Last week I noticed that a new world record had been set... and I'm afraid I'm not talking about the common wealth games, but rather about another advance in private cloud computing. The record was set ( or rather reset, as they already held it ) by Autonomy Corporation Plc, and the press release is here. So now there is a private cloud that manages over 17 Petabyts of storage space ( for Autonomy this is emails and multimedia etc. ). To put this in perspective, they say this is equivalent to "11 times the size of the 10 billion photos on Facebook, or 226 years of HD video" or alternatively, or as I'm tempted to think, one particularly good version of War and Peace... More seriously, this reminded me of a talk I heard recently at ECCB, where Elaine Mardis from The Genome Center at Washington University, in describing the compute setup at Washington University, mentioned that they have 6 Petabytes of storage space. Impressive in relation to the new record, but also very much putting into perspective the demands that the surge in biological data production is putting on compute resources.
So, as our company is a regular user of the public cloud in relation to biological data, this got me interested in the private vs public ( or indeed hybrid) cloud issue. There are many articles on this already, so here are just some of the more interesting ones:
Interestingly, it seems that a hybrid cloud approach is suggested a lot as the future approach. Though if you have some machines 'permanently' and expand to the public cloud as required, as we do, then in a sense your are already using a hybrid cloud approach.