In this video from DEMO Traction in San Francisco on April 22, 2015, Jeffrey Rothschild, VP of infrastructure software at Facebook, and Lance Smith, CEO at data visualization company Primary Data, explore big data’s impact on enterprise data centers and the new tools for managing it. (To watch the video, register or log in below.)
From the video:
One of the problems plaguing the data center today is the cost of over-provisioning. IT runs into problems that are unknown to them when it comes time to deploy a particular application, so they have to build in this over-provisioning at a cost that is sometimes quite overwhelming for some budgets.
The second problem is the mismatch between the compute and the storage resources. That mismatch is driven by the rapid pace of change from the type of data and the business requirements. Those change — when you first start out, it may be a very hot application, then drop off after some time, then get popular again.
The last problem is resource sprawling. We’re very application-centric in these architectures that get implemented in a data center. There’s resource sprawling where you design a system for an application and you bolt everything with it — the compute, the networking and the storage — and then you instantiate it multiple times. But what happens with that first one you implemented? It might drop off, it might get cold, it may not be used as much — but it’s still being utilized. All of a sudden you’re replicating and these data centers are getting much, much bigger than they should be.
To continue reading this article register now