1. Ginormous Growth of Bio IT good news/bad news for data centers - By Doug Mohney

    Views and Opinions on Green IT (Dec 8 2010)

    1. Ginormous Growth of Bio IT good news/bad news for data centers  - By Doug Mohney

      An evening at the Hudson Alpha Institute for Biotechnology (www.hudsonalpha.org) in Huntsville, Alabama last week left me ready to laugh and cry.  

      One of things Hudson Alpha does is sequence the building blocks of life - DNA, genomes, all that good stuff.  A sequenced human genome consumes about 8 GB of data. With today's latest technology, it roughly costs about $15,000 to $20,000 to do and takes around a week or so to finish, but there's a big push to drastically cut cost and increase speed of sequencing. 

      The target goal of the $10 million dollar Archon X prize is to sequence 100 human genomes in 10 days at under $10,000 per genome, but the bio community feels pretty confident that it will hit a target of around $1000 per person and do it within a day or two.  The advances here make Moore's Law look like a snail, since at the beginning of the decade it took about $3 billion and about thirteen years to sequence the first genome.  With not-so-crazy-talk that sequencing may dip into the hundreds of dollars in the near future, everyone will ultimately get their genome sequenced and the human race will see explosive advances in medical treatment.

      Good news: Better health care.  Bad news: Huge amount of data, requiring advances in storage, access times, computations – the works.

      How big is big?  For starters, the building LAN runs at speeds of 10 Gbps -- a speed that many entire corporations and data centers are just starting to use for WAN connections.  Storage? It's all about multi-petabyte sized disk arrays -- and since one petabyte is 1,000 terabytes, imagine how many racks of 3.5 inch disk drives that means.

      If your head doesn't hurt yet, now think of backing up all that data to a safe place. 

      And that's even before you start mining all that data in statistically significant quantities by comparing sets of 10,000 to 25,000 people and cross-checking for specific diseases and conditions and the potential that certain drugs/treatments work better than others.

      NOW does your head hurt? 

      Sure, there are other scientific applications that fall into the petabyte range of data, but genome crunching is one of those classic/cliche killer apps that universities, drug companies, and even nations are lining up behind for the potential to improve the human condition and make a whole lot of money.

      Did I mention that Hudson Alpha has a pair of 1 megawatt backup generators when the main power goes out?  There's no fast green solution on the horizon for genome crunching and the hockey-stick of growth is coming as fast as the price of sequencing is falling.

      -->

      -->

      Bookmark or Share this article

    Login to comment.