Wind River, in its role delivering software for intelligent connected systems, claims leading performance with its accelerated virtual switch (vSwitch) integrated within Wind River Carrier Grade Communications Server, which is designed for network functions virtualisation (NFV). Carrier Grade Communications Server is, WR asserts, the only commercial platform delivering six 9s (99.9999%) reliability for telecom networks, delivering up to 1,000x uptime improvement compared to enterprise-class solutions. The accelerated vSwitch can deliver 12 million packets per second to guest virtual machines (VMs) using only two processor cores on an industry-standard server platform, in a real-world use case involving bidirectional traffic. This performance represents 20 times that of the standard Open vSwitch (OVS) software used in typical enterprise data centres. Providing unlimited scalability when instantiated on multiple cores, this performance is achieved using up to 33% fewer CPU resources than other commercial solutions, with no requirement for specific hardware acceleration.
As service providers adopt NFV architectures in order to reduce network operating expenses and accelerate the deployment of new value-added services, the vSwitch plays a key function within the infrastructure platform that runs virtualised network functions (VNFs). The performance of the vSwitch determines the bandwidth of network traffic processed by the VNFs and directly influences the overall cost performance of the platform. Since service providers will ultimately derive revenue from services based on virtualised applications, it is important to minimise the processor resources consumed by the vSwitch function and maximise the resources available for VNFs.
The accelerated vSwitch within the Wind River Carrier Grade Communications Server can switch 12 million packets per second (64-Byte packets) using only two processor cores within a dual-socket Intel Xeon processor platform running at 2.9 GHz. This measurement is based on bidirectional network traffic running from the network interface card (NIC) to the vSwitch, through a VM and back through the vSwitch to the NIC. This represents a real-world NFV configuration, rather than a simplified configuration where traffic runs only from the NIC to the vSwitch and back to the NIC (bypassing the VM, so that no useful work in performed). The performance of the vSwitch is fully deterministic and scales linearly with the number of processor cores allocated to run the function, providing the scalability required for NFV deployments with ongoing changes in bandwidth requirements.