Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > Networks

Examining the bottlenecks in virtualized networks

Posted: 30 Jan 2015 ?? ?Print Version ?Bookmark and Share

Keywords:Network Function Virtualisation? NFV? virtual machines? VM? VNF?

In November 2012, a handful of the world's leading service providers founded the ETSI ISF for Network Function Virtualisation (NFV). There are now over 220 participating companies. When surveyed, nearly all carriers worldwide are expecting to begin virtualizing network functions (VNF) within the next few years. Telefnica has declared that they expect to virtualise 30% of their network functions by 2016. While straightforward in theory, the move from network appliances to running network functions in virtual machines (VMs) on x86-based servers is far from trivial.

One of the requirements of running a function virtually is having "predictable performance." This is kind of like going car shopping and requiring any vehicle tested to go 0-to-60 in less than 5 s, get at least 35 miles to the gallon, have sports car handling, and stop on a dimewhile not restricting how many people are in the car.

Over the past two years, four key areas have arisen that create bottlenecks in performance for a virtualized function. By identifying them, tuning them, and tweaking them it's possible to get predictable performance in highly controlled environments.

1. Servers and NIC(s)
Not all COTS (commercial off-the-shelf) servers are created equal. The most fundamental aspects of the server is the CPU (brand, generation, number of cores) and system memory. An Intel E5 v2 has more than four times the packet processing performance of the older Xeon 5600. The NICs (network interface cards) are also critical in terms of performance when choosing 10GbE versus 40GbE NICs and ensuring the proper driver is used which is optimised for the OS and server. For example, a NIC without support for DPDK (data plane development kit) could have significantly worse performance.

Source: Intel

2. Hypervisor and the vSwitch
The hypervisor is used to virtualise the resources of the underlying server. If predictable performance is required, the resources must be strictly allocated but not highly oversubscribed. Some hypervisors do a better job than others in that they don't expose shared memory or cache, which could affect other VMs on the system. It may also be required to allocate CPUs from the same socket to avoid bus bandwidth issues. The virtual switch is also a critical component because it is providing the networking connectivity for the virtual machines. If the hypervisor is not accelerated or optimised for packet processing it could become the bottleneck. Also some overlay technologies require the virtual switch to perform encapsulation to tunnelling technologies like GRE or VXLAN (virtual extensible LAN) which can create significant overhead.


3. Host OS and guest OS
The guest operating system used by the VM/VNF may or may not be compatible or optimised with the host OS/virtualisation layer resulting in a communication bottleneck. It is important to know how this communication channel works and if it needs to be optimised for various workloads supported by the VNF.

The Virtual Network Function itself could be the bottleneck if it is not written properly to handle multiple cores and the memory allocated to it. In some cases the VNF provide will have specific limitations of what can be allocated to it. For example, it may be limited to support three cores and 2 GB of RAM. Some VNFs will have stated limitations like a vBRAS that supports 64k PPP (point-to-point protocol) sessions.

In conclusion, don't expect to drop a VNF on a COTS server and get wire rate performance. If the system is highly optimised and the resources are explicitly allocated, it is possible to get high performance and predictability.

About the author
Michael Haugh is the product marketing director at Ixia.

Article Comments - Examining the bottlenecks in virtual...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top