 Rwy'n gweithio gweithio'r gweithio i'r dyfodol i ddechrau'r cyfan. Rwy'n gweithio'r cyfan ddominion i ddechrau'r cyfan. Rwy'n gweithio'r bydwyr yn ddiweddol gael y Cyngor, a'r cyfnodd maesaf ymddangos. Felly yw Alun Fiol. Mae ardal yng Nghymru i Eirloedd Cymru. Rwy'n gweithio'r gweithio'r gweithio'r panel. Yn cael y cyffredinol, yw'r gwneud o blaenol, ac oedd yn dwylo'r gyffredinol yn fnwysig. Ac rwyf wedi gweld i'r dda i'r dweud yr adeiladau yma o'r yma i fynd i'w adeiladau sydd ar y dweud. A bydd wedi'u ddweud o ddweud o'r ddweud. Mae'n dweud, rwy'n rhaid. Felly, mae'r ddweud. Mae'r ddweud. Mae'n rhaid i'n rhaid i fynd i'r dweud, ond mae'n ddweud. Dyna'r ddweud o argynnu'r ddweud mae'n cael ei gweithio o'r cymryd ac yn ymgyrchol i'w arferrath. I've been working closely with OpenShift, but judging by the people in the room, I don't think I'd ever call myself an expert because we've got a lot of people, users, people who write the things, so, yeah. So, I think, well, we've basically heard directly from the financial services and customers. I think they talk about speed of the market, we talk about how quickly things have come about. Actually, this event in itself has come about very, very quickly and it's amazing to see such a turnout. So, on that note, we've got some fantastic guests with us, but obviously, public sector in terms of speed and agility in trying to get organisations to join us. We've actually taken it slightly on its head, so what we have here is four organisations that work with customers in the public sector. So, without further ado, we'll start at this end, if you'd like to introduce yourself. Hi, everyone. My name is James Curran. I'm an exile servant that worked at Department for Work and Pensions, Biz and Department for International Trade. I currently consult into government and we were using OpenShift at Department for International Trade and we're looking to expand it out across different departments, looking at how we could do some cross-government collaboration. Hi, I'm Bill New. I work for UK Cloud. We're possibly the largest supplier of cloud services into the UK public sector and we also serve the health sector as well. We have a multi-cloud strategy providing a whole selection of different cloud platforms, but I've been a keen collaborator with Red Hat, both in our OpenStack platform, but also in OpenShift as well, where we're seeking to work with government departments as they move services and work loads to the cloud. Justin Cook. I'm Justin Cook. I'm a distributed system specialist that worked on some projects with Department for International Trade, business innovation and skills and Department for Education. Hi, everybody. I'm Cal Busahi. I'm the CEO of a company called Dree that provides infrastructure consulting to governments in the Middle East, in Central America and the UK. I used to be a civil servant as well. I built something called the HMRC tax platform, which was a Microsoft services platform that started in 2011 and the 31st of January never fails to fill me with panic because I was the operations manager for that for two years. If you paid your tax today, shame on you. You should have done it weeks ago because there's people stressed out on the other side. To get the conversation going then, we're seeing digital disruption across all sectors and we know that public sector organisations would like to see themselves alongside their periods in their commercial counterparts. However, they do have their own restrictions and their own limitations due to the special nature of the departments. What are you seeing? What are your experiences of the customers that you're working with? One of the really beneficial things that we saw was cross-government collaboration. We were looking at how every department needs to have hosting of some sort and very much is insulated in one area. What we noticed was that other departments all wanted to do the same thing. They all wanted the opportunity to spin up development environments, that type of thing and potentially share it across different departments. So one of the things that we looked to share was our leverage between international trade and base and that gave us a good opportunity to see how a cross-government model might work. We supported the project that James has just been talking about and it was a great example of cross-departmental collaboration. Unfortunately, it was an unusual example of cross-departmental collaboration because much of Whitehall is incredibly siloed. If you look at overall cloud adoption in the public sector, it is still lagging the private sector. If you take a private sector at about 20% of workloads in the cloud, central government is probably about 10%. But then if you get into areas like local government or the health sector, it's down at 1% or 2%. Therefore, they have a long way to go and one of the things holding them back and it was called out by the Kings Fund recently in regard to the health sector is the fragmentation. I mean look at the number of trusts out there, look at the number of different local government organisations. There are collaborative bodies trying to bring things together but that sort of inter-departmental collaboration is unusual in central government. It's just almost nonexistent elsewhere. We need to be able to drive that and some of the technologies that we're talking about here and around containers and shared services and they are the way forward. It has to happen at some point but we're not quite there yet. Government silos are really good for innovation. It's also bad for innovation. Mindshare is the biggest problem that I encountered really with the change that we're seeing with Kubernetes and OpenShift. It's really difficult to see a rapid adoption across so many silos that have their own independence with the disruption that it causes to deliver. I'll decay some of the things that people said there about silos. I think the other thing that I may bring to the table is that government has been and still is I think in a large sense a really bad technology purchaser and some of this I think my own experience has been there being a low level of genuine technical skill within government departments and so actually the ability to evaluate and get value out of products has been pretty low and has been dependent on vendors and dependent on historically quite large vendors. There are areas and pockets where that's changing but the majority of cases in IT purchasing and technology purchasing is still driven quite far away from the people who have to deliver or operate or use products. That to me is quite a big challenge and I think that some of the seeding of departments with internal development capabilities has pushed now what works for some developers so departments that have got in internal development capabilities and developers are driving discussions about what kind of tooling there should be, what kind of platforms, what kind of things people should be using but that doesn't really take into account often the operation or need of a department and B is often driven by quite parochial concerns because government departments don't pay very well and find it hard to hire people who are invested in the long term outcomes that those departments are looking for. There's a reason that two of us are ex-civil servants on this panel. I imagine James has reason to be similar to mine and if we were invested for five, seven, ten years in making those technology decisions I think we'd see different outcomes to being invested for say a year or two. I said I'm lucky that I work with OpenShift a lot within my team and for a long time I found it really cool technology so I've considered myself lucky to play it. You guys have all had experience from various times, various views on this so what is the thing that drew you towards that from your experience? If we start this and with James and then work down again, give you a chance to have a ring? Sure. My role was more of a product manager and programme lead so what I saw the benefits were so Justin who was working in my team was spending a lot of his time having to update Kubernetes Docker and actually do that manually whereas the advantages that we saw was that you could use that in a more automated way and be able to use the deployments of OpenShift versions rather than having to, I wouldn't say waste our time but spend a lot of time having to maintain our systems rather than getting something automatically updated. From our perspective, we were always sought to serve the market and the clients that we operate in and we are purely focused on government sector and health and initially we had largely a VMware based environment which catered for the lift and shift of virtualised workloads very soon after that there was a call for OpenStack environment and we partnered with Red Hat for that and indeed we worked with these guys because they were one of the departments that demanded that of us and very rapidly after that it became apparent that again the clients that were some of the more forward thinking in central government were looking to move to Kubernetes and again we did an evaluation and went with Red Hat and OpenShift but this is what it's being called for but again we have to look at reality here this is only being called for by those that are really at the leading edge and really trying to push the envelope there are a vast majority of workloads in the public sector which are way behind this and there's an orientation around shared services and the government mindset about what shared services is is a very old world perspective and there's been a big debate in the press recently about how they deal with the large Oracle and SAP estates without really looking at where they need to be going with microservices there have been some notable projects and advances around key services like gov.uk pay and notify and the identity management platforms but these are isolated examples it's a very slow type but we're talking to some people here at the leading edge we were doing a green field project and that was really cool didn't have to worry about existing code bases the infrastructure convergence patterns are really attractive and they just make a lot of sense we were able to abstract away the platforms developers loved that their artifacts were immutable again loved that everything is well you have data and everything else is code so it all really made sense I actually ansibilized the Kubernetes rollout went to Australia for a couple of weeks came back they loved it that was cool over shift B3 was at 3.2 at that time it all just fell into place so I grabbed the ansible playbooks did some upstream commits and it worked and here we are I'm not 100% sold on OpenShift are you allowed to say that here is that okay I've built two homespun platforms of my own I'm not really sold on anything right now I feel like there's so much turbulence in the infrastructure space in the past I still am a very big fanboy for Heroku I don't know if people in the room are aware of Heroku but the idea like to me for a developer the development experience can sort of almost be magic and I feel like a lot of the implementations of platformy type stuff that I see falls very short of magic and we have the automation capability to make that happen it's almost always like organizational will or something called security or something like that that stops that from happening Heroku is the closest I've ever been to magic the thing I like about OpenShift is that it's kind of not hugely opinionated which is odd given that I care about the thing I really like is Heroku which is incredibly opinionated but when you're working in organizations that are very messy and have lots of different things going on it's almost useful to have something that's a bit more I can bring a service breaker that's a nice one I can plug and play a bit more than having to anticipate the exact needs of an organization and kind of like work out which technology is going to fit that I like the fact that I can use and as your Docker registry or the internal Docker registry or some other Docker registry depending on whatever my organization cares about and that's what I like about OpenShift but I still feel that most products in this area are falling short of like magic and I feel like there's more and more magic coming but it's not there yet Okay so in an effort to make my job easier then so my job is to push OpenShift to my customers if some of them have taken it they're fairly happy how would you sell OpenShift to see some of those benefits if you're not sold on it what are the bits that you would want to improve you want us to go back and look at? Everybody everywhere you go and we have a lot of debates about this but everyone everywhere you go wants Beth and the very idea of best practices some oasis in there, some chimera or something right but there's some kernel of truth in that and making it easier to consume patterns that other people have created in the past and identify them as patterns so the great thing about the open source community is there's lots of stuff but actually the idea that some of this stuff fits together to create an outcome somebody needs to do that curation of stuff that's my opinion and that's what I spend a lot of my time doing I pull a bit of helm, put a bit of this and then put together a thing not really any of my original engineering talent in it but I've seen a few pieces that solve a problem for someone because I've been around the block a bit the other thing that I would say that is coming up more and more and actually is driven more by the work that we're doing in the Middle Eastern in Mexico is the sort of question, the nascent question I guess some people are actually grappling with but we're more sort of playing with is can you containerise everything I'm very interested in the idea that operator experience is like a burden if I have to use 27 tools to do my day job that's like a burden I spent a lot of time running an operational team dealing with outages, dealing with postmortems and lots of things are, hey it's confusing there's 17 different ways of doing this there's 12 systems, I had the recent example of this who I alert where it was like those four buttons that were labored almost the same thing and I kind of feel like being able to consolidate around one user interface for operations there's some value in this and if that user interface could be something let's say let's call it OpenShift but through OpenShift I could run the stuff that I currently think I have to run using say Puppet or I have to run using some homespun bash scripts or something like that if I can consolidate that around one operator interface I feel like I can get benefits where things are currently falling between the cracks of lots of interesting useful bits of automation that haven't been consolidated I don't know how you make that a feature I'd be happy to talk about it more but you know So I think just before this Justin we had a conversation and you talk about OpenShift Kubernetes has been disruptive technologies so can you go into that how you've seen that be useful to you so you've had projects get off the ground based on these disruptive technologies how do you feel that's worked in the departments that you've worked with Overall I feel it was a bit of a struggle because there's a lot of... the silos are independent and to get them to adopt something so different comes down to the mindshare changes like for instance I prefer Debs I've been doing Debs or RPMs for 20 years everyone rolls the same I've been doing it so long that my scripts just do that for me why should I go ahead with this because I'm not going to be as forward thinking or it just may be that I'm working with a technology team or developer team that's going to take a lot of my time for me to research, develop train them, get them up to speed a lot of them just don't have the resources to handle with that I think it's been brought up a couple of times around the lack of resources the lack of will and I see OpenShift as a... those being as a cool technology is something that helps people want to actually work for a government department so in that way do you feel there's any value in trying to attract talent based on a technology let's give that to James because I haven't spoken for a while so government has different ways of doing things they throw a lot of people at things so you may find out that you could go to one department and there will be 20 web ops engineers where if you go to another one there might only be five or six and I think being able to adopt something like OpenShift could actually help change that culture so you could really find something where you have maybe a smaller web ops team but you can really specialise in something and be able to provide the opportunity to them to learn more about OpenShift and also look at saving the department money at the end of the day it's all taxpayers money so we should look at the most efficient way to do that is the most efficient way to look at something that's automatic you might have to pay licensing for it but you can do some more automated work with it or do you then decide to pay all of the web ops engineers that you might need to maintain a fully open source product that you don't need to buy licensing for I think that's a question that government really needs to tackle in my perspective we're talking about magic here there are a few magicians out there just in one of them and doing great magic but magic is that much easier to do in the right environment and greenfield environments are that much easier most of the different departments are still spending most of their time struggling with the legacy equipment and the cultural barriers in reality if you've got an environment we were talking earlier about the skills and the recruitment and retention of staff if you've got an environment where most of the type of sea level executives are career civil servants but the CIO is typically a two year rolling contract there is no great longevity there and the government has had real difficulty not only retaining those skills and providing the long term road map and the longevity and continuity in those environments but also having skills further into the departments because it's been overly reliant on a few of the major what people sometimes refer to as the oligopoly of large companies that always said look we'll just do it for you don't worry about the technology thankfully this is starting to change there are teams in DevOps within some of the departments that are actually making real progress here and we as one of the disruptive companies in this particular market are helping change things but it has a long way to go and I think many of the challenges that some of these guys have faced have been faced in very similar circumstances before we're constantly seeking to reinvent the wheel but constantly facing the same challenges because of the turnover and because of the in some cases the repetition of the same mistakes okay so I think actually we've covered a little bit of this anyway so we talked around then or we just touched on the cultural adoption do we think that is the largest barrier to entry for this is the cultural adoption and the different ways of working taking on containers and tools like open shift is that the key piece or do you think these organisations are able to overcome it or are there other factors so I think government's always afraid of lock-in so there's fairly or unfairly there's a reputation of being locked into certain suppliers for quite a long time I think they're afraid of getting locked into other services open shift provides that flexibility so you can deploy on to many hosting providers maybe we looked at AWS as a UK cloud so there's many options there I guess it's what the next barrier is do they then see open shift being locked into as well so you can only use that type of platform and actually trying to degrade it down is a lot difficult it would cost you a lot of time to be able to make that into a more of an open source product overall it's probably more of a cultural adoption that you need to look for as I mentioned previously there is a lot of we'll recruit a load of people and we'll try and do it as open as we can but is that the most cost effective way of doing it there could be other ways and that's obviously where we start talking about open shift and I think lock-in comes in many different forms with the large suppliers that have dominated this particular market for a long time lock-in has come in the form not only of technological lock-in to particular platforms but in contractual lock-in with monolithic contracts that went on for a number of different years I think a brexit is actually having a dual effect here in some ways a lot of organisations are actually putting off proper consideration of opening up these contracts because they're so obsessed with needing to deal with brexit that they're just going to continue rolling over some of their existing contracts rather than seeking to break out but in other ways brexit is also acting as a disruptor here because they have to do something different so it varies from department to department depending on what they're doing but eventually we need to reach a point where they are avoiding the lock-in and I think if we were looking for a buzzword in the last debate we would see in the government the buzzword being around multi-cloud where so many of the government departments are having to deal with legacy workloads they've got large Microsoft estates they've got large Oracle estates they want to do some stuff that's cloud-nated they want to be doing some green-filled stuff in containers you need that multi-cloud opportunity in order to do that and you need to do that without bumbling into another lock-in in the future Justin I don't really have anything to add sorry you can't stop me from talking the each as I pointed out each department is very different from each other department they have over time have emerged their own unique culture when I went to Nature Marcy people described it as Game of Thrones looking to build their own sort of little empire and then fight someone else and that was kind of how you made your name in that particular department everyone has their own ways of doing things but interestingly technology is often driven by what the overall purpose of that department is for the time and in HMRC if you've ever interacted with the tax system you'll know there's lots of forms so lots of things are about repeatable small services that had to just we had thousands of them so you just had to build a kind of a factory for building those kinds of repeatable small services and people as soon as you offered something that was going to solve that sort of problem people jumped on it very very quickly and it sounded very much like the previous panel the problems and the issues and the kinds of things we built sounded very similar to those things where I've worked in other departments they have a much more heterogeneous set of things that they do and offer to a more heterogeneous set of people and so there it's kind of very difficult just sort of from a technology perspective to bring along a single kind of approach to those things and people are sort of looking for something to attack those number of different problems but actually the character of those problems is pretty different so in Department for Education you have some things where people are trying to roll out things that will affect every every school child and also that will affect every school and also that will affect people who are applying to become teachers and things like that and each one of those constituencies each one of those user groups is quite different and has different kind of needs and often the technology that has gone in the past to support those different users is quite different and so the story is more the question in some ways the barrier is often like how do we get from where we are today to where we want to be when you start off with a bunch of different things that all have very different maybe different suppliers, maybe different contracts maybe different technology stacks those things have become even inside a single department become their own silos and it's very hard to even parse the question of how do I go from all of this different stuff to one thing especially in places where that as we were sort of saying before that kind of level of technology literacy and longevity of thinking is not always there and so those sorts of things they are cultural barriers but they're kind of it doesn't mean that there aren't like people who are kind of committed and invested trying to solve these things just actually some of these are quite challenging problems that have been built up over a long period of time okay thank you Diane I think we're about on time are we okay well I can say I think FY18 then year of the container 2018 on a challenge against Chris's the financial services will be 2019 before them so I've got a challenge to a couple of my customers who are in the audience that I think that we can start playing with this service mesh this year and get ahead of the financial services and we'll probably put a bet on with Chris at some point as to who gets that into production first so have a chat about that over be later but I'd like to thank everybody on the panel for taking part and hopefully being interesting for everybody here thank you so before Justin steps down I just really want to thank Justin Cook I don't know how many of you interacted with him over the past month that we organized this event but Justin was pretty key to getting this space and helping us get this off the ground so thank you very much Justin for all this