As more IT organizations start to become conscious of the implications of in-memory computing appliances, a lot of people are starting to wonder what the data center of the future is going to look like. The answer is a lot smaller than it does today.
What will drive that shrinkage is going to be the collapse of a few tiers of computing that is typically associated with data warehouse applications. In the place of storage-intensive traditional servers that come loaded with disk storage that takes up a lot of space while consuming lots of energy, we're going to see much smaller in-memory appliances running our data warehouse applications.
Oracle this week signaled its intention to compete aggressively in this emerging sector with the launch of an Exalytics platform that can run the company's namesake database, the TimesTen database and the Essbase online analytical database in memory. That move follows SAP's release of a High Performance Analytics Appliance (HANA) platform, which in turn followed a Smart Analytics appliance from IBM. Somewhere lurking out just over the horizon is an in-memory instance of Microsoft SQL Server, codenamed "Denali."
According to Sanjay Poonen, president and corporate officer of SAP Global Solutions, in-memory computing appliances will be rapidly adopted because they not only make analytics applications run faster at a much lower cost, they also enable the development of a whole range of new real-time applications.
Naturally, SAP dismisses the Oracle announcements this week as an attempt to retrofit decades-old TimesTen and Essbase database technology on a modern processor architecture. HANA, by contrast, has been built from the ground up to optimize the performance of the database technology that SAP embedded in it.
In addition, Poonen notes that HANA platforms are based on x86 server platforms that are being developed in conjunction with partners such as Dell, Hewlett-Packard and IBM. Of course, IBM is betting on HANA, its own appliance and the ability to run analytics in real time on the mainframe. Oracle, in contrast, is basing its in-memory strategy on Sparc processor technology.
It's too early to say who will win the next big server war in the enterprise. But it's clear that the dynamics of the data center are about to rapidly change with the rise of in-memory appliances that are not only capable of processing terabytes of data in a matter of seconds, but also compress that data in a way that greatly reduces the amount of storage required to support it.