High availability of computer systems, especially internet computer systems, is a necessity in today’s world when the absence of a computer system and/or web site will not only affect the immediate revenue stream, but will affect revenue to come by decreasing the likelihood of future use of the system by those users who encountered the absence of the system. Merely installing failover software and hardware will not ensure that high availability will be maintained under load conditions. High availability testing must be conducted wit h analysis and a structured approach to provide an accountable confirmation of reliable high availability functionality. This testing will return to the system owner not only a metric of availability versus conditions for functionality, but a means to calc ulate cost versus risk reduction for system reliability.
The paper will define availability and “high” availability. A testing typology will be discussed in detail that explores availability in terms of software, system (software and hardware), and process (software, hardware, and procedures). A testing taxonomy will then be reviewed in terms of induced failure analysis by means of increased load, singular system component termination, and multiple system component failure.
Test strategy derivation will be discussed next showing traceability to the logical system architecture and the physical system architecture, with special attention to how to discern points of criticality for system availability. Test documentation will then be reviewed with emphasis on particular points that must be documented in an availability context. Finally regression testing of high availability will be evaluated as to its value for maintenance of high availability after the initial deployment of the system in both a periodic sense and in system updates.
Mr. Jones has a Master of Science in Computer Information Science and a Master of Business Administration. He has been through software engineering training at Boeing, Texas Instruments, and McDonnell Douglas in the past. He has worked in the software industry since 1976 and is currently the Chief Consultant at Dromedary Peak Consulting, which provides analytical direction and operational support for business. He is also an Assistant Professor in Information Technology at Western International University. His courses include Advanced Software Engineering, Advanced C Programming, Information Resource Management, Internet Business Strategy, and Web Application Development. Some of his published articles include: “Pragmatic Software Configuration Management in the E-World”, “Pragmatic Software Testing in the E-World”, “Software Configuration Management for the Web”, “Report from Captain QA from the Web”, and “Four Conceptual Attributes for Successful Web Applications”.