Browse Prior Art Database

An Algorithm for Ranking SLAs for Importance Based on Trends, Violations, and Quantity.

IP.com Disclosure Number: IPCOM000016682D
Original Publication Date: 2003-Jul-09
Included in the Prior Art Database: 2003-Jul-09
Document File: 3 page(s) / 7K

Publishing Venue

IBM

Abstract

An enterprise IT shop may manage several hundred resources spread across thousands of SLAs. A method is needed to differentiate SLAs by customer so that those violating or about to violate contracts can be dealt with first. The advantages of using this invention are: 1) increase the usability of large scale SLA applications by displaying SLAs by importance to an SLA Administrator 2) enable SLA administrators and operators with limited time and resources to focus on solving IT problems that are predicted to breach or have breached SLA contracts with their customers.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

  An Algorithm for Ranking SLAs for Importance Based on Trends, Violations, and Quantity.

Disclosed is a formula for ranking the results of an SLA evaluation process. To fully understand the basis of the formula, on the Service Level Agreement process will be briefly described. The process starts with a customer and an IT department negotiating a contract for the IT department to provide the customer with a certain level of service. A Service Level Agreement contains one or more Service Level Objectives or SLO. An SLO contains a resource in an IT infrastructure that is being monitored for a level of service defined by a metric value. An SLO can be internal or external. An internal SLO is not part of the negotiated agreement between the customer and the IT provider, but provides useful data on whether an external SLO which was part of the negotiated agreement will succeed. An example of an internal SLO could be the response time of accessing a certain application on a web application server and an internal SLO might be the free paging space on the host on which the web application server is running. As the amount of free paging space gets smaller this may have an indirect impact on the response time of applications running on this server which would have an impact on the customer.

Ranking Calculation

For purposes of this article, we assign a value of I to represent an internal SLO and a value of E to external SLOs. To achieve a proper ranking value, the constant chosen to represent E should be greater than the constant chosen for I. Over time, an SLO may produce one or more trends t as well as one or more violations v. A function of E, say E(V) is used to represent the number of violations from external SLOs and the function E(T) is used to represent the number of trends associated with an external SLO. Similarly for internal SLOs, we have I(V) for violations and I(T) for trends. Given this, the following formula can be used to produce a ranking value. Rank for SLAs is the sum of the following three values:

1) For all External SLOs across all SLAs sum the following two values:
a) multiply number of violations v by E(V)
b) multiply number of trends t by E(T)
2) For all Internal SLOs acr...