19th september 2008 -Best practice generation for virtualized envrionments
Adherence to best practices is essential for successful configuration and deployment of complex systems. In such scenarios, experts rely on experience as well as repositories of best practice guidelines to proactively prevent any configuration problems while deploying systems. Reasons:
- the entities in a data center are associated via complex physical and logical relationships, which may be an order of magnitude more than the entities themselves.
- the diagnosis of a problem requires the collection, filtering and correlation of a huge amount of data from various sources over different periods of time
- most deployments are not integrated enough to present a single, usable text or graphical interface to navigate through the problem diagnosis process. So most times some one has to correlate data across multiple point tools using techniques that are tedious, time consuming and prone to errors.
- technology related to systems keeps on changing( or evolves continually) due to the need for product differentiation. This hampers the ability to diagnose configuration problems because of the "lag" behind the state of the art
- standardization process that dictates inter vendor interoperability may be immature and continuously evolving leading to a hard-to-diagnose configuration problems.
This becomes even more apparent when virtualized environments are involved. virtualization helps in simplifying the complexity of IT environments. And it is increasingly being adopted and used by IT. True to the fact it helps with flexibility and ease, the way complex pieces of software works in virtualized environments is not fully known. There are no studies on what kind of problems one may encounter in virtualized environments.Nevertheless, best practices.
So, there is an inherent need for best practices in virtualized environments.
automatically generate best practices for virtualized environments -(a particular domain - db servers, application servers, web servers, NAS or SAN ??)
- automation -> because the manual generation of best practices is costly. In most cases, it requires many man years of data gathering and analysis.
- the best practices are dynamic and change as new technology is introduced
- systematic way to gather all the data needed for problem diagnosis.
- quality and quantity of data
- since data sets have a large no of entities, attributes and associations, there is a need for effective dimensionality reduction so that the data sets can be analyzed efficiently.
- purest subset of entities, attributes and associations that contribute to a configuration error
- require the use of a highly accurate data classification tool that can overcome incomplete dimensions as well as noise in the datasets.
- a combination of IR principles, entity ranking algorithms and classification schema to statistically infer the best practices relevant to a set of virtualized environments
--Pavan 08:27, 19 September 2008 (UTC)