New 100Gbps bandwidth to support data intensive research
Imperial College London are the first University in the UK to have a connection speed of 100Gbps!
Imperial College is largely acknowledged for their excellent connection to external networks through the Joint Academic Network, also known as Janet. The network has enabled and supported much of the successful, world-leading research accomplished through the college.
Pressure on this connectivity has increased, through inter-institutional collaboration, which involves the processing of ever-larger datasets that are accessed and shared widely. The College’s current external connection stands at 2x20Gbps to Janet. With a basic load of 8-10Gbps for “normal business”, this has ultimately left an estimated 30Gbps for research data exchange.
To date, the bulk of this (if not all) has been utilized by the High Energy Physics Group and only recently has the network capacity become a limiting factor for their data processing. HEP’s expectation is that they will be seeking 80Gb/s when back online in one year, a figure which will double in two years and double again two years after that. To maintain our pre-eminent position, network capacity needs to maintain pace. Various departments are also beginning to require capacity and so the demand for bandwidth will continue to increase.
Imperial approached Janet with a request to increase our connection bandwidth. Taking into consideration the department’s inclination to support research data exchange, Janet were keen to facilitate the delivery and application of this increased capacity. The project has therefore been structured to transform the College’s ability to support high speed data transfers and exchanges between external institutions in the academic community.
The project holds the following goals:
- Increase the current connection bandwidth from 2x20Gbps to 2x100Gbps.
- Absorb the increased capacity into the College infrastructure to deliver to research groups.
- Develop the concept of Research Data Transfer networks, exploiting the new methods of working that are enabled and encouraging new groups to participate.
This project will provide the following benefits to Imperial College:
- Ensuring the College academic community is on a competitive trajectory to conduct world-leading data-intensive research with both internal and external collaborators in the UK and internationally.
- Building the infrastructure required to ensure Imperial College has a competitive advantage for future data requirements.
- Addressing the challenge of linking increasingly affordable but high data volume networked scientific equipment, such as electron microscopes and gene sequencers, to national centres for processing and long-term storage of data.
- Enhancing the ability of remote scientists to carry out ‘real-time’ research activities such as remote experiment control at national and international experimental facilities, much like Diamond Light Source and remote telescopes and observatories such as SKA. Rapid access to data outputs enable efficiency gains that are obtained through the ability to identify and resolve errors or make parameter adjustments during the experiment window.
- Delivering the bandwidth required to support High Energy Physics, to ensure that they are able to collect the data they require, as the current bandwidth may not be able to transfer as much data as they are expecting to receive in 2021.
- Providing an open source high-capacity data transfer toolkit that will ensure its users can easily exploit the UK RDTZ infrastructure to its full potential.
- Making regional and national research facilities more accessible and hence, more attractive to researchers, increasing quality of research and collaborative opportunities.
How you can get involved
- We are looking for a few volunteers with data in Diamond Lite to participate in our proof of Concept and transfer large data between both sites.
- If you believe you have a strong demand for large bandwidth, please get in touch with Matthew Harvey who will discuss your requirements.