NHS Spine2: Q&A with Basho, the US developer behind Riak

Basho's software engineer Sean Cribbs and marketing manager Jeremy Hill answer questions on the new infrastructure

US developer Basho's open source database Riak was selected as the NHS's preferred choice to underpin its efforts to rebuild its Spine infrastructure.

The new project, dubbed Spine2, will replace the existing infrastructure that was implemented as part of the ill-fated NHS National Programme for IT (NPfIT) - which was scrapped but cost taxpayers nearly £10bn.

Computing sat down with Sean Cribbs, software engineer at Basho, and Jeremy Hill, the company's EMEA marketing manager to find out more about the project...

Q. Why has the NHS chosen Basho instead of a bigger, more recognised vendor?

Hill I think part of what drove this was the government's drive to use open source software, which from their perspective is more cost effective, but at the same time they needed software and an organisation they could depend on, and Riak is a dependable technology - it has been proven with other organisations like [cloud-based meter data management company] Temetra, Rovio and CNN. They have been able to test it and see that it works for them as well.

Q. Do you feel that there is pressure on you as a vendor, after the failed NPfIT programme?

Hill We realise that this system has to work, and we'll work very closely with the NHS to build the project to their requirements. I don't think we feel an unnecessary amount of pressure, because we know the technology works and has been proven and deployed elsewhere. We know it simply can't afford not to be available - for example, in instances where a patient's life is at risk, we know if a server fails over the system will carry on running and we are not concerned about that

Cribbs That is a common story with our customers - sometimes they suffer hardware and software storage outages, but they don't notice because [the system] is resilient and it's still available to use even with a degrading capacity.

Q. The Riak NoSQL database replaces Oracle's relational database - what are the benefits of moving from a big player like Oracle to a smaller player like Basho?

Hill Generally speaking, the advantage of moving from a legacy-type platform is that previously people had to buy a box to store all of their information on, and as they upgraded the specification of that box, they would need to buy a bigger box and typically that gives people exponential cost increase both in terms of licence fees and hardware.

Typically, Riak gives people an alternative approach and that's a distributed approach, so instead of buying one big box as your requirement scale, you just add another commodity server and another licence to it. With a legacy system, if it fails, it doesn't matter how reliable it is, you can't help it from falling over. One of our customers, Temetra, mentioned that they hadn't even realised until they looked at their error logs that one of their servers had fallen over. So in terms of reliability it gives them a huge advantage as they don't have to worry about something failing.

[Turn to page 2]

NHS Spine2: Q&A with Basho, the US developer behind Riak

Basho's software engineer Sean Cribbs and marketing manager Jeremy Hill answer questions on the new infrastructure

Q. Do you take a different approach from a technical point of view when you're dealing with a public sector customer like the NHS?

Cribbs It's related more to the size of deployment [than the sector]. We have some customers which have an insane amount of traffic, and we are frequently in contact with them to make sure we are meeting their requirements. We've already learnt a lot from our interactions with the NHS to make the product better.

What about your approach to security, is there anything you do differently with the NHS?

Cribbs No not necessarily, we have generic services for that. The main thing is that they are interested in always having the data available. They are focused on how critical the data is, and other customers are focused on the reliability of the service itself. This ends up serving the same needs because [we] want it to be responsive to their load but also ensure that we don't lose data or corrupt data.

Q. Ovum analyst Laurent Lachal warned that agile is "quite a challenge for large, established organisations", do you agree?

Hill No I don't think so at all. I think it's a positive step that the NHS has taken this approach. I don't see why it should be more difficult for them to do than anyone else, but it is obviously working well for them and they should be applauded for what they've done. They have been able to build a very visible project in a short period of time.

Q. The overall strategy for the NHS is to share more data within the system and potentially with other parties - will Spine2 make this easier?

Hill I think the openness will allow it to be easier but in a secure way. The obvious concern is whether it will be secure - and the answer is yes, and it will make it easier where it's appropriate to share information.

[Turn to page 3]

NHS Spine2: Q&A with Basho, the US developer behind Riak

Basho's software engineer Sean Cribbs and marketing manager Jeremy Hill answer questions on the new infrastructure

Q. How does your relationship with the NHS work - who does what?

Hill I like the NHS's new approach to building large projects. Part of it is that they outsource where it makes sense. So [IT consultancy] BJSS is helping them to develop the software, and the NHS continues to manage the project. We work closely with them and the people developing the software.

It doesn't make sense for the NHS to employ a larger team of programmers just to build Spine2, it makes more sense to outsource that to people who have built similar projects already. But at the same time it doesn't make sense for them to outsource all of the ongoing maintenance of the database and therefore they will do that internally - it's a good balance.

Previously, the NHS outsourced [everything], which doesn't eliminate risk but passes on the risk to a third party, meaning it can penalise them when they mess up. The other option is to bring it all in-house but that assumes you have the skills internally to be able to do that on an ongoing basis. I think it's recognised now that neither of those approaches is particularly the best thing to do.

Q. How do you measure whether the project has been successful?

Hill I believe agile allows and encourages people to embrace change as part of the project, so that you can modify things if necessary rather than having to change something that would break up the project. It's very much something the NHS will be reflecting on in time, as their requirements change.