Annual Scientific Report 2015

Systems and Networking
The Systems Infrastructure Team manages EMBL-EBI’s compute servers,
storage, virtualisation, data centres and networking, including managing the
campus Internet connection. The team works closely with all project groups,
maintaining and planning their specific infrastructures, and plays a key role in
managing the technical frameworks supported by the UK Government’s Large
Facilities Capital Fund.
Major achievements
In In 2015 our team maintained EMBL-EBI’s growing
compute infrastructure, which is now based on
approximately 30 000 Central Processing Unit (CPU)
cores. As our user base is very broad, with wide-ranging
technical requirements, we organised two training days
for approximately 70 members of staff to get up to speed
with EMBL-EBI clusters. To support internal users
working in a distributed computing environment, we
built a new, 36-node Apache Hadoop cluster.
We maintained EMBL-EBI’s secure, high-performance
GridFTP data-transfer protocol in 2015, and as part
of this work we completely rebuilt and automated
the installation process. This makes it much simpler
for users to share large datasets on our networks. In
addition, we installed MongoDB machines to provide an
alternative database platform for our internal users.
Security is paramount to the integrity of the controlledaccess
European Genome-phenome Archive (EGA),
which EMBL-EBI jointly develops with the CRG in
Barcelona. Our team re-implemented the vault firewall
for the EGA in 2015, using a software approach. This
improves the performance of the system substantially.
Our team put considerable efforts into maximising
flexibility and optimising compute performance in 2015.
We rebuilt our compute clusters’ General Parallel File
System completely, improving high-speed file access
for multiple applications executing on different nodes
within the clusters at the same time. We also continued
to develop Elastic clusters, and plan to complete this
work in 2016.
Given the scale of the storage challenge for EMBL-EBI,
our team’s focus is always on identifying and
implementing the right technologies to maximise
efficiency. For example, Docker containers create a
wrapper for an application, with all of its dependencies,
which provides a more standardised and softwaredevelopment-friendly
unit. Importantly, they also
Detail from the Genome Campus data centre, with Dawn Johnson, data
centre engineer..
guarantee that the application always runs the same,
regardless of its environment. In 2015 we enabled
internal users to run these containers in our clusters,
which resulted in improved efficiency and more
consistent performance.
Our team is responsible for maintaining the raw data
storage capacity of the institute, which exceeded 60
Petabytes at the end of 2015. A substantial portion
of this represents raw sequence archives. Although
159
2015 EMBL-EBI Annual Scientific Report