Method for Autonomic Software Integration Testing
Original Publication Date: 2002-Jul-14
Included in the Prior Art Database: 2003-Jun-21
The invention of autonomic software that enquires on both the complete environment (including network connected components) the software that is being installed into it and a central knowledge repository of coexistence, fix and performance data. A report containing a list of the necessary coexistence products, expected throughput and risk analysis is supplied by the operation of the autonomic software integration testing process. The autonomic software integration test would: a) query the installation environment for product levels and throughput capability. The boundaries of the environment to be checked would be dependent on the product, but it is envisaged that any component of the overall system could be checked, even those components over networks. b) check a database for product and fix level dependencies and performance capability/tuning rules of thumb c) report back on any action required for a successful software installation and some measure of the risk associated with installing the software The problem has been addressed on a very small scale by software installation wizards that report, for example, how much local disk space is required for installation and how much computer memory is needed to function as designed. This solution improves on that method by commencing a probe of the target installation network (in virus like fashion) thus establishing a-priori the configuration of the complete installation environment. The autonomic software solution then queries an existing knowledge base to gather and report information to enable a successful installation. The virus like search could be done in a number of ways, for example an agent could be installed on each node that returned the software configuration of that node, or each product running on that node could be architected to return its software configuration when 1 queried.