In-Memory Analytics

One of the hottest emerging trend in BI that holds great promise for this year is In-Memory Analytics. Imagine the entire data residing in the memory of the server, queries that do not require any disk I/O. This would definitely translate into the most dramatic performance improvement that a data warehouse could ever hope for. The truth is that In-Memory Analytics is not a new concept at all. We always knew that things would run much faster in memory, but the memory costs were prohibitive and a processor could only access so much memory in a 32 bit architecture. However, things have begun to change with the advent of low cost 64 bit computing (resulting in larger addressable memory space) and declining memory prices. As a result, in-memory analytics is no longer limited to small deployments and holds a promising future for functionalities where query performance is a key measure for e.g. dashboard analytics and performance management solutions. It may eliminate the need to store pre-calculated data in the form of an OLAP cubes or aggregate tables.