Ten Things You Need to Know About Software-Defined Storage
Slowly but surely the compute, storage and networking elements that make up enterprise IT have been converging as part of an effort to both boost performance and lower the cost of IT.
Pluribus Networks took that idea one step further today with the announcement of the general availability of Netvisor 2.0., an implementation of a network hypervisor that combines an X86 server and switch functionality into a single Pluribus Freedom server-switch platform.
The Pluribus Freedom architecture eliminates the need to rely on network integration cards (NICs) to connect servers to the network. Instead, commercial application-specific integrated circuits (ASICs) are used to tightly couple network switching within the server environment.
Pluribus Networks CEO Kumar Srikantan says this approach not only boosts overall performance of the data center environment, it also lowers costs by making it simpler to manage servers and the network under a common framework.
One of the other immediate benefits of this approach, adds Srikantan, is that the Netvisor running on the Pluribus Freedom platform eliminates the need for separate appliances to deploy network monitoring applications. All the tools needed to monitor the environment are baked into the network operating system.
Applications for the Pluribus Freedom platform can be developed using a Freedom Development Kit (FDK) that makes it easier to create applications in Java or C using a series of Unix-style application programming interfaces that enable them to build applications.
A massive amount of disruption is taking place at the network layer in the form of everything from network virtualization to software-defined networking (SDNs), but the ultimate impact of these technologies goes well beyond the network. Instead of thinking in terms of compute, network and storage resources that all must be managed across different silos, convergence is leading to a level of integration that will ultimately blur the lines between what were once considered separate IT disciplines.
As that process continues, the total cost of managing enterprise IT as we know it today will decline, and the latency issues that have prevented certain classes of applications from attaining higher levels of application performance are starting to disappear.
