AMS needed to create a new data-mart to give their users the in-house capabilities and freedom they needed to create their own ad-hoc queries. The solution would have to be easy to use and offer extremely high performance -- without soaring IT costs.
After conducting a proof-of-concept to test solutions offered by traditional database vendors and in-memory databases, AMS found that Sybase IQ stood out as the clear winner for two primary reasons. First, the system was easy to handle, giving users complete freedom to query historical data in a quick and straightforward way. Second, Sybase IQ did exceedingly well in performance testing, without any need to tune the system.
Complex Data Access Limits Freedom of Analysis
The business intelligence suite that AMS had been using since 1998 began limiting customers. Standard reports offered by traditional tools severely restricted expert users. Cross-referencing historical data was almost impossible with extensive use of intermediate calculations. Ad-hoc and detailed data reports had to be outsourced to AMS’ service provider. Developing these reports could take days -- with every hour of development costing AMS additional money.
The situation created unacceptable delays for users and extra expenses for AMS. The decision was made to implement a new data-mart, one which was optimally suited for such ad-hoc query requirements.
Sybase Finishes Ahead of Competition in POC Testing
AMS conducted proof-of-concept (POC) testing with solutions from three types of database technology providers -- traditional, in-memory and columnar store -- to find the right solution for its new data repository. Sybase IQ proved to be the clear winner in tests pitting it against traditional and in-memory databases.
AMS strictly carried out two types of tests on all three database technologies during the POC process. First, AMS user experts tested usability over a number of weeks. Second, meticulous technical testing measured ease of installation, speed of data loading and query performance.
“The results were clear,” said J. Ernst Oberklammer, AMS Head of Business Intelligence. “Sybase IQ was easy to handle and in performance testing it did exceedingly well -- up to 10 times faster than the other databases.
“Comparable performance was only possible on the other databases with a large amount of tuning,” Oberklammer continued. “Sybase IQ delivered these results with absolutely no tuning.”
Speed was also key for AMS. “Loading data into Sybase IQ took only minutes,” Oberklammer said. “While loading the same data into traditional databases was taking up to seven hours.” Sybase IQ was the clear winner in 85 percent of AMS tests. It was runner-up in only one case, and never once fell into third place.
Sybase IQ’s compression capabilities really impressed AMS during testing. Reducing data volume instantly showed that such compression rates would dramatically shrink storage needs -- and the company’s IT service provider charges per KB of hardware needs.
Deployment to Lower IT Operating Costs
The scheduled deployment of Sybase IQ as the new data repository for AMS will yield significant cost savings and enhanced performance. Ease of implementation was a key criterion during the initial POC testing. Deployment of Sybase IQ will apply additional tuning parameters to the database, realizing even greater performance than originally achieved during the POC.
Performance gains aside, AMS expects to reduce its overall IT operating costs. The company estimates that Sybase IQ will compress the company’s 3 TB of raw data by 80 percent, directly saving AMS $10,000 per year in hardware requirements.
Sybase IQ will remove the query constraints of the existing system. Users and AMS partners will have complete freedom to cross-reference historical data. And they will carry out ad-hoc reporting directly on large data sets. And AMS will no longer incur additional charges for outsourcing the development of such ad-hoc reports to their service provider.
Ultimately AMS is building a system that can expand over time to allow more users true freedom of analysis with fast response times, while maintaining lower IT operating costs -- clearly achieving the goals that Oberklammer set out to achieve.