2005-11-03 09:38:41天光雲影

A Guide to Successful SLA Development & Management

Critical Successful Factors

◎Fail SLAs are based on what is easily measured, rather than on what should be measured.
◎SLAs that lack buy-in from end users : 1. Involvement. 2. Relevance
◎The ultimate measure of performance is customer’s perception and satisfaction.
◎The traditional operational quality metrics (availability & response time) are still the most important quality characteristics to measured.
◎Change & increasingly complex technology interfaces pose challenges to define metrics, and these must be recognized and addressed early in the definition process.
Best Practices

◎ Use service-level definitions that :
-ARE Business needed
-Can be defined, measured, examined, corrected, and continuously guided
to maintain gains achieved
-At both meaningful to end users and relevant to their needs

◎Ensure that provide-side definitions and metrics- which both define the roles and responsibilities of the various parties and aid in managing overall service delivery -support end to end service levelmeasurement

◎Completely and unambiguously all the ways in which service-level metrics will be calculated and penalties or incentives will be applied. Include the data sources and calculation formulas to be used, and precise definitions of terms. (note that even a “commonly understood “ term like “business day” may vary in its meaning from one location to another.)

◎Keep metrics to minimum needed to accurately monitor and manage each unique IT service

◎Align pricing with service-level priority differentials, and with the service provider’s achievement of service-level targets

◎Use exception reporting on a daily basis to trigger problem resolution practices, and use summary (I.e., Dashboard) on a monthly basis to report overall performance. Supporting details should be made available on an as–needed basis

◎Service-level problem reports should include
-A description of the problem as reported
-Steps taken in response to initial problem report
-Identification of the underlying problem, if known, and the steps to be
taken to resolve it