In this third installment in our “BHL and its developers”-series, we interview Lee Namba from Atos. Topics include the role he plays in the project, the Fedora-system and his favorite beer!
Lee, can you tell us what it is that you do for BHL-Europe?
Atos has been part of the project since the very beginning, which – compared to most people involved – is a very long time. And over the course of the project, I’ve done many different things, but in general, I (and Atos) took on the role of giving technical direction to the project and working on the architecture.
This included thinking about how to globally architect the system working with all the developers. We had to ask ourselves what the interfaces between components should look like, how we would use Fedora in the OAIS architecture, how to integrate with the content provider systems on the ingest-side … We really tried to find out the different possibilities and the benefits or drawbacks to each different solution.
On the development side, we also worked on establishing a shared development environment. This included recommending github as an SCM and issue tracker and setting it up, organizing the servers, VM’s, and environments, establishing a best practices development environment with the development wiki and so on. I also did a training session on continuous integration using Jenkins and on different types of testing for BHL-Europe.
You mentioned Fedora. Can you elaborate on that a little bit?
Fedora is an open source software that is used for digital preservation. In BHL-Europe the digital objects need to be preserved for a long time (50, 100 years or even longer!), and during that time files might need to change because of evolving computers, operating systems, and file formats. Other things that might change are the storage mechanisms (hard disk, tape, dvd, RAM memory …).
Fedora creates a way of organizing files so that in the future, if you lose your whole system, you can recover just the files and recreate the system. It is used in libraries around the world and allows objects to be transformed and accessed over the web. If, for example, there’s a high resolution TIFF file stored in Fedora, and you want to access it over the internet, you may prefer a smaller JPEG file because it’s faster to download and easier to view in the web browser. Fedora allows you to transform this large TIFF file into a JPEG file but we only use this function in a limited fashion.
Why in a limited fashion?
For scalability reasons. If Fedora had to do not only preservation but also transform and serve up these files, then it would be serving two purposes. For smaller libraries with a smaller volume of users, that’s fine. But we’re hoping many users will use BHL-Europe, so Fedora could become a limiting factor and access would be very slow.
Can you tell us something about creating the architecture from the very beginning?
At first, a large effort was made to prioritize and decide on the requirements. We had to think about security questions, site functionality, whether it was a high-traffic site or not… This was followed by thinking about all sorts of non-functional requirements, such as making sure that the components in BHL-Europe are robust and don’t require much maintenance and also administration requirements.
Then, we created an initial architecture diagram using google docs, and it has continued to evolve after testing and prototyping different solutions and changing requirements. From the architecture diagram, we created different project deliverables which were architectural and technical documents.
You have tasted beers from many different countries over the course of the project. What’s your favorite one?
Kozel from Prague!