to a summit and you start noticing that you can talk actually to people which is great you will probably notice that you can also talk to people working in big companies there was talks by paypal for example and they were happily to share their experience with all the other guys so i would count paypal as a big company and they probably do stuff which is confidential in some ways but at least what what's on open stack side they are happily to share their experience and so if you ever think about doing your open stack projects behind the curtains you can't do that it's no problem but think about that you probably will end up in doing all the work twice or even more times to fix your stuff instead of just going to a summit meet up something like that talk to the people and get some feedback maybe get the right pointers to fix your problems or issues or even you think about doing something very big with open stack nobody has done this before because you don't find it on the internet and then you end up in coming to a summit and three other companies just did the same so if you work openly with open stack it's probably in my point of view the best way to do it because then you can easily adapt to the upstream releases you can easily adapt to new changes on all the projects and you can easily integrate other stuff in your open stack projects so try to exchange with others try to integrate other ones in your thinking and project because that one is really a key driver in all this business if you take a look how many companies do quick releases and stuff open stack still hurts big companies with their six month release cycle because they say oh six months release cycle that's ridiculous we can't do that because we usually do something like a year maybe more if you stay the same way for next two years you will probably end up in running open stack in an in a version which is not only not supported anymore which is more counted like a dinosaur and doesn't provide anything you need today so try to move in the same speed like open stack does or at least try to adjust your speed a little bit more than just a couple of years number four also a thing which is quite common people think that if you run your product or something like that on open stack instead of traditional hosting platforms it will be a success because you can do all the fancy open stack stuff yes in some cases this might be correct but most of the time it's not only about the platform you use and open stack is is a step to your success but probably not the one bringing you up to the top of your path to success so when you plan about running something on open stack because that one is the only problem you have good on you but most of the time you will see that if you move on to open stack they will raise new problems new issues and new possibilities maybe which makes you even which enables you even to be more successful with your application with your platforms with whatever you plan to use with open stack so don't rely on just open stack open stack is just part of your solution never ever is open stack the only solution and if you say it but company number x did it the same way and they were successful that might be correct but it's company x not yours so don't just think that success from other ones can be copied over to your companies or copied over to your project because you don't see the whole effort behind a project or platform and you don't see what's what happened on the way to success from them so you probably will only see the top of their success and the rest is also very important and if you skip those parts you will have a hard time later on so let's come to number five this is something widespread because open stack is often seen as an enabler so if you have traditional developing teams in your company they always often complain about oh we would like having some self-service platform to be enabled to do DevOps stuff DevOps is something everyone likes today if you don't listen to Gardner but an open stack platform is not the only thing you need to be able to do DevOps or agile development or something like that because this is only the platform and a platform itself basically does nothing for you it just gives you the right starting point so if you rely on open stack as an enabler that's okay but the enabling itself is done by you or more important your people so if you want to make sure that open stack enables you to do something then make sure that the company is also changing or adopting depending on what plans you have because this one is yeah it's it's like if you start a new project and you do it the same way and you you have seen in the past that some stuff doesn't work like you don't get releases quickly or you don't get the feedback from the customer just replacing your old stuff by open stack or an open stack enabled platform probably won't change anything of those problems you can just maybe develop quicker but if you don't get the feedback what is the profit of it and something which is also very important if you work with people in your company and i bet you do so don't assume your people are stupid if you assume your people are stupid then ask yourself first why you hired them and secondly if you tell them like we go daffops we merge teams everyone will work like before just better and everything everything will work smoother and everything will be nice uh i don't know anyone who is believing in that bullshit sorry yeah it is just bullshit because what is behind that is like oh we have 30 people working together now then we have 60 and then they count on numbers we don't need 60 people doing that work 30 is enough maybe even less because we do more together and we work more efficient and people will notice these type of changes it's not about providing them a better working environment it's a company companies look on numbers and successes and so companies are moving on in this direction to shape their own employee count a little bit and if people notice that one they will stop working like you would expect it to do they will not like this daffop stuff anymore they will try probably to sabotage your daffop stuff because they know what it will end up with maybe in another company so if you think your people are stupid they probably aren't and hopefully they aren't and it will not work that way because you will have a lot of people working against you in your own company this is not only bad this is the worst you can have so um yeah what are possible steps to success in this case because we heard now what what's probably not working but you will be hopefully hearing what is working so be open minded is just like open stack just as a platform or just as a yeah we just run a bunch of instances on it it's nice uh to to think about it but there's much much more to do you can do much more than just like running instances on it you can run sdns with it you can run like services on demand on it in in in scale and in in a in a way that it was not possible before and if you start thinking about what is possible with open stack you will probably end up in getting new ideas or new projects new applications whatever you can imagine to run into your own company instead of just buying something from amazon maybe um communication and exchange is also very important because um this is where all the knowledge comes from it may be that you know people already from the foundation or maybe not or from other projects or other companies and at least for the summits here in in the open stack area i have never seen that a company denies on exchanging with some other company even if they are working on the same uh platforms issues whatever um normally you will find nice people over here to exchange with them and i would recommend to do so because it's the best thing here you come together with like people from 60 other countries and from hundreds of companies why not using those four or five days to exchange to those people it is quicker and better than whatever you can do otherwise no email no no phone call will be ever that good like meeting people in person over here so go to summits if you have projects in your companies enable people to go to the summits this is key important i was talking to someone yesterday uh they started an open project couple of weeks ago not much more than a than a month and they send up two people over here because they so see it's important to meet other people to exchange um i have seen companies they denied people to go here not here in particular but to summits uh because it's just a waste of time it's one week conference nobody needs that you can look up all the stuff on youtube yes you can but you will not be able to communicate on youtube you can probably post comments but most of the people won't read them at least not the ones you want to reach so enable your people to go to the summits second second thing is we have mailing lists for open stack projects they are also very useful because you have a lot of interesting and intelligent people on those lists and a mailing list is is still old-fashioned but it works we have archives you can look up and on on google and stuff you will find answers over there and you will get answers over there so open communication is the key to success in my eyes next thing is what i personally like is to go to meetups i'm not sure how many meetups you have here around the area or in your home cities but at least in germany we have a couple of them and i can say at least the ones i have attended to they are useful and nice because you met a lot of other people interested in this project and working in this area even from big companies again and they are a little bit smaller than a conference like that where you have seven and a half thousand people we probably it's hard to get to the right ones but if you meet local people that's also a good starting point to exchange to grow your own community to grow your own knowledge and in case everything else fails you can still talk to the foundation the foundation is nothing like a another big company you may not talk to them they are happy to answer your questions they can link you to other ones working in the same area they can link you to the developers they can link you to companies working with the same issues and stuff like that so don't forget about the foundation that's one of the best things we can have an open stack other point is to ask vendors vendors are usually around here like in the marketplace and of course they want to make business everyone wants to make business but they are also happy to help you with your projects i have worked in the past with a lot of vendors and most of them are here right now and i know they are nice people i can promise that they will help you with your projects they will help you with support they will help you with getting in contact with developers on their side so if you have issues they can probably solve that and one thing i have seen in the past was like we requested something and it took two weeks to get a full blown update for a solution from a vendor and two weeks for an open stack release is amazing so don't think there's just sales guys wanting your money they are nice people behind them and they are happy to help you and of course you can ask experts some of them are working around here and if you don't find one you can ask me because i'm most of the time nice and i'm also happy to help you and to connect you to other ones because i personally like the philosophy behind open stack and the way it works and i would love to see this project ongoing for a lot of more years maybe even better than it is today so get support from experts because that's the easiest way to get knowledge into your company i mean growing knowledge in your company is probably hard but at least what open stack provides you with all the documentation it's a little bit easier to do so than with other projects i have been involved with because it's already there you have all the documentation and you have all the stuff you need so you can do your own research and with the right pointers you will quickly be able to do stuff you probably never have imagined before and when you are successful with your open stack project and you are able to do please contribute because many stuff is yeah just keep into the lockers because they say oh this is not important that's probably not right because as i said in one of the failures before many people work on the same issues and if you keep your stuff behind the curtain nobody will profit from it and profit means not probably money but you get recognized and people which get recognized usually get also help so put it on your list in your open stack projects to plan ahead for contribution to open stack it must not be code in some way it could also be like you fix documentation people from the documentation team i'm always happy to get fixed for their own documentation or if you have ideas you can still just support the community with your ideas because some people will probably adopt them and make them possible for you so any questions and comments and please no eggs or tomatoes to me but if you have cookies it's okay i don't see that much because the lights are so bright but i don't see anyone having comments and questions so any no okay so i did and anything wrong okay one final quote this is my personal one because i like to tell people uh technical stuff in the way they probably will understand it and transfer to other ones because people just if you talk about clouds people still look at the sky this will store my data no if you go with open stack and think about another as a car think about open stack is an electrical car one not the traditional ones you have with your gas in it it's a new car with new possibilities but it will only be your new car if your needs and habits fit to that car like if you start going on a trip and you fill up your gas tank like two minutes before you will have a bad time doing that with an electrical car one because in two minutes you will probably not even charge a person so you need to change your habits and you need to be uh you have to need sorry you need to have the right needs to fit to that car so if you go open stack check first if that one is the right way to do it and if you see issues take a research about the issues you probably can solve them with additional projects or additional vendors um before you start moving on because otherwise as said you will probably end up with a bad time so thanks if you have any questions afterwards you can contact me by email or on twitter or if you meet me somewhere around this area just talk to me i'm happy to help talk and have a good time over here so that's it now you can plow it again