Mechanism for Operational Risk Monitoring, System Testing Optimization and Root Cause Analysis for Complex Applications
Publication Date: 2014-Jul-09
The IP.com Prior Art Database
Software solutions often suffer from the following problems which is being addressed in this publication: A. Monitor operational risk and reduce impact of production anomalies B. Improve testing accuracy between testing and production phases C. Improve root cause analysis of problem systems Using a number of solution profiles discussed in this publication, how to resolve the problems in bullets A, B and C is shown.
Page 01 of 17
Mechanism for Operational Risk Monitoring , System Testing Optimization and Root Cause Analysis for Complex Applications
Software solutions often suffer from the following problems which are being addressed in this invention:
A. Monitor operational risk and reduce impact of production anomalies
B. Improve testing accuracy between testing and production phases
C. Improve root cause analysis of problem systems
Enterprise systems are becoming extremely complex. Enterprise systems embrace more programming models and include more middleware with increasing set of capabilities. Enterprise systems use more layers of computing through virtualized and cloud platforms. Enterprise systems continue to add more access channels such as smart phones and sensors. Enterprise systems are experiencing radical growth of big data and associated analytics. Economic realities further drive complexity with more partner enterprises involved in solution delivery.
Traditionally Enterprise systems are composed of many applications. In many cases, an application is not fully tested under conditions similar to the conditions in production. The reasons for testing and production differences usually include cost, skills, and time. The following are only examples of differences that often exist between the testing (i.e. pre‐production) and production environments:
Number and types of user requests
Different qualities of service enforced in production
User request data
System of record data
Mixing online requests with database batch jobs
Performing administrative tasks in production such as taking database backups
Production APIs or services are not fully available in testing.
The application executes in a shared environment with hundreds of other applications
If such differences are not taken into account, a production solution may experience service level agreement problems such as a long response time or slow down.
This problem is further exacerbated with enterprise systems becoming more complex as companies do not fully understand how or what to test. Applications are being delivered into production less tested simply because the tester doesn't understand how to test everything.
Untested applications are deployed into production and wait for outages, then move into crit‐sit mode. Commercial products are released with inadequate level of testing and rely on customer PMRs to debug
Page 02 of 17
and fix problems. The feedback to testing on what to improve to make production system behavior more predictable is usually simplistic.
Root cause analysis of outages or ...