A Large Health Services Entity Based on the West Coast
As a healthcare provider, their clients are far-reaching and have a wide range of technology backgrounds and budgets. These clients range from hospitals to doctors, offices, clinics, and individual users, so acquiring specific software to suit all of their needs is impossible. The best that can be done is to set guidelines and recommend software that is known to adhere to industry standards.
The overall system had both real-time system updates and batched updates that were sent from the client sites to the main data center. Standard FTP (File Transfer Protocol) transfer software was used for the batched data that would be sent at close of business each day to the healthcare provider. From most offices these data were transmitted in one hour with no problems.
With no known data center host system changes, the mainframe began behaving very erratically with slower and slower response times, which became evident when using the CleverView for TCP/IP monitoring tool. There was a ripple effect with the slowdown impacting more and more socket-based mainframe applications and subsystems. A pattern emerged showing the slowdown occurred between 7:30pm and 9:00pm (Pacific Standard Time) Monday through Friday. The operations center knew they had a problem on their hands.
Using CleverView for TCP/IP historical reports and investigating system logs, key discoveries were made including
| identification of FTP as the offending application environment,
dramatic decreases of ECSA storage (at this time socket control blocks were part of ECSA),
socket control blocks that were dropping to zero (all socket connections were then blocked),
discovery of the offending system location to be in an office in Honolulu, HI. |
While this information was useful, it still didn't tell operations why the system was reacting abnormally. Why were they losing so many SCBs? The technician had heard of CleverView for cTrace Analysis and its ability to simplify the process of collecting TCP/IP packets on z/OS then quickly analyze them. Within hours of receiving the product it went live and within a few days it was deployed across multiple processors, efficiently collecting traces to analyze their FTP application issues.
What the technician saw with the trace was that the MPUT statement was not responding as specified in the FTP standards documents. They talked to the owner of the PC sending the information and discovered their use of an FTP application never tested by the Data Center technicians. The trace indicated that wait time logic was inconsistent with FTP standards. The data center team had the office use a known FTP application and the problem was resolved. Contacting the offending FTP development team (a college student), the operations team provided the traces showing the problem area to the developers, allowing them to change the application logic to be compliant.
AES focuses on the development of network performance and availability Business Service Management solutions. Through its strong emphasis on listening to customers, foreseeing far-reaching trends in the marketplace, and providing stellar support systems, AES continues to maintain its lead in providing state-of-the-art, easy-to-use performance tools for the very backbone of eCommerce.
|
