James Kobielus of Forrester just put out a fantastic post on how social network analysis will be the reason for petabype data warehouses. He makes a convincing argument on how data sizes will actually double every few years.
From our standpoint, the key question is how to balance the analytics system such that not one part of the architecture becomes the bottleneck. Sure you can place lots of data onto cheap disks, but if you only have those disk controller and you want to analyze the social network then you are limited by that those disk controller, even if you have 100 CPUs.
Tags: data mining, Marketing, Social networks
