Talk:Service-level agreement/Archives/2011

Measurements

Some discussion could be lent to:

misleading sampling techniques - e.g. if I sample a service once every 5 minutes to see if a service is alive, then I don't know if the thing has fallen over during the 4 minutes in between.

availability calculations - e.g. the well known "five nines" standard in telecommunication (99.999%) uptime translates to unplanned downtime of less than 316 seconds (5 minutes) per year

exclusions (planned work) - e.g. it is well known in industry that service windows are usually afforded in a contract ("we have the right to take the service down at 1am-2am every Sunday night", or "we can arrange planned outages with 30 days notice", etc). Such excluded time does not contribute to the calculation of adherance to an SLA

Typical Contents

I am not sure why someone has added a tag questionning this section. In my view it is one that could do with expanding, not deleting. Very odd.

Experience Level Agreement

I've just checked the "Experience Level Agreement" reference and I found it to be not relevant to the subject. It only partially explains a 'research project' with little information even it announces it has been finished, and search engines mostly return references to that website when the term "Experience Level Agreement" is introduced. Those facts, joint to the lack of knowledge about such term or concepts in the SLA enviroment lead me to think that this reference was introduced with self-promotion aims, and therefore I have removed both contents from the article. —Preceding unsigned comment added by Angarg12 (talk • contribs) 14:59, 16 November 2010 (UTC)