Today we have the fortune of announcing another amazing DDN customer deployment. In the heart of London, there happens to live the very forward-thinking and quite large research institution that we know as UCL, or the University College London. UCL has been a longstanding DDN customer and partner for many years and we’ve done several projects within their high performance computing (HPC) team to support the simulation and parallel processing applications which run on scalable cluster computers. About two years ago, UCL began to spur new and exciting discussions about an altogether different project they were considering—this was a project with the potential to both complement their computing efforts and dwarf anything they were doing in HPC storage in terms of scale.
As one of the largest academic research institutions in the UK, UCL recognized that the amount of data they were generating was exploding in volume and the university’s infrastructure for storing this data was not always being utilized. Academic research projects could span multiple partner institutions across multiple continents and could last for as many as 3-5 years. Once complete, the output and simulation data from these projects would often find their way to their final resting place—which could be anywhere from a university data storage facility, to a thumb drive, to a tape. This issue of research data retention and curation is no new challenge to researchers around the world… but the interesting thing is that the challenges stem, often times, from limitations associated with conventional storage models.
Enter the cloud. Enter massively scalable object storage. Enter my hippocampus analogy.
To support the needs of up to 2,000 researchers and extensive, data-intensive research projects, UCL realized they needed to rethink their approach to building a scalable information storage infrastructure that could capture and manage the extensive volume of institutional knowledge that would be amassed through its efforts. With cloud storage technology, underpinned by object storage tools, UCL now had the opportunity to build an altogether more-scalable and cost-compelling infrastructure to enable wholesale ingest and retention of institutional research data where data growth could reach 100PB of research data. What’s more, selecting the right platform presented the promise of opening new methods of data sharing to facilitate faster collaboration and accelerate the discovery process. Finally, all of the system needed to be serviced and supported by the existing information services team (consisting of four people).
Following a highly competitive procurement, UCL selected DDN’s Web Object Scaler (WOS) platform via our partner in the project (Dell) to begin the build-out of a cloud that could quickly approach 100PB in scale. The decisions behind their selection-making process were extensive—too long to discuss here. I’ll summarize the key benefits they gained from a very sophisticated solution that leverages the power of open source tools and the speed, scale and simplicity of storage for which DDN is synonymous.
- A single platform that features versatile cloud storage access to serve a one-platform-fits-all agenda: REST APIs, NAS Protocol support, iRODS open source data management and smart-device clients
- Integrated support with DDN’s GridScaler system, where their HPC simulation data can be transparently migrated to a WOS cloud storage platform
- The WOS platform ensures robust data protection that eliminates the risk of storage media loss, data corruption (aka: bit rot, silent data corruption), data format changes and technology obsolescence
- Built-in support for storage replication enables low-cost disaster recover
- Policy-based, self-healing storage infrastructure that enables the storage managers to avoid hair-on-fire trips to the data center
- A scalable architecture which can collect and manage all of the institutional research information of the college – i.e. the Hippocampus for the Campus
- And more… (see the case study here)
So – here I send a big thank you to UCL. Thanks for all of your hard work in making the decision to partner with DDN on the start of this journey with you. Thank you for trusting DDN with your institutional knowledge base. Thank you for embracing cloud technologies in a way that will showcase the power of scalable and portable infrastructure that is designed from the ground up for the (100) petabyte era.
P.S. Yes, I recognize the blog title is fairly nonsensical 🙂
Thanks for noticing.