The Coming of Next-Generation Information Platforms

Michael Vizard

One of the thorniest issues in all of enterprise IT is the simple fact that applications take and hold our data hostage. With each successive application that is deployed, IT organizations wind up making copies of the same basic sets of data. The end result is huge amounts of data, the vast majority of which is either duplicate or simply no longer actively being used.

Of course, the IT infrastructure required to support all that mostly useless data needs to be acquired and managed. So in addition to huge amounts of mostly irrelevant data, enterprise IT organizations wind up hiring huge staffs of people to manage woefully underutilized servers and storage systems.

This situation exists because most IT organizations don't really manage data. Instead, they deploy one application after another and then focus on managing the machines the applications are running on. In order to simplify the management of the machines, IT organizations for the past several decades have been deploying one application per server alongside some dedicated storage, regardless of how underutilized that hardware actually was or how much power was consumed.

The good news is that there are two trends starting to converge that should put an end to all this. The first is the rise of virtualization, which makes it easier for applications to share infrastructure, and the second is the convergence of multiple types of application workloads within the same composite application. As enterprise applications continue to evolve, it's pretty clear that analytics, for example, will have to be embedded alongside transaction processing workloads. The rise of these applications is already influencing the design of next-generation IT systems, whether it's in the form of the Fusion accelerated processing unit (APU) and the Bulldozer Opteron processors from Advanced Micro Devices (AMD) or the distributed architecture that anchors the new zEnterprise mainframes from IBM.

But in order for this transformation to come about, Sybase CTO Irfan Khan correctly observes that IT organizations are going to have to really manage data. Allowing multiple copies of the same sets of data to be strewn across multiple applications is just not going to be practical and trying to resolve which set of data is the most recent and relevant would be nearly impossible. In response to this need, Khan says we should expect to see companies such as Sybase, a unit in SAP, developing next-generation information platforms where data is managed within a distributed grid architecture running in Level 2 cache that will then allow developers to create applications that can pull data from any source in the enterprise. In other words, a true n-tier architecture.

To make that vision a reality, the data itself is going to have to be a lot smarter. Khan says that a next-generation information platform will essentially act as a metadata server that keeps track of all the relationships between different sets of data. This new approach to managing data, says Khan, will need to incorporate principles from what is known as a Magnetic, Agile, Deep Analysis (MAD) framework through which sets of data will be able to continuously monitor the rest of the enterprise for updates to related sets of data by querying the information server.

The end game is to not only create more intelligent applications, but also fundamentally reduce the amount of useless data in the enterprise and along with that the actual number of machines we need to manage. It's our collective addiction to storing and processing data that makes enterprise IT so costly. This is why enterprise IT today resembles the auto industry from the 1990s. Everybody wanted the biggest SUV they could afford, but once the price of gas started to climb, you now rarely see a Hummer on the road. Similarly, as the cost of enterprise IT has spiraled out of control, we're beginning to see an entire industry starting to focus on efficiency, rather than celebrating who has the biggest IT infrastructure on the block.

It will take a few years, but the entire enterprise IT landscape is on the cusp of a profound transformation. And the IT leaders of tomorrow will be the people who saw these changes coming today.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.