Data Centre

The Data Centre forms part of the new Bristol Digital Future Institute research facilities at Temple Quarter, providing a rich environment of real-world research and experimentation, advance telecommunications technologies, open data platforms, powerful data centre infrastructure and immersive technologies.

What is it?

The new Infrastructure is split into one Data Centre and two comms rooms. The Data Centre hosts compute, data and advanced network services and applications, whilst the comms rooms are the gateway of all network connections internally and to the wider campus and beyond. Both the Data Centre and comms room have approximately 30-40% headroom / growth allowed for any future expansion.

Why is it special?

The Data Centre will provide a significant compute and storage facility, with a 5G and ultra low latency optical network backbone. It will be connected to a national dark fibre network, thus extending its ability and facilities to a wider audience of researchers and partners. The solution provides a private cloud within our facility allowing us to grant private research “tenancies” for experimentation that are entirely isolated from each other.

As well as a significant amount of CPU processing, the Data Centre will also provide a large amount of cutting-edge GPUs and FGPAs processing power, allowing it to run advanced Digital Twinning, and Reality models for users in Realtime.

In addition to this BDFI and the Smart Internet Lab will provide a KPI measuring solution allowing users to make interventions and measure the performance gain or penalty of the underlying services.

How does it work?

The Data Centre has 14 racks at current, with room for four more in the future. 10 Racks for BDFI and the Reality Emulator and four racks for MyWorld.

The racks, including power, are designed to allow hyper dense compute and storage services to reside within.

The Data Centre and comms rooms support the latest standards in high availability. All services and distribution paths are redundant, including redundant power supplies and generator. The Data Centre also has an all fibre backbone, allowing for exceptional high data throughputs and low latency.

The Data Centre has a remotely managed monitoring solution, with over sensors within the Data Centre monitoring cooling performance, availability, power usage, battery backup, early waring fire and smoke detectors, water leakage, humidity and an overall temperature map.

The racks are all water cooled, and the heat gained by the water-cooling system will be extracted via a heat exchange and used to heat the buildings.

What difference might it make?

The University’s commitment to sustainability took a significant step forward this year with a £2.5 million grant to embed carbon reduction technologies and practices at the new Temple Quarter research and innovation. Thanks to this new funding a comprehensive range of environmental measures will be introduced in the new home for BDFI, significantly decarbonising the energy use of the Institute’s datacentre and buildings.

An artificial-intelligence-driven Data Centre energy management system to reduce energy consumption of the Reality Emulator and its datacentre hardware. The use of sensors and our ability to analyse vast amounts of data will allow for quick, automated decisions on those parts of the system that can be put into sleep mode, shut down or turned on again.

Sensors for real-time data monitoring of building occupation and status, and heat metering is used to provide heat, temperature and flow rate data from the Data Centre chillers to export heat elsewhere.

Get involved

The datacentre itself is an experimental platform, meaning researchers will be able to bring along experimental hardware and perform experiments / research. The network infrastructure is all oprtical which will allow for cutting edge network research, such as using Quantum Key Distribution and even adding neuromorphic devices into the fabric.

All aspects of the Data Centre service will use open standards, allowing researchers to bring in existing solutions and take advantage of the scale of compute and network resources the Reality Emulator will supply.

Edit this page