Subscribe
Share
Share
Embed
The Public Sector Executive Podcast is the new podcast bringing you closer to the public sector leaders in the UK. Covering everything from the environment to the economy to transport, our podcast will bring you the latest news, views and insight from the people responsible for shaping the country's future.
We need governments to ultimately be following this mindset. You know, if we want to be as a very kind of developed nation, kind of on the bleeding edge, we have to be using all of the data that we have. Ultimately could optimize tens of thousands of government jobs to make them more innovative rather than, you know, doing quite mundane activities.
Let's not bury our head in the sand, right? Legacy is not going to go away, the problem is going to get bigger. So we've got to tackle head on. This is the public sector executive podcast bringing you views, insight and conversation from leaders across the public sector, presented by Dan M Benn.
Today I'm joined by Martin Dolce and Dan Wesley from Oracle UK. Martin is sales director for the uk public sector, central government and defence, whilst also being the UK regional director for the Northwest. Dan is cloud engineering director for Oracle and we're going to be talking about cloud data and digital. I just want to kick things off by talking about a hugely successful partner case study that Oracle has with regards to data. Dan, would you tell me a little bit more about that please?
Today's world is all about data, right? And ultimately data is used in things like machine learning models, but also for these artificial intelligence models as well. And if we look today at some of the, where we're helping customers, one of the best examples I guess that resonates with a lot of people is around Red Bull racing. So we've had a partnership with Red Bull Racing now for, well probably I think this is our third or fourth season and people often think that this partnership if you like, is just a commercial partnership. So we pay money to put our brand on their vehicles. as I race around many Formula One tracks globally, but actually it isn't. So just to kind of give you an insight, this is from Red Bull themselves. One of the massive parts of ah, Formula One is around the data strategy. So how you set up the vehicle, the car itself, how it then responds to the track itself, then what your competition are doing, what the weather's doing, what the track surface is doing, and as you can imagine when you put all those variables into the mix, there is lots of different outcomes that could occur based on maybe the track temperature change and et cetera. So one of the things that Red Bull have to do with data, ah, essentially is do predictions. So before practice, if we set up the car in this configuration, what does that actually mean in terms of an outcome? And I think the number is, and don't hold me to this number, they do something crazy like a billion race simulations before practice, another billion before qualifying, and then billions more during the race. And you know, a real famous, example, much to the annoyance of much of the british public, is they actually use one of these simulations running on Oracle cloud to really decide to put the slick tires in the Abu Dhabi Grand Prix, which ultimately resulted in max winning the championship that year, rather controversially. But as you can imagine, it's all about data. So having the right data, having the right models, but more importantly, having the infrastructure and the cloud services to support that is really important. So a bit of a commercial example, but I guess it leads us really into, we need governments to ultimately be following this mindset. You know, if we want to be as a very kind of developed nation kind of on the bleeding edge, we have to be using all of the data that we have across government to ultimately do things like predictions, you know, such as if we dig up this road here, what's the knock on effect, you know, a mile away of traffic, you know, all the way through to, you know. And one thing we cherish in this country a lot is the National Health Service. There's no reason why, given that we're one of the most developed nations, we shouldn't be doing proactive healthcare today, not reactive healthcare. And this is all on the foundation of, data.
And Dan, if I may. Sorry. I'm Martin Dolce, I'm the sales director looking after, UK public sector, for Oracle's cloud business. But, just to add to Dan's comments, really, it was described to me the step change from, if you like, pre Internet to Internet. It's hard for us to imagine that, those days now, particularly when I describe it to my children, they think that I'm 100 years old. But it was described to me that that's the kind of the step change that we're looking at as we move into the AI era. And I think that's what's really exciting for me is the opportunity that comes with that because it really is kind of limited by our own imagination. It's applicable to every facet of society, whether it's a commercial business, whether it's a government organization, whatever it may be. It's really looking at how we operate today and identifying where and how we can improve on that. But ultimately, underneath all of that, to your point, Dan, is that data is the fuel that drives that?
Absolutely, Martin. I mean, you know, just reading the news, keeping on top, there's a feeling that artificial intelligence could save around 5 billion a year from the UK budget. And that 5 billion could then be reinvested in things like the road network as an example, or our healthcare, and ultimately could optimize tens of thousands of government jobs to make them more innovative rather than doing quite mundane activities. But I think the real challenge we have in the United Kingdom, is because we're a developed nation. The side effect of being such a developed nation is that we have a hell of a lot of technology across the whole of government, which is out of date. It could be classified as legacy. And ultimately these challenges around these legacy environments could ultimately really disrupt how quickly we can adopt artificial intelligence. And, you know, and ultimately the cost involved of running those legacy environments, we would like to see them reinvest that money into things like the new artificial intelligence, technology that we have in the cloud network right in today. But ultimately, when we spend a lot of money on keeping the lights on for these kind of old systems and m as well as, a lot of these environments, probably make it really difficult to get the data out of them, which, you know, these models, these artificial intelligence models would require for them to be more accurate. You know, obviously the historical data you put in can depict what prediction you get out at the back of it. So ultimately, you know, I think government has to really think about how they go about resolving some of these legacy challenges and almost moving away from the term legacy a little bit because I think it's got this real negative connotation. So I guess realistically, what we want to do today is kind of talk about a real story, actually, of a customer that has undertook this journey to modernize their legacy world, to make it more, I guess, modern, to bring it into the new world to support these AI legacies. And we want to talk about that journey. And then more importantly, what we advise as a corporation that Oracle has been a custodian of secure data now for over, I think since our 47th year. so we feel we are in a good place to talk around how customers can ultimately get the most out of data. Ah, but ultimately modernize these data platforms as well, where a lot of this data sits on.
I think just adding to some of the comments you said there. So obviously in my day to day job, I talk with, across government, lots of the government departments, and I think you're right. I think legacy, or sometimes it's referred to tech debt, has kind of been, almost an embarrassing secret within some departments. They know it's there and they've buried their head in the sand and that's probably because frankly it's still working and maybe that's to Oracle's testament. You know, a lot of the stuff that we produced isn't broken. and given the financial constraints that we've faced over the last ten or 15 years as a UK public sector, they probably prioritize spending that money elsewhere. But I think to your point they're saying exactly that. They're saying, look, it's prohibitively expensive. Some are telling me that it cost them three times more to maintain the old stuff than it would to bring on board new stuff. But there's finding the cost to make that migration. So that's certainly one of the challenges. I think there's a recognition around security vulnerabilities and again let's be open eyed about the world that we live in today and the threat of nation states. Any kind of security vulnerabilities are going to leave us susceptible to attacks or disruption. So I think there's an element of that and particularly as that software gets older and older, that makes us more and more vulnerable. They're very conscious of that. They're very conscious of the cost of the spending and to your point, the conscious of the data silos, right. That actually some of the new software doesn't necessarily plug into the old and therefore to use AI as the example, it's fueled on data and if that fuel isn't able to be accessed then there's a limited amount which you can do. So I think speaking to the government departments, they're very much aware of it. I think up until now they've focused on, on the, dare I say, the easy stuff. But I think they're fast becoming open eyed to kind of recognizing that they need to tackle that old problem, that legacy debt, that tech debt, whatever you want to call it. And I think you're right. I think we're seeing more and more government departments now approaching us to say, okay, can you help us with this?
Yeah, I'd agree Matt. I mean we have seen, if you talk about the security angle, we've seen some real extreme examples of this where customers have kind of waited for this new replacement of this legacy service that they're offering using modern cloud, native kubernetes, postgres and all these kind of new modern open source technologies. And they basically leave that legacy, sometimes critical national infrastructure that unpatched. So because they feel that this nirvana is two years away of this replaced system, but actually five years on I still haven't done anything with it. And you know, a lot of the time it essentially is left unpatched. So these vulnerabilities you talk about that nation states would absolutely love to exploit are existing in these legacy environments. So, you know, I think it is super important and when we, when we talk about the customer that, you know, is going, undertook this journey with us who are department of work and Pensions, it will kind of give you all, I think, a bit of a realization of their journey and how probably similar it is to your environment today. So I guess if we go on to the story, Martin, so I know this has been quite close to your heart probably for the last 18 months now. So tell us about where the department for work and pensions kind of started, on their journey and then we'll talk about, I guess, what they did and what some of their pain points were, etcetera.
So, yeah, you're right, we engage with the problems of working pensions. About 18 months ago, they had a stated objective around their journey to cloud. They had about 30% of their business in the cloud, they had 70% on Prem and they wanted to flip that around. But they recognized that a lot of their real core applications that touches us all today. So whether it is the pensions payment system or the benefit systems, that was still highly dependent on Oracle legacy software, and to be honest, they had, as I referenced earlier, they had deferred kind of upgrading that platform because they were challenged on budget. And so that had rolled on and rolled on and they were in a situation where there wasn't a natural migration path. And that's where they approached us. They came to us to sort of seek our help, to say, look, we know we've got this problem, we can't see a way through it. There was definitely a reticence to engage with ourselves as oracle. And they were very open with that, the outline look, you know, generally Oracle hasn't got a good reputation across UK government. We've not had a very good experience. Their experience quite often related to the cost of maintenance or an audit, say that Oracle would approach them with, and which resulted in them getting a bill at the end of it. So there was real hesitancy in engaging with us. And if I said that they approached pretty much everybody else that they could before coming to us, I don't think I'd be far wrong. But having made that engagement, we really did go on a journey with them of an, outline, an approach, that was agreeable to them, whereby we undertook a proof of concept that was to take their very old application and to migrate that into the cloud, and again without going into the weeds, and going into the detail, that's ultimately what we've done over the last twelve months to build confidence within departments, of work and pensions, both to the IT team that ultimately own the delivery of service to the business, but also to that department within DWP that was ultimately responsible. There was a real kind of nervousness, you know, let's face it, right? If that payment system goes down and people don't get the pensions, it makes an awful lot of noise, it causes them an awful lot of problems. And there was an understandable nervousness about making sure that we got that right and there was no disruption. and so that was ultimately what we've done. We engaged Oracle's kind of expertise around migration services and about the cloud tenancy that we looked to stand up for them. and we created that and it's gone really well. It's gone really well. I was super proud. We cloud world, which is Oracle's kind of annual conference that we hosted in London. Paul Booth from DWP stood up on stage and outlined the journey we've been on and outlined the change that he'd seen from Oracle. He referenced it being a completely different Oracle to the one that he'd engaged with before. He talked about it as Oracle 2.0. And really the complete turnaround approach that we've had with them, working with them in partnership, recognizing that maybe our timelines are different from their timelines and and working with them to tackle obstacles as and when they came up. So it's been on a personal level, it's been really rewarding and I think what's really nice is that they accepted that this was a partnership, that the nervousness initially about doing this, they were understandably kind of quite guarded about sharing this information whilst they were on the journey. But having done so and come out the other side, they're now open to go and talk to other government departments and share their experience. It's been really good, I think just.
To drill into some of the kind of facts if you like, around what that environment is. You know, this was running on Spark, right, which is a, I wouldn't say it's a legacy, but it's a completely different architecture to what, you know, all of the clouds offer in x 86. So they had that challenge around, well we can't just move it as it is because it's on spark architecture. You know, they were using databases which are no longer in support, as an example from a version perspective. And I believe that they were actually having to virtualize Internet Explorer six because that version they were running of the application only supported ie six. Being a previous Microsoft guy that I used to support ie six back in the day and you know, it wasn't a great browser. So to keep, ah, having to keep that around, just to support this application. And Martin, I know when we've been chatting there was quite a number of global service integrators who actually told Department for Worker pensions this journey is impossible, you're too old. You've got all these things I've just said around old version of the database all using Spark, not x 86 using ie six. You know, this was a tough journey in their minds more than anything else because if, you know, they went and spoke to a lot of GSI's who said it was basically impossible and don't touch it because, you know, like the famous saying is kind of, if it's not broke, why fix it? Kind of thing. And this wasn't essentially broken, it was just full of risk. And I think that's essentially what you kind of have to think about is when you look at these environments, it's a risk to the future in terms of the AI vision which we, you know, we kind of opened with. But more importantly there's a risk in terms of if you're not keeping up to date then there are, there will be security vulnerabilities in that system one way or another. There's the famous saying, you should always assume you're breached rather than think you're not and work backwards from there. So, you know, I wouldn't underestimate the kind of mindset and the almost courage that DWP had to have to come to Oracle, say, can you help us with this? So, you know, I think it's a great story, Martin, really of a you know, it's not a smaller part of the government either. I think as budget goes it's one of the top two, I think along with HMRC in terms of where the money's spent, our taxpayer money. so, you know, and these environments are critical national infrastructure, which you know, comes with a little bit of risk and nervousness when you're starting to think about moving them. So I guess if we step forward, Martin, about what we actually did, with them, it probably would help. I had a thought frame, this kind of mindset that they had to undertake. So I think first of all I guess it's like anything, right? And, you know, if you talk about addiction or anything like that, the first thing you have to overcome is realizing you have a problem, you know, and ultimately when you realize it's a problem. And I think that was one thing that I think DWP did quite well, is, you know, they kept getting told that the replacement for this service is two years away. Two years came, it was another two years. Two years came, it was two years. You know, and I think after five or five years, the service owner said, this is not coming anytime soon, is it? You know, and I think they had to accept that actually it's still probably another five years away, before this service is actually replaced. So it really, they had to draw a line in the sand at some point. That's the one thing I would, I guess my advice to a lot of government organizations that we speak to is we still have this conversation with them around, no, no, no, we're not doing anything because it's legacy and we're waiting for this new, modern service. And in essence, what happens is that, legacy service, if you like, is basically left to rot. So I guess if we'd think about what we would advise customers to do, first of all, and I think one of the things we see a lot of, when customers go to, to move is a lot of the time that they decide to move is because they maybe have a compelling event. So maybe their, data center is due for renewal, they have to buy new hardware. There's a huge capital outlay involved in that. Maybe they have to change service provider, et cetera. So your compelling event is almost a time bound problem. So we need to be out of this data center by this date or our contract with this SI expires in a year's time. What are we going to do? So you kind of have that very time sensitive challenge. What we then often see is customers say, well, what we'll do, because we've got this compelling event, we'll also use it to do a complete transformation of that service whilst we move it. So it could be running on VMware today, as an example, in your data center, in a coal off facility, etcetera. And then we see some customers and there's one good one comes to mind who remain nameless. And, you know, they had a great idea of moving their environment, again, critical infrastructure at the same time as moving to postgres away from Oracle. And, you know, we advise them, don't do it. Not that we want to keep you on Oracle with visa, you're mixing different problems together. And one thing I'd always say, and it's kind of an interesting kind of point really, about, I guess, when we have a public sector as large as what we have is, I remember speaking to somebody before who worked for the NHS, was their customer. Ah, and they were saying, if you look at this trust, you know, you've got 150,000 kind of people who kind of work in this area. You know, if you actually look at an IT system, generally speaking, when you introduce a new service, you generally want a pilot or proof of concept to cover around 10% of your organization. 10% of 150,000 is 15,000 users. But are on a pilot, right? I guarantee you now, having been on these pilots with FTSE 250, who've maybe got three and a half thousand employees, so you're talking 300, 5350 voices, opinions of people who say, this isn't working the way I said it works. I don't like this new system. It's not fast enough. Imagine scaling out to 15,000 people. This is before you've even got to any form of service transition or anything like that, you know. So when you think about moving, ah, an environment and then transforming it at the same time, at the scale that I've just talked about, there's no absolute surprise to me why a lot of these fail, these projects fail a lot of the time when you're operating at this level of scale. So I think for me, my first piece of strong advice based on the wisdom I've had in my career, don't mix your goals together. If you have got a compelling event, keep it as simple as possible. Move the environment as it is. Maybe keep it on VMware as a good example. And then when you've moved it to the cloud, then look at doing the modernization. And Gartner have got a really lovely mindset around the five r's, you know, do you replace that environment for SaaS? Do you rebuild it to cloud native? Do you maybe re platform? So you maybe put it on a new database or platform or whatever, do you retain it on VMware or do you retire it? So replace, rebuild, replatform, retain and retire once you're in the cloud. And I think that would be one of my big pieces of advice here. And I think that's if you look up, the DWP journey, that's exactly what they did. They're moving, they have to do a little bit of transformation just to get it into the cloud. As an example, the application they're running on was a spark version of the application. They've now got to deploy the x 86. That is technically transformation, but it's a small jump. They're not turning around and saying, well, we want to replatform from Oracle to postgres at the same time and having all those data challenges. So think about should we just spin up a vmware in the cloud? And all the big, cloud providers have vmware today. Oracles is quite a unique offering because it mirrors what a lot of customers have in private clouds. on premise, we can mirror that with our offering. So move that first of all, as it is with as little change as you can. And then at that point I think is think about what you can do to modernize that, you know, and a real great example, I think, of what department for work and pensions have done is when they've moved it to the cloud, their next step is to move it from a, traditionally customer managed database into Oracle's autonomous database. So the autonomous database, if you like, takes away a lot of the kind of, I wouldn't say mundane tasks, but the BAU tasks that your database admins would have to do. So in that journey, by the end of it, when they go live with this service, it'll be running on an autonomous database, which ultimately means that they've reduced their overhead, they've reduced their costs, etcetera, but they've only done it at the point they've moved. So I think it's a real, real key point of don't mix your goals, ah, together and try and keep it as simple as possible.
It's a really interesting point actually, because again some recent conversations, and I say recent over the last three or four weeks, is on the back of obviously the acquisition of VMware, and the change in business model. And I've seen a lot of clients, understandably across government, have seen an increase in their vmware, costs, which has caused them to maybe accelerate their timelines that they had in mind in terms of migrating away from VMware or doing something with their vmware estate. And it's interesting you say that because as you say, some are keen to go down a kind of cloud native migration. and that's great. But as you say, that takes time. And I think when we've shared the oracles ocvs offering, we do give them that stepping stone option to do a lift and shift, take their existing environment, put it into ocvs and then migrate at their pace. And again, we connected two customers, at Cloudwell, actually, and they were having the conversation about doing just that and how it allowed them to move at their pace, at their speed that they were comfortable with, that their organization was comfortable with, rather than kind of into panic stations and try and move off as quickly as possible. Because ultimately, when you're moving at that kind of pace, sometimes that comes with an increased risk and mistakes can happen. Would you agree?
I would wholeheartedly agree in mine. I think one of the central issues when customers move to have a mindset of moving everything to cloud native. And Amazon is a really great example of, they've actually moved one of their cloud native apps back to being a monolith app, because not everything makes sense to be cloud native, you know, and I think where a lot of organizations struggle a hell of a lot, you know, it's the famous people process, and technology people think the move to cloud native is a technology jump. It absolutely isn't. It's an organizational jump. You have to basically create your, in that cloud native mindset. So everybody's responsible for their own little part of the puzzle, and then they've got to almost be, then have the process of, okay, how are we updating all these different component parts of this application? How many of these component parts do we have? Do we have, 50 different moving parts that everybody in the organization have all got to be aware of each other? And from where I've seen, lots of public sector are just not set up to do it that way at the moment. Not saying they won't get there, but I'm saying today. And that's why for us, this kind of mindset of move as is using things like oracle cloth, vmware, and then look at what makes sense to kind of branch out, if you like, and into all these kind of new kind of modern, mindsets, etcetera. And I think, you know, if we kind of step forward on kind of what some of the outcomes of doing it this way, you know, if we, if we go back to how we open this, this podcast, it was, how do we make this data available to these machine learning and artificial intelligence models? First thing is gaining into the cloud, gaining it closer to these, the compute that ultimately runs these models. So, you know, when, when you look at moving from it a legacy version of Oracle database as an example into autonomous, what you then do is expose that data to a whole host of modern, APIs. A lot of them today are all rest APIs, which is how a lot of these environments all communicate. First of all, at the point you move that into say, autonomous database as an example, you then modernize it to make that data available, which is the first, I guess, challenge of trying to overcome the kind of, I guess, paralysis that a lot of organizations are in at the moment. So do the small jump and then make the data available. What they essentially do, you will see a reduced cost. We see generally when customers do that short step, they get around a 30% cost saving. You know, it does vary. this is just an average cost that we see. but also what we see in DWP are a great example is they can do things today that they couldn't do previously. There were parts of the application that just didn't work previously and they'd almost accepted it didn't work. and the funny thing is they've now been able to move away from all of that, virtualized Internet explorer six, and they now can use chrome. So they only have to manage one browser today rather than two, which is a crazy thing to place to be in. But these are the benefits that customers are getting when they do this jump. And I think if we tie this back to our security aspect as an example, what a lot of cloud providers offer today is complimentary services that surround these data platforms, as an example. So we, or have a good one called data safe, which you basically point it at your data platform and then it will look at, ah, who's running these queries that look a little bit suspicious, who has these privileges that, a lot of challenges we see a lot of the time is things like joiners, movers and leavers processes where you've got people with privileged access that when they leave the organization, they don't always tidy up that privileged access. These tools can give you a real clean dashboard view about who has had the ability to run these privileged queries, which if you look at data exfiltration, one of the key things they'll always do is try and get an account with the most amount of privileges to run the queries. They want to get the data out. So, and a lot of the time it's just awareness about who does have those, privileges. So, you know, if you can have surrounding services that surround that data, even though it's classified as a legacy system, you've modernized a part of it in a very tactical fashion, and you then get increased security posture from that using a bit of a defense in depth approach. So you've moved it to the latest version of the database as a replatform, exercise. So you then tackle all those maybe CVE's that were existing in your environment, but then you then surround it with a modern data kind of wrapper, which give you then a lot more peace of mind around your kind of security posture for these platforms. As I think Martin said when we opened around the nation states and the way the world is at the moment, it's probably never been, a more unsafe place, especially in my lifetime. Anyway.
No, it's really interesting. So I guess wrapping up the three things that, that I think has come from the conversation today, particularly for UK government, is let's not bury our head in the sand. Right. Legacy is not going to go away. The problem is going to get bigger. So we've got to tackle it head on, and we've got to, and in doing so, we've got to get it into the cloud, right. And that's probably about looking for the right workload to go into the right cloud, recognizing that it's a multi cloud world. so I guess that's the first point, and I take your point about getting closer to the modern tools and making use of AI. I think the other one is eating the elephant. One bite at a time, if you put the expression. But let's not try and do all things at once. So let's not try and migrate and then move, but actually let's get into the cloud early and then that allows us to modernize at whatever pace is right for that organization. Right. Whether some might have lots, of resources and a high appetite for risk, some might want to do it gradually, but at least we're taking it at a pace that everybody is comfortable with. So I think they're probably the key takeaways that I've got from our discussion today. Is there anything that I've missed, Danielle?
I think, first of all, is don't fear. Don't have fear about tackling these environments, and certainly don't have fear around coming to Oracle to talk to us about them either. You know, and, it's probably worth a chat, really, around kind of this new oracle. I guess, that, you know, a lot of customers are experiencing DWP one, but I'd just like to kind of just say don't fear. And, you know, a lot of these, a lot of our, these cloud and, providers are all here to help you, right, one way or another. So don't fear and certainly don't accept it's impossible, because there's always a way forward, I think.
No, agree. So, yeah, I guess wrapping up, please come and speak to us. We're here to help. We're not scary. we're not going to try and sell you something you don't need. certainly we, as Oracle, are a different organization than we maybe have been in the past. and we want to work to help, particularly across UK government.
That's a really good place to wrap up. I think the messages at the end are really, really nice and strong because a lot of this does seem like there's a lot of work to be done. You've touched on the fact about, you have to use your phrase, eat the elephant one bite at a time and not being scared of taking on that challenge because the benefits that will be gathered from making that jump, eating that elephant, will in the end be so much more because, you know, you mentioned the security aspect, which is such a crucial thing for central government or local government as well, but also just making the efficiency and the way that these organisations run so much better.
You've been listening to the latest podcast from Public Sector Executive magazine. Don't forget to like and subscribe to make sure you receive every new edition. It.