Q. Welcome Graham and Rich. Please explain your roles.
A. Graham: I’m head of technology for the Met Office. So that means I’m responsible for all IT development and support, IT-related project management. That amounts to what — about 350 members of staff and an IT budget of around about £35 million per annum.
Rich: And I’m Rich. I’m an enterprise IT architect looking at essentially our high performance computing platforms.
So in terms of the Met Office, we are a govern — what’s called trading fund. So we are part of the UK government and we’re an arm of one of the big departments in the UK government, which is the Department of Business Innovation and Skills. We are — it’s this strange entity called the trading fund. So what that means is we have individual contracts with individual organizations to bring in our revenue. So our turnover is about £230 million per annum. And about half of that comes from something called the Public Weather Service, which is a collection of government departments who fund the core of our infrastructure.
Q. How does that relate to Meteorology?
Meteorology is a slightly odd discipline in that you need to invest huge amounts of money and energy into just setting up your basic infrastructure. That amounts to around about £100 million per year. And beyond that, we then start producing forecasting products and the rest of the money comes from again individual government departments, the military, international organizations such as aviation and commercial customers around the UK and the globe.
Q. What are the core values for Met Office?
A. We have three broad goals. The first one is saving lives, protecting infrastructure and the natural environment. So we are the UK’s national weather service and one of our key remits is the national severe weather warnings service that we provide for government, the public and the emergency services.
The second thing we do in our remit is to try and stimulate UK economic growth. We do that by working with public bodies and commercial organizations to access and make use of our products and services to effectively make them more efficient.
And the third one — it’s a slightly odd thing. It’s about improving well being. If people know what the weather is going to do, they can sort of live their lives accordingly.
That’s probably a decent overview I think.
Q. Okay. Now, Richard can you give us just a brief understanding of what kind of development techniques you’re using today? So — meaning are you moving into service-oriented or micro services-type applications built in containers and that sort of thing. Or is that something that’s still on the horizon for you?
A. Richard: That’s still on the horizon for us. We’ve made a big push in the past couple of years to become more disciplined in how we do our agile development, and that’s been a big focus for us. We’re making use of the Spotify model of organizing our teams into squads and guilds to give us agile we can scale effectively within the organization.
We’ve been doing a lot of essentially traditional three tier style applications. We’re starting to see a lot of pen – of converting some of those applications to produce finer grain services, so more of a micro service based approach that allows us to move away from having sort of big monolithic that releases into smaller targeted services that can be released on more of a rapid update. But we’re a long way from being at — moving into a sort of a dev opp style culture. That’s there. That’s still quite a way down the road for us I think.
Q. How about with regard to the application development environment that you use — are you primarily developing – for example, are you primarily developing applications for deployment on Linux and assuming that you would be, what kind of frameworks and applications development environments are you actually using there?
A. So, yes. With — we’ve pretty much developed Linux. We have essentially two sort of streams languages that we use on — for the operational systems. On a more scientific basis, we make use of a lot of Python. We’ve got our own open source projects that we’re sort of moderators for and the scientific toolsets. And that’s something called IRIS. And so we make a lot of use of open source Python packages there.
And then on our more sort of traditional operational side, we’re heavy users of Java, making use of a lot of the spring framework. So we’ve started to use spring boot and those sorts of things to help us go on the sort of micro service journey.
Very into sort of like Apache Tomcat and those sorts of Web services, making a lot of use of messaging, so ActiveMQ and those sorts of middleware technologies as well. On the sort of — the data side — traditionally been an Oracle house for our databases. We’re now starting to make a bit more use of some other technologies. So we’ve been doing a lot of work with PostgreSQL — an alternative in the relational side. And then we’ve been doing some work with MongoDB as well on the no SQL and unstructured side for our data services.
Q. Where does the IBM LinuxONE system fit in the Met Office mix?
A Richard: The IBM LinuxONE machines–they have a mix of both Linux on z and z/OS on there as well for our sort of legacy — essentially back to production system.
Graham: So the both the z/OS element and the z Linux element is fundamental to what we do. So a lot of the z/OS stuff is quite all legacy applications but they’re used by a lot of the down — the follow on downstream systems and also by some key customers, such as aviation. And on the z Linux side, we have used that as a platform to consolidate all of our databases and also we’re right in the middle of migrating a lot of our infrastructure to that. And further to that, we’ve got some other exciting applications we want to port onto that environment which we can obviously pick up in a few minutes I guess.
Richard: It’s where we put what we call our systems and record. It’s the systems that are core to providing the forecast to all of our customers and it’s the ones that we know we just – they can’t go down. If we lose those, then we lose the ability to send out all the information we have to the people who need it. And so it’s — what it provides for us is sort of the resiliency and reliability and trust that we can get our data out whenever we need to.
Q. Can you summarize the benefits of LinuxONE to Met Office?
A. Richard: We have to be able to run 24 hours a day, 365 days a year, and being able to deliver our services day-in and day-out is really critical to our brand and our reputation.
Graham: In terms of quality, we’re able to use LinuxONE to underpin many of our services. If you take something like the emergency services, if we have severe weather coming in, they need to know what’s going to happen, when it’s going to happen and how severe it’s going to be
Graham: We can get the answer, the key element, the key weather warning, the key location, the severity of it to the right people at the right time. And, we’re talking seconds and minutes here can actually make a quite a big difference to, you know, people’s lives to infrastructure, to the natural environment.
Richard: The LinuxONE platform delivers this resiliency, reliability, and it’s just a platform that we’re able to trust and make our business depend on.
Graham: You know the system’s going to be there, it’s going to be up, it’s going to be available, it’s going to be reliable.