This article originally appeared in the Feb. 2011 issue of INTERNET TELEPHONY Magazine.

The topic of carrier-grade virtualization first emerged back in 2006. Granted the focus at that time was on carrier-grade Linux, but the seeds were planted for the future of carrier-grade virtualization. Today, the migration to more packet-based services such as voice over IP and high-definition video are finally making carrier-grade virtualization relevant. The latest carrier- grade platforms now have more processing cores, memory, and I/O than ever before, but can telecommunications original equipment manufacturers realize the benefits of using carrier-grade virtualization software? The potential benefits are reduced hardware infrastructure costs, improved time to market, and better energy efficiency.

Service providers rely on carrier-grade redundant systems to provide high availability for communications equipment in their networks. To still meet the demands of the network and capture the full benefits of carrier-grade virtualization, the OEM’s virtualization software must meet three critical properties. The first is performance. The software must have low overhead and operate in real time. The second property is availability. Fault isolation and high availability management must be present to support watchdog, error handling and notifications. Third is serviceability. The health monitoring of virtual machines must provide advanced warning of VM failure events. But just having highly available virtual machines is only part of the solution. Adding elastic provisioning to virtual machines allows the service provider to realize improved energy performance during low usage times.

According to Austin Hipes, director of field engineering at NEI (News - Alert), “through intelligent use of carrier-grade virtualization solutions and elastic provisioning techniques, OEMs can create highly scalable platforms and eliminate unnecessary over provisioning of resources for peak usage. Current advances in multi-core processors, packet processors, and high throughput Ethernet silicon make it possible to consolidate what previously required multiple specialized carrier-grade server platforms into a single private cloud.”

Final Score:

In 2011, we will see another leap forward in the number of cores available to software developers. Combining elastic provisioning with highly available virtualization software will allow OEMs to develop carrier-grade virtualization platforms with the level of performance, availability, and serviceability needed for next-generation carrier networks.