We are delighted to announce that 2bm are finalists at this year’s DCS Awards.
Whilst we have a wide range of case studies available, it was our Swansea University Bay Campus data centre modernisation project that has been shortlisted for an award.
We would really appreciate you taking a few minutes to vote for us. Simply use the drop-down menu in subcategory to select ‘Data Centre Consolidation/Upgrade/Refresh Project of the Year’, then select 2bm and vote!
The deadline for voting is November 26th.
Award: Data Centre Consolidation/Upgrade/Refresh Project of the Year
SWANSEA UNIVERSITY BAY CAMPUS DATA CENTRE MODERNISATION PROJECT
BACKGROUND TO THE PROJECT
2bm was the successful bidder for the Swansea University Bay Campus Data Centre modernisation project located within the existing Bay Campus Data Centre. The basis of the project was the relocation of an HPC data centre from a facility three miles away to their on-campus Bay Data Centre, consolidating the university’s IT footprint and facilities. Also, the university required a refresh and update to the existing facilities so they would be future-proofed and capable of accommodating the wide range of computing used by the university and its various research faculties.
As well as meeting Swansea University’s key objectives for a resilient, reliable and high-quality facility for critical equipment, we also considered ways in which to contribute to the sustainable development of the local environment, economy and community in which a facility would be located. This included addressing any environmental and energy concerns as part of our decision-making process, designing and building a data centre that was both future-proofed and sustainable. In addition, we acted in a socially responsible manner when it came to our supply chain and the management of our internal operations and logistics.
2bm designed a solution that had both a high degree of fault tolerance as well as a high level of resilience on both the electrical and mechanical systems – ensuring a high level of expected uptime. Alongside the importance of a resilient solution, we aimed to deliver one of the most energy-efficient facilities in Europe, in-line with 2bm’s own data centre design ethos to always explore ways to reduce mechanical and electrical power consumption and, ultimately, lower the OPEX of a data centre.
WHAT EXACTLY IS NOVEL ABOUT THE PRODUCT/PROCESS OR INNOVATION?
In designing a solution which consolidates two data centres into one, we not only delivered the most modern, efficient and highly resilient solution possible but also maximised the existing space as well as providing the opportunity for future expansion.
The modernised facility, which was delivered in a live working environment with zero downtime, was designed to house both HPC racks (SuperComputing Wales/Engineering) circa 25 kW per rack and ‘standard’ density racks, in a fully room-neutral setting. In addition, our solution allowed any individual rack to be suitable for HPC at an IT load of up to 27 kW, whilst maintaining a temperature-neutral environment, cooling all IT heat at the source and eliminating any hot spots within the facility.
Our solution incorporated 100% free cooling with a compressor-less design which meets ASHRAE guideline recommendations. The refresh also included the reuse and recycling of existing equipment from the redundant data centre, all of which helps to meet the University’s request for a low carbon, environmentally-friendly facility.
HOW DOES THE PRODUCT OR PROCESS BREAK WITH CONVENTIONAL IDEAS OR PROCESSES IN ITS FIELD?
During the initial consultancy work we undertook on behalf of the university, we identified areas of risk and concern in bringing together HPC and standard equipment in a single environment. What the consultancy work also successfully identified, were those areas that would meet the required criteria for the consolidation project.
By introducing the HPC computing to the existing data centre, we modified the environment to consolidate the two facilities. This not only enabled the university to better utilise its existing assets, but also removed the rental and operational costs associated with the redundant facility. The resulting cost savings meant the university no longer had to consider the option of co-lo facilities, a compromise that would have been extremely expensive when trying to accommodate specialist computing equipment.
The new mechanical system, at a minimum N+1 redundancy, provided the necessary resilience together with an element of fault tolerance, with all control systems dynamic and standalone, ensuring there were no single points of failure. As this system avoids the need for any mechanical cooling, it meant we could take full advantage of the introduction of an adiabatic process – a technology we have deployed on several occasions including our 2019 DCS award-winning cooling solution. This provides ultra-low PUE and OPEX to our client, whilst still maintaining the environment within ASHRAE recommended guidelines for 100% of the year.
Electrical works included the upgrade of the existing UPS to provide an N+1 resilience against 400 kW of load, complemented by batteries that were upgraded so they could maintain at least five minutes of autonomy at the same 400 kW of load.
All rack locations were designed with the capability to move to 27 kW of load, with the theoretical limits of each section being a total design IT load of 35 0kW. Which means the entire room is completely flexible as to where the load is located, currently with 14 racks at 25 kW and 24 racks ranging from 1 kW to 6 kW.
Despite working in a live environment with zero downtime, the power and cooling were successfully upgraded from fixed-speed DX cooling with NO free cooling to a hybrid chilled water system. This meant we were able to install the two systems side by side whilst we phase-swapped over the systems. All the existing racks were relocated and fitted with new rear-door cooling units.
HOW DOES IT GO BEYOND MARGINAL IMPROVEMENTS ON SOMETHING THAT ALREADY EXISTS?
The original data centre had a PUE in excess of 2.0 and, due to existing cooling technology, no HPC could be installed in that environment. With 2bm introducing new cooling technology as part of its proposal, it was able to reduce the PUE to below 1.15, creating a world-class, market-leading solution.
The Bay data centre had a fixed power feed size that had limited the IT system’s maximum deployment, as any remaining energy was required for the cooling system. With the introduction of our low energy cooling solution, the University could now use the energy saved to power additional IT and, in doing so, increase the facilities computing capacity by up to 60%.
Bringing the HPC equipment on-premise, reduced latency within the overall IT service, as well as the risk from external interference. For example, in the event the internet went down it ensured all internal communications would remain unaffected and also that the facility could be utilised to its maximum.
By closing the legacy data centre, the University was in a position to make further savings in terms of removing the rent, energy and operational costs associated when running a second building.
HOW DOES THE CUSTOMER BENEFIT FROM THE PRODUCT/PROCESS OR INNOVATION?
Swansea University now has a facility that it can be proud of for the next decade and beyond. It is a facility that is already regarded as being one of the most efficient data centres in Europe, with a 250 kW load that requires only 5 to 6 kW of power to cool it!
In addition to now having a greatly reduced carbon footprint, Swansea University enjoys energy savings in excess of £185,000 per year, which significantly reduces payback periods and also saves 888 tonnes of carbon. And, having previously been co-located in a non-university environment over three miles away, the new facility is now conveniently located on-premise, enjoys far better security and is far more accessible for staff, visitors and research students.
The University also has the opportunity to utilise the facility to its maximum as well as save money on rent and travel, lower latency and significantly lower running costs.
By creating a mixed environment with compressor free design and working to ASHRAE guidelines for new and legacy equipment, means the 2bm solution means the University can apply for CEEDA accreditation – keeping it in line with the University’s wider environmental strategy and policies.
And because the university can fully utilise the facility, it is able to maximise its machine learning and support for the wider institution through its advanced learning and study of artificial intelligence and IoT.
The upgraded equipment’s extensive plant monitoring was installed and configured to the campus-wide BMS system, this allows the IT team to be alerted day or night in the event there are any issues within the facility. The system has additional back-up in the form of 2bm’s ‘24/7 emergency response service’.