Newsletter: Compute equipment purchase, and lunch with the RLP
12 Dec 2018
Welcome to our final newsletter of 2018, in which we talk about our next business case, let you know when some of our new technology is expected to arrive, announce our ‘Lunch with the Research Lifecycle Programme’ event and talk about restricted data
Business case update
We’re about to present our last business case of 2018 for approval.
This business case has a particular focus on our projects concerning:
- Enhanced capabilities for professional services for research,
- Facilitating networking and collaboration,
- Incrementally building compute capacity,
- Developing a service to manage restricted data.
The latter two projects that deliver components of the e-Research Infrastructure are built on your inputs from our workshops held in the last few months.
Incrementally Build Compute Capacity, and Provide Orchestration (Projects M and K)
Following from our successful community engagement held in October, proposals have been confirmed to purchase key equipment that addresses the top priorities and concerns of the University’s research community.
We aim to deliver increased compute capacity, pending funding approval, which will start to be implemented in the New Year, as new technology is expected to arrive around April 2019.
The investments are:
| Graphics Processing Units (GPUs) | 16 Nvidia Volta (GP)GPUs — four × quad GPU nodes. |
| High-Memory Nodes | Three very-high memory (1.5 TB RAM) compute nodes. |
| Cloud-based Interactive Compute and High Throughout Computing (HTC) | Amazon Web Services , Microsoft Azure and/or Google Cloud Platform based computational resources: Web-based (eg Jupyter, Shiny) and other interactive; HTC accessed via Condor. |
| HTC | Incremental increase in on-site HTC capacity, accessed through either Condor or the Computational Shared Facility (CSF). |
Develop a Service Offering to Manage Restricted Data (Project S)
Working with the community we have categorised the University’s secure data needs into four use case scenarios:
- University of Manchester owned data,
- Established third party data relationship,
- Unestablished third party data relationships,
- Blended data; a combination of any of the previous cases.
We plan to rationalise and extend current facilities to meet the requirements, and evolve these based on the opportunities of the cloud approach to infrastructure.
This will secure current needs and expanding researchers’ capabilities to deal with projects involving sensitive information.
Lunch with the Research Lifecycle Programme
We will be holding a lunch meeting in the New Year so that you can come and have your voice heard on the progression of the RLP so far, and where you would like to see it go in the future. Refreshments will be provided.
Stay tuned to this newsletter and the RLP website for further information.