Intel® SBC, Powered by Dual-Core Intel® Xeon® Processors, Shows 2.8X Improvement in SIP Performance for IMS

2.8X Platform Performance Improvement for IMS
Service providers are working feverishly to find their own unique set of “must-have” services, such as multimedia on cell phones, network-based gaming, and IPTV. Service providers need a robust, scalable infrastructure on which to deploy these new services while at the same time controlling costs. The IP Multimedia Subsystem (IMS) standard is quickly moving to the forefront as the best alternative to meeting these challenges.

Because of the heavy technical demands placed on an IMS, Intel has developed its third-generation Intel NetStructure® MPCBL0040 Single Board Computer. Using new multi-core technology based on Dual-Core Intel® Xeon® processors LV 2.0GHz, this SBC provides a 2.8X platform performance improvement over its first-generation Intel NetStructure® MPCBL0001 Single Board Computer, creating an ideal platform for IMS infrastructure deployment. Intel has documented the 2.8X performance improvement in compute- intensive Session Initiation Protocol (SIP) call processing, which is the basis of an IMS environment.

Key Factors: Parallelism and Cache
Although many factors affect performance, two key factors come into play when explaining the significant performance improvement between the MPCBL0040 and the MPCBL0001. Hyper-Threading Technology in the MPCBL0001 creates two threads on two logical processors within each physical processor package, while multiprocessing in the MPCBL0040 creates two threads on two complete execution cores in each dual-core processor package. When thread stressing occurs (competition for resources among threads), the MPCBL0040 with Intel® mobile microarchitecture and multi-core technology provides a significant performance gain. Cache architecture and memory access affect performance, as well. In the MPCBL0040, Level 2 (L2) cache is shared by the two cores in each processor package. Because of this shared cache, threads can pre-fetch for each other, greatly improving memory access and performance.

Performance Test Result Summary
To measure the performance improvement, Intel engineers compared the MPCBL0001 and MPCBL0040 in a compute-intensive IMS environment using Busy Hour Call Attempts (BHCA), a common measurement for call processing performance. BHCA is the number of times a call (or session) can be attempted during the busiest hour of the day in a simulated call environment, and the object of the measurement is to record the maximum number of calls a given platform can handle in that busiest hour. All attempts were SIP calls, because IMS uses SIP to handle calls.

These performance tests demonstrated that the MPCBL0040 can handle 2.8 times more SIP messages. See Figure 1 for a comparison of the raw number of BHCA. Performance per watt is an even more important measurement for service providers because power represents a significant portion of the total cost of ownership of a system (Figure 2). When normalized per watt, the MPCBL0040 shows a 3.2 times improvement over the MPCBL0001.

Note: Performance tests and ratings are measured using specific computer systems and/or components and reflect the approximate performance of Intel products as measured by those tests. Any difference in system hardware or software design or configuration may affect actual performance. Buyers should consult other sources of information to evaluate the performance of systems or components they are considering purchasing. For more information on performance tests and on the performance of Intel products, please visit intel.com/performance/resources/benchmark_limitations.htm.

The complete white paper, published by Intel, summarizes the results of a series of benchmark tests comparing SIP performance of the Intel NetStructure MPCBL0001 and MPCBL0040 SBCs. To read the full paper, please visit: intel.com/design/telecom/papers/9936.htm.