SHARE in Pittsburgh

16209: Sicoob’s Case: Linux on System Z

Tuesday, August 5, 2014: 3:00 PM-4:00 PM
Room 404 (David L. Lawrence Convention Center)
Speaker: Claudio Kitayama(SICOOB - Sistema de Cooperativas de Crédito do Brasil)

The server consolidation and virtualization in Sicoob, in accordance with Strategic IT Planning, were impelled by a necessity of a flexible architecture that would maintain and broaden the processing power. As well as attend the expansive credit cooperative growth in Brazil, associated to Brazilian financial market dynamism. This also required operational systems that could follow the market and technological growth.

This tendency was to replace blades for middle-range servers, mainframes and enterprise servers, and also creating a new virtual structure that would adopt an operating system which was common to either Mainframe or x86 (distributed through VMware platform) environments.

The chosen operating system that met consolidation and virtualization requirements, and reduced software licensing costs as well, was Suse Linux Enterprise Server (SLES). This platform allowed Sicoob to achieve the required flexibility to support its business needs and IT infrastructure.

This flexibility provides many benefits, such as increasing the availability, scalability and security of the IT environment. Besides that, it also reduces points of failure, taking into consideration that any of the environments may become unavailable, either due to scheduled maintenance or hardware problems. This way the business core wouldn’t become unavailable.

Using Linux on System Z, the performance gain was noticeable. It demonstrated the efficient use of the computational resources, which reflected directly on the application servers (Java, Websphere and Jboss) and high-speed connectivity database servers based on DB2.

One case that reflects Suse Linux’s high performance was the SLES 10 update to version 11 on System Z environment that supports Sicoob’s business. This requires high processing power, such as compensation checks.

Before the upgrade, thirteen servers were used to process one million check images every day, operating CPU and memory on its high capacity usage. After updating, the check image processing performance improved 25%, even though the number of server had dropped to nine machines. The current structure holds twice the processing check image capacity. (Benchmark test).

Financially, the processing improvement meant 25% less IFL mainframe usage, which represents U.S.$1,354,808.05 of savings, that yielded on licensing investment return.

Tracks: Linux Systems Management and User Experience
Share |




See more of Project: Linux
See more of Program: Linux & VM