From our virtual 4m below sealevel studio in Almere we caught up with Mark Parsons, Director of the Edinburgh Parallel Computing Centre (EPCC), to look into the recently announced selection of HPE by EPCC to power the Edinburgh International Data Facility (EIDF), Europe's first regional data innovation centre.
Before we go into the details of this collaboration with HPE, can you tell a little bit about what the Edinburgh International Data Facility is?
Mark Parsons: The Edinburgh International Data Facility is a new investment funded by the Scottish and UK governments. It has actually come about over the last three years. The UK government has funding, called City Deal Funding. It is about investing in a city region. Edinburgh is in the southeast Scotland city region. The International Data Facility that we are going to build is designed to increase the use of data analytics, data science and, in particular, data driven innovation across the whole region over the next 10 to 15 years.
What is the current status of the facility?
Mark Parsons: The whole programme is a very large programme. We are talking well in excess of a billion pounds. We have been building the facility itself over the last two years. We have created a new computer room which is going to open in September 2020. The recent announcement with HPE is a very important step in the process. Over the last year, we have been running a large public procurement for the IT provider that is going to provide us with all of the IT equipment, initially over the first five years, but if it goes well, it is a contract that could be extended for up to 10 years. It is worth a 100 million pounds in total. A key thing about this is that we are not just going to spend 100 million pounds at the start of the process. The aim is to have continuous investment in the infrastructure over a 10 to 15 year period.
The facility will be starting somewhere in the fall of this year, is that correct?
Mark Parsons: Yes, in fact we have already built parts of the infrastructure as we have been developing how we want it to come together. It is a large private Cloud designed from the outset to support data science projects, digital analytics projects, AI projects and IoT. We have designed it so that, although it is a private Cloud, it can interoperate with all the major Cloud providers.
Can you give some examples of the usage that is made of it, or the intended usage?
Mark Parsons: The idea is to use data-driven innovation to transform our city region. We realized that Edinburgh is a city that needs to invest in in its economy. A key way of doing that is to use the data science revolution to help us do that over the next decade. A key part of that is training people. One of the things we are going to do is that, in a city of about 600.000 people, we are going to train 100.000 of those people over the next 10 years in different aspects of data science and data driven innovation. So, if your boss walks into your office and asks you: 'What is data science?', you will be able to tell your boss that and think about how you can apply it, either in your business, in your government department, in your school or your university.
That sounds like a lot of people that will get trained then, at least in Edinburgh. What is the international component of it?
Mark Parsons: Edinburgh is a very international city. We realized early on that there are a number of other cities around the world that are taking forward similar plans. Our idea is to work with them. Particularly, we have been talking to the city of Copenhagen. We have also been talking to the city of Chicago in the United States. I think the other use of the word 'international' obviously is Edinburgh should still be seen as an outward facing city welcoming people over the next decade and more.
What is the time plan for the whole project?
Mark Parsons: The initial infrastructure that we are putting in place at the moment will be in place over the next six months, with a full service available in the second quarter of 2021. We have taken a building block approach to the infrastructure we are putting in place. As people use more bits of the infrastructure, the bits that they use the most will buy more off. We are not trying to second-guess people at the start. I have funding for a decade from now. With the expectation that it all goes well, it will run for considerably longer than that. This is a really important investment for the city region of Edinburgh.
If you look at the size of the systems, what are we thinking about? What are the current ideas?
Mark Parsons: The initial big investment we have made is a big 20 petabyte cluster data storage that has been coupled with another 20 petabytes of block storage and storage that we can configure ourselves, plus an initial set of about 50 large VM servers, etc. We are starting off quite small to meet the needs of the projects we have but I think certainly in the next two or three years, we will be dealing with between 50 and 100 petabytes of data storage, all properly backed up, and all fully set up with a disaster recovery. It is a big undertaking.
Can you tell a little bit more about the deal that you had with HPE?
Mark Parsons: Yes. It was a competitive procurement. We have got a number of excellent bids, I have to say, for a number of suppliers, but HPE won it. We have had a strong relationship with HPE over the years. Obviously, they are providing the next national supercomputer which is being installed at the moment and we already own a number of pieces of equipment, for example HP A600. However, this is really branching us out into a different area of HPE's business which is the data side of their business. We have been working very closely with a number of their technical teams to think about how you use the the technology that is available today, in order to build a large private Cloud. This means we are looking at new technologies they are offering, for example this Ezmeral Platform that they have which helps us to manage containerized environments in a large private Cloud. At the moment it is very exciting because we are just experimenting with lots of different ideas with them as we complete the first part of our design.
You very briefly mentioned the Archer. That is a completely different system, of course. It is the new national supercomputer in the UK. Everything is going fine with that, we assume. It is just being installed now?
Mark Parsons: There have been delays, largely related to COVID-19, but I am very happy to say we received the first four cabinets of Archer II in July. They have been installed and over the next few months we will install the remainder of the cabinets. So, we are looking forward to having a new national supercomputer in the UK.
Thank you very much for this interview.