The Broad Institute is relying more and more on the cloud for its genomic research, writes Steven Norton at the Wall Street Journal.
The Broad began building its cloud capabilities about four years ago, he says, after realizing that traditional computing methods would soon be insufficient. William Mayo, the Broad's CIO, tells him that, prior to the cloud, researchers relied on high-performance computing "islands," which led to both data duplications and technical differences in the computing environments. The cloud allows for processing large amounts of data and secure sharing of data among researchers.
According to the Journal, the institute has partnered with a number of companies like Google, Amazon Web Services, and Cloudera for its cloud projects, but it relies on Google Cloud Platform, though also others, for computing and storage. "The goal is full multi-cloud as we know not everyone will ever agree on a single platform," Mayo tells the Journal. It adds that the Broad handles about 70 petabytes of data.
The Journal says that the institute estimates that it now costs about $5 per genome to process it on the cloud, down from $45 when they made the switch. Mayo adds they hope to get that cost lower, to $3.