The edge compute project team wants to hear from you. This month, the project team tells us more about edge compute, what the project involves and how you can help influence the outcomes of the project. 

Firstly, what exactly is edge computing?

Until now, Research IT has managed only central platforms, such as the Computational Shared Facility (CSF) and Research Data Storage (RDS, aka Isilon). And we have encouraged researchers to use these central platforms if at all possible.

But sometimes there is a good reason to build and run local infrastructure — workstations, storage (NAS) and even fast local networking. Examples might include pre-processing of data taken from a lab instrument or graphically-intensive post-processing. We call this edge compute and satellite storage.

Can you tell us a little bit about the edge compute project?

The project aims to create a service which will provide edge compute capabilities to the research community for computationally/data-intensive work. In particular, where the use of centrally-hosted platforms is impracticable or must be augmented with additional functionality.

A common use case is where researchers have a large amount of data, generated locally, which should be screened, pre-processed or analysed interactively, or visually, before uploading to central Research IT platforms. Another is where local processing is required for instrument data before forwarding to central data stores.

This service will:

1. Formalise the design and costing of project-specific IT platforms and support within research grant bid development.
2. Provide a simple model for provision and charging for professionally supported local IT platforms.
3. Speed up the provision of such platforms upon bid success.

A successful outcome of this project will eliminate the requirement, in almost all cases, for researchers to design, procure, implement and manage their own IT platforms. (There will be exceptional cases, including those where the IT platform itself is the subject of the research).

Components of the service will include:

• A standard “menu” of high-performance workstations, local storage and fast network infrastructure.
• Professional support options including remote management of local platforms by IT Services.
• Consultative co-design of edge capabilities to address the requirements of any given research project.

How will the project benefit a researcher at the University?

The edge computing service will enable standardisation and professional support of local and project-specific platforms, and reduce shadow IT. This will result in a reduction of risks to productivity, data loss and data breach. It will also yield cost savings through lower procurement costs, more efficient support for edge computing and leverage IT Services’ upcoming End User Compute (EUC) service.

What has happened so far in the project, and what’s next?

The project has just started, so we are holding two workshops in November (one in south campus and the other in north), and a third workshop in January. The two workshops in November will mainly focus on gathering feedback from the research community on what this service should and should not do, in order for it to be useful to them.

You can sign up for the workshops via the below links:

• Thursday 7 November – 9:30am – 11:00am (University Place, 2.219)

   

• Wednesday 13 November – 2:00pm – 3:30pm (Sackville Street Building, J15)

The third workshop, taking place in January, will review the outputs from previous workshops of the user’s perspectives from the previous workshops and detail what the service will look like. Further details of this will be communicated in future RLP newsletters.

If you are also interested in this service and would like to become a use case for the project, you are more than welcome to contact the edge compute Project Manager, Daniel Sanchez.

Related