CASE STUDY - Good Technology Boosts Transaction Processing Application by 93 Percent
Commentary by Simon Ha, Database Administrator, Good Technology
California-based Good Technology is the industry leader in enterprise systems for corporate message and mobile application access. They offer corporate customers the ability to access various mission-critical enterprise applications via PDAs and other handheld devices. Mobile workforce employees can access Microsoft Office, Oracle, salesforce.com, and Siebel through wireless networks.
When funneling transaction usage data from a variety of live collection points, however, Good Technology ran into a processing bottleneck. This transaction data—which is similar to an individual’s cell phone usage—was needed for analysis by Good Technology’s accounting and marketing departments. They needed to process and analyze the data to extract critical business intelligence. Because the company’s operations were 24/7, raw data was continually streaming in, sometimes at the rate of 35 million rows per day.
Industry leader in enterprise systems for corporate messaging and mobile application access
- Minimize the elapsed time of daily transaction processing application
- Apply various transformations to high volumes of data to allow for simplified
- SUN 420R with 4 CPUs
- Oracle 9i data warehouse
- UNIX Utilities
- SQL Plus
- 93% performance improvement in the processing of data
- Join, append, and lookup operations efficiently transform 35 million rows of transaction data per day
The problem was that the processing, which was done using SQL Plus and some native UNIX utilities, was taking 16 hours. Good Technology’s staff needed to minimize the elapsed time of daily transaction processing and apply various transformations to the high volumes of data to allow for simplified analysis.
Good Technology requested a proof of concept from Syncsort—who in turn committed to a demonstration of a significant elapsed time savings with the implementation of DMExpress on this mission-critical application. Using 18 million rows of actual production data, which would have taken about eight hours’ processing time, the engineers at Good Technology ran a sequence of DMExpress steps and were able to bring the processing time down to just 30 minutes, a reduction of 93 percent. Very pleased with this outcome, Good Technology licensed the Syncsort product.
“Not only were we able to improve the performance of our transaction usage application; we also gained functionality through the use of the DMExpress GUI,” says Simon Ha, database administrator at Good Technology. Join, append, and lookup operations are now easily and regularly applied to up to 10 gigabytes of data per day. Results have been so satisfactory, in fact, that Good Technology is planning to expand the use of DMExpress to other departments within the company.
This article originally appeared in the issue of .