Optimising large-scale data transfers between sites
Dr Tim Chown, Network Development Manager, Jisc
While research computing support focuses on the design and implementation of high performance compute and storage facilities, it is also important to consider best practices in optimising large-scale data transfers between such facilities. In this talk we’ll explore the ’Science DMZ’ model which includes consideration of the local network architecture, well-tuned data transfer nodes, appropriate software tools and efficient implementation of security policy with a view to creating a ‘friction-free’ end-to-end network path to enable high performance transfers. We will also briefly cover network performance monitoring through the example of the open source perfSONAR platform that is widely used by the CERN experiment sites.
A One Team Approach to Research Infrastructure Software Deployments at KCL
Operations Director (Research & Researchers), Kings College London
Research & Information Technology: How IT supports the needs of Researchers
At KCL, software deployments have been impacted by the usual issues that can often flag an IT project as failing, resulting in a legacy of blame and mistrust between IT and the business. At KCL, taking a ‘One Team’ approach has eliminated the divides between IT, Professional Services and Research Academics which has resulted in the most successful IT deployment the College has seen for several years.
Dr Martin Kirk, Operation Director (Research & Researchers) at Kings College London sets out to provide an overview of how an institutional wide approach to engagement, cross functionality and collaboration has removed boundaries, built trust and reminded everyone that success comes from seeing beyond their traditional team to the One Team that is KCL.
Building Trusted Partnerships with Researchers
Jon Lockley, Deputy CIO (Scientific Computing), Institute of Cancer Research
Researchers depend on digital services both directly and indirectly. This talk will highlight some of the ways ICR’s Scientific Computing service has built strong partnerships with our researchers and where contentions are most likely to arise with a “traditional IT approach”. The talk will also cover what our biggest emerging challenges are such as the growing demand for large scale collaborations, and managing huge volumes of data against a backdrop of increased cyber threats and regulation.
