Everyone is trying to build scale into their data infrastructure, but the question is: How much is enough?
To hear some vendors tell it, cloud computing, mobile traffic and Big Data require nothing less than your own private Google. But even if your aspirations are not that high (yet), you can build scale into existing infrastructure without requiring a complete make-over of the data center.
Nutanix, for example, is out with a new version of its Nutanix Operating System (NOS) Ver. 4.0, designed to turn today’s virtualized data center into a Web-scale operation. The fact is, many data environments are already equipped with the resources to provide significant scale, but they lack the ability to harness them in a coordinated fashion. NOS provides a broad set of system analytics and multi-cluster management tools that enable single-pane control over disparate resources. As well, new deduplication technology helps ensure that systems don’t become bogged down with redundant data and services.
Eucalyptus Systems has released a new version (4.0 again) of its software that boosts scale in private and hybrid clouds and delivers built-in compatibility with Amazon Web Services in case traffic exceeds local capabilities. The system includes scalable object storage that is suitable for commercial or open source environments, as well as a hybrid user console that allows users to manage their cloud environments from a PC, tablet or mobile device. The overall platform is designed to provide dynamic scaling capabilities by coordinating the needs of compute, storage and networking in an integrated fashion so as to avoid data overload in any one component.
Of course, at some point, the rubber has to meet the road and you have no choice but to expand physical resources. Fortunately, this is not the costly process it used to be. New low-power microserver architectures based on the ARM and Intel Atom platforms are making it possible to build out Web-scale resources quickly and easily. At the same time, they are helping the enterprise match both the underlying technology (the ARM is the most popular cell phone processor) and the small-bit, high-volume traffic patterns that characterize today’s mobile-based economy. Already, leading platform providers like Microsoft and Red Hat are tailoring their environments to new enterprise-class micro platforms.
This need to go big is not lost on the main platform providers as they manage the transition from yesterday’s silo-based infrastructure to the dynamic, scalable models of tomorrow. HP, for one, is pouring resources into its Helion platform to provide not only a scalable architecture for enterprise customers, but one that supports the growing OpenStack ecosystem on the cloud. A key component in the strategy is tight integration with many existing HP products like the 3Par and StorVirtual portfolios, allowing HP to maintain a footing in both the old and new data environments. HP also recently broadened its relationship with systems OEM Foxconn to develop new lines of commodity servers.
Scale may be the new black in the data center, but it is by no means the full ensemble. Agility, flexibility and reliability must be built into cloud-ready architectures or else the enterprise runs the risk of finding itself atop a mountain of resources with no way to control them.