Blazegraph GPU
Date:
By: Philip Howard
Classification: InDetail
There are relatively few companies that specifically address graph analytics as opposed to analytics within the context of operational applications, which tend to be much simpler and at smaller scale. Of those that do address these sorts of problems some (for example, Teradata and Oracle) are limited to environments where relationships are known in advance. In other words, these products are not suitable for discovery, data mining (or science) and other complex analytic problems. Of the remainder, these are either extremely expensive and out of the practical reach of most organisations or, in some instances, vendors have attempted to build low cost solutions but these have failed the test of performance and scale. To summarise, prior to the introduction of Blazegraph GPU there was no affordable product on the market that could analyse very large graphs with appropriate performance. From this perspective Blazegraph GPU is in a class of one.