Intec, one of the world’s leading providers of Business Support System (BSS) solutions, recently announced the completion of another successful performance benchmark of its award-winning charging, billing and customer care system – Singl.eView v7.0. The performance results reaffirmed Intec’s commitment to providing software that offers exceptional performance on platforms that are a fraction of the cost of traditional UNIX servers. Further, it demonstrated that Singl.eView, running on Intel® Xeon® processor 7500 series, easily supports millions of convergent prepaid and postpaid subscribers, executing real-world service and business processes.
The benchmark, carried out by Intec on Intel’s latest Xeon® processor 7500 series, validated the performance of Singl.eView’s critical functions:
• Real-time rating and balance management
• Batch post-event rating
• Bill cycle processing
Singl.eView is relied upon by many of the largest and most innovative communications service providers (CSPs) worldwide. Singl.eView’s efficient, distributed architecture allows CSPs to lower their operating costs by supporting all rating, balance management, charging and billing tasks on a single platform, serving postpaid, prepaid and convergent customers for all classes of product and service. The Intel Xeon® processor 7500 series ushers in a new era in performance and scalability within the fast growing x86 server market. Singl.eView, running on the Intel Xeon platform, provides a powerful and cost-effective solution for today’s service providers, allowing an attractive price-performance ratio with more than adequate headroom to support the growing volume of increasingly complex demands of modern transaction processing.
Commented Wolfgang Petersen, EMEA Director of Intel Software & Services Group, “The results of this benchmark prove that BSS customers need no longer be dependent on expensive proprietary servers, and can be confident that Intel-based affordable processing solutions powered by the Xeon® 7500 series can easily manage highly-demanding workloads, and deliver the real-time performance necessary for a modern, converged communications service provider.”
The benchmark outcome demonstrated excellent performance and linear scalability across a range of real-time convergent rating, balance management, on-line transaction processing, and charging and billing tasks typical for a large service provider environment. It also showcased Singl.eView’s horizontal scalability, executing real-time and batch-based charging and billing scenarios typical of a convergent service provider.
“In today’s highly competitive markets CSPs are looking for solutions that will scale with their business, delivering consistent high-performance, at predictable cost, without sacrificing flexibility,” said David Heaps, Chief Product & Strategy Officer for Intec. “As evidenced by this benchmark, CSPs that use Singl.eView for their real-time charging and billing can be confident that they have chosen a solution which meets and potentially exceeds all these objectives.”
The benchmark was conducted in early 2010 at Intec’s Brisbane, Australia R&D facility with Singl.eView v7.0 on Intel Xeon® processor 7500-based servers running Oracle Enterprise Linux 5.3(1). The testing focused on various conditions for voice, data and content services, typical of today’s dynamic operator environments.
• When used primarily for real-time authorisation, just a single 4-socket (32 core) Intel Xeon® 7550 server demonstrated the capacity to handle up to 13 million high-usage subscribers at only 50% system load(2).
• When used primarily for postpaid billing, that same Intel Xeon® 7550 server demonstrated the capacity to handle up to 25 million high-usage subscribers(3).
For more information on Singl.eView’s Intel Xeon® 7500 series benchmark results contact [email protected]
(1) Singl.eView is also supported on Red Hat Linux 5.3.
(2) For a system in which prepaid subscribers are dominant, real-time authorisation requests can be performed at the rate of 13,000 requests per second using 32 cores at full CPU load. For real production loads, processing of the expected average busy hour load should not use more than approximately 50% of the available CPU capacity. This allows for additional OLTP load during the busy hour in addition to providing more capacity for occasional spikes in the real time request rate. Thus, 32 cores would be used to handle real-world production loads of around 7500 requests per second as the average busy hour rate. For real time request loads which are dominated by services like SMS, and assuming 1.5 Busy Hour Call Attempts (BHCA) per subscriber, this rate of 7500 requests per second corresponds to being able to process real-time authorisation requests from 13 million subscribers.
(3) For a system in which postpaid subscribers are dominant, billing of subscribers governs the system sizing. 32 cores were able to generate 315,000 invoices per hour at full CPU load. In a production system, 100% of the CPU can never be dedicated to billing. At 80% CPU utilisation, and assuming 15 bill cycles per month each with a billing window of 10 hours per cycle, the 32 cores allows for billing of 25 million postpaid subscribers, with 300 events/subscriber/month as the average monthly usage across the 25 million subscribers and which includes itemised call details on the invoice image.