Johannesburg, 17 Jul 2013
Dimension Data, the $5.8 billion global ICT solutions and services provider, today announced independent testing carried out in May 2013 by The Tolly Group reveals that Dimension Data's Public Compute-as-a-Service (CaaS) cloud servers showed consistently high performance across CPU, memory, storage and networking measures.
The report shows dramatic differences between providers and highlights the importance of performance in determining the true cost of a cloud computing solution.
The Tolly Group has been testing systems and networking infrastructure for over 25 years, and uses industry-standard testing methodologies and tools to benchmark system and networking performance. In this test, commissioned by Dimension Data, the Tolly Group tested the system performance and network throughput of US-based Web/app servers running on the public cloud platforms of Dimension Data Public CaaS, IBM Smartcloud, Amazon Web Services and Rackspace Cloud using three different cloud server configurations: small (1 vCPU and 2GB RAM), medium (2 vCPU and 4GB RAM) and large (4 vCPU and 8GB RAM) systems.
The configuration details, test set-up and methodology are documented within the report. The testing showed performance benefits of Dimension Data's cloud solution versus the other solutions tested.
The Tolly Group benchmarked cloud servers using publicly available solutions from each vendor. The report shows that when compared to other cloud vendors, Dimension Data's Public CaaS provided:
* Faster processing: Dimension Data showed consistently high performance across configurations. In particular, when conducting a test on small systems, the market-share leader took 50% longer to complete a series of compute-intensive tests.
* Faster memory throughput: When measuring system memory performance on large systems, Dimension Data delivered more than twice the memory performance, as measured in memory operations per second, as the nearest competitor, and over five times the performance of the market leader.
* Faster file access speed: When measuring average transactions per second on a local disk using the default file system, Dimension Data delivered 3.6 times the market-share leader on small systems and more than six times one of the solutions tested on large systems.
* Greater network throughput: When measuring the bi-directional network throughput between two cloud servers on the same internal data centre network, the report found that Dimension Data was the only provider that delivered true Gigabit Ethernet throughput.
Commenting on the results, Steve Nola, Cloud Business Unit CEO at Dimension Data, says: "Our cloud was built to support more than just testing and development or non-critical applications. Our Public CaaS clients use our cloud for running production Web and software-as-a-service applications, as well as enterprise applications like SAP. Our unique network-centric architecture ensures security, reliability and high performance by leveraging best-of-breed technologies and best practices. The performance of a cloud provider can have a big impact on the quantity and size of cloud servers organisations need to run an application."
Says Kevin Tolly, founder of The Tolly Group: "The Tolly Group thoroughly tested the cloud solutions across the parameters that matter most: system performance and network throughput. Across the board, Dimension Data's solution illustrated significant performance benefits among the solutions tested."
Puneet Pandit, President and CEO, Glassbeam, says: "Our blue chip clients require strong SLAs, backups and expect their systems to be up and running when they need them most. They generate large volumes of data and rely on scalability on demand. Dimension Data has provided us with the enterprise-class cloud performance our clients demand."
An IDC report, titled "Dimension Data: A Network-Centric View for Cloud Professional Services" by Gard Little (Research Director, IT Consulting) and Leslie Rosenberg (Research Manager, Network Lifecycle Services), says: "IDC believes Dimension Data offers a differentiated network-centric view of cloud and is using its professional service offerings to highlight its position. Strongly rooted in network engineering, Dimension Data has applied this filter on its approach to offer cloud services to its global enterprise customers."
To obtain a free copy of the report, visit www.dimensiondata.com/tollyreport.
Note to editors
There were some differences in system configuration between service providers, and not all providers had a small system option to test. Ubuntu 10.4 was used as the operating system tested. Since IBM does not offer Ubuntu as a standard offering in its public cloud service, Red Hat EL 6.3 was used for IBM. For Amazon, m1 instances were used for small, medium and large systems. All Rackspace Cloud Servers tested were their Next Generation servers. The Tolly Group used the open-source Phoronix Test Suite (PTS) to benchmark CPU, memory and file capabilities. iPerf was used to benchmark network throughput.
Share