Customers became frustrated by the poor performance of Inpho’s SMS batch process which took upwards of 40 seconds to complete. In addition, they had several new SMS, Chat and IVR applications to roll out with different versions of MySQL to balance. InPho partnered with Pythian to resolve customer-facing performance issues and reduce service delivery risk for the new applications. Pythian delivered a cost-effective redundant solution and reduced application response time from 40 seconds to 2 seconds. Customers and users became frustrated by the poor performance of Inpho’s SMS batch process which took upwards of 40 seconds to complete. This performance bottleneck forced InPho to compromise on the volume of SMS batches the SMS gateway could process. To address the problem InPho engaged Pythian to conduct a remote database performance and tuning audit of the main Oracle database, which supports InPho’s SMS gateway applications. In addition, InPho had several new SMS, Chat and IVR applications to roll out with different versions of MySQL to balance. All of these new services needed to scale gracefully and offer redundancies in the event of a server failure. Pythian was engaged to audit the runtime environment to reduce service delivery risk.

To address the SMS performance issues Pythian conducted an audit and found several major bottlenecks in the database architecture that were limiting the throughput of the mainSMS batch process. To resolve the issue, Pythian used Oracle’s function-based indexes to allow the SMS batch process to complete much faster. Pythian conducted the necessary code change to the index structure and the programming changes to the Oracle PL/SQL packages. The entire operation was remotely coded, tested, and transitioned to production with no disruption. Pythian’s MySQL experts worked with InPho’s team to determine the business and scaling requirements for the new  SMS, Chat, and IVR applications. Pythian then provided detailed guidance on structuring InPho’s MySQL database architecture to deliver performance and availability under load. To ensure scalability, Pythian recommended the separation of applications onto individual servers and MySQL instances. For cost effective redundancy, Pythian recommended a secondary slave server running multiple MySQL slave instances that also ran specific MySQL versions as required by each application.