Large scale deployment of clouds to support scientific research
I'm responsible for the team at CERN which manages the operating system and infrastructure services. I previously worked as a Unix kernel developer at IBM along with managing large scale Unix production deployments and services for Deutsche Bank in Europe.
One of my team's responsibilities is running the CERN OpenStack cloud which has been in production since July 2013 and is currently around 27,000 cores. This cloud provides processing power for the physicists working on the Large Hadron Collider and its experiments, producing around 35PB a year.
As a representative of a non-commercial organisation, I speak regularly on our experiences and approach at OpenStack, academic and industry conferences along with sharing details of our deployment at http://openstack-in-production.blogspot.fr/.
I'm involved in the following OpenStack projects: Nova,Glance,Keystone,Horizon,Cinder,Ceilometer,Openstack-manuals
I have been a member of the OpenStack management board since the foundation was launched and am also a member of the user and elections committees.
As part of CERN's expansion to a second data centre, we are implementing a large scale Infrastructure as a Service based on OpenStack aiming to support 15,000 hypervisors by 2015 to help analyse the 35PB of data a year from the Large Hadron Collider. The CERN OpenStack use case and experiences were described in the Essex, Grizzly and Icehouse summits.
OpenStack provides a key part of our strategy to adopt common tools in use by similar scale organisations and contribute to the community. As an open research centre, we can share our experiences only and benefit from the work of others towards a common goal.
Examples from the recent summit include our techniques for scaling out OpenStack (http://www.openstack.org/summit/openstack-summit-hong-kong-2013/session-videos/presentation/deep-dive-into-the-cern-cloud-infrastructure/). I also lead the CERN side of a research project with Rackspace which is contributing code to OpenStack on aggregating multiple clouds into a single resource (http://www.openstack.org/summit/openstack-summit-hong-kong-2013/session-videos/presentation/hybrid-openstack-clouds-cern-research-project-aims-to-solve-federation-for-the-real-world/)
Growing the user base is key to OpenStack's success. I have presented our activities at user groups and meetups in London, Paris, Budapest, Berne and Zurich during the past year and industry events such as GigaOm Structure. I am currently organising the Swiss and Rhone Alpes user group meeting in December which is expecting over 100 people.
As a member of the OpenStack user committee, we run the confidential survey of users and deployments, consolidating the results as input for the design summit and the technical committee.
I was elected in 2010 to the OpenStack board of directors as a member representing the individuals of the foundation. While it has involved many long phone meetings and mail discussions, I have found it highly rewarding to be part of the OpenStack movement and contribute my experiences as a user to the board decisions. Equally, frequent outreach activities and summits give the opportunity to discuss with community members from development through to operators and ensure that I represent the broad range of opinions of the membership.
Since my daily work is for a non-profit organisation, international collaboration and community consensus is part of the IT management team culture. From the invention of the world wide web, CERN has been heavily involved in open source development activities, sharing our improvments and benefitting from the work of others.
Given the continued expansion of the membership and OpenStack related developments, the board's role to ensure correct governance and set strategic direction for the foundation becomes more critical. The individual membership role allows the community to a voice and should reflect the varied electorate from developers, operators and end users.
With the aim of clearly identifying the central components of OpenStack, we need to ensure a prolicy for the correct use of the OpenStack trademark is needed so that operators and end users understand what an OpenStack cloud should provide and we can achieve the goals of multiple compatible implementations. Innovation and functional expansion is welcome but this must be within a clear definition of what is OpenStack and without diluting the continued improvements in the maturity and ease of deployment of those components.
Directing the OpenStack foundation funds towards the enlargement of the community, especially outside the US, along with encouraging further production deployments and user stories will increase the deployments even further.
Firstly, there is a need to clarify what OpenStack is (and what it is not). The ongoing work around the definition of core is key but we also need to encourage adoption of the OpenStack APIs in other products while allowing innovation in the ecosystem around OpenStack such that projects can thrive without needing to be core.
Secondly, there needs to be further support for production deployments. Reference architectures, user stories and focus on the areas which improve the installation/migration/configuration experience.
Lastly, continue the drive for international expansion. With the summit in Hong Kong and upcoming Paris summit in 2014, OpenStack has become a global community. Assisting ground roots growth with programs such as the user group Ambassadors and the travel bursary for those who needed financial assistantance to attend the summit need to be further built on to ensure that developers and users worldwide can easily contribute their skills and energy to the growth of OpenStack.
I hope to have the opportunity to continue to represent the individuals of the foundation in 2014.
Tim has already been nominated by: