tract, and many companies are now using these agreements as tools to enhance the business relationship between the service provider and the buyer. In some situations, service level agreements are being developed for internal providers of IT services, as well.
When they were first developed, service level agreements were based on easily quantifiable metrics such as the amount of time that a system would be available. Thus, a service level agreement might have provided for 90 percent uptime. As the sophistication of networks increased, along with the demands that users made on those networks, the sophistication level of service agreements increased, as well. Buyers learned that having a system available 99 percent of the time did not improve the response time of the network. Remote users who sought to access e-mail when they were on the road might technically have access to e-mail, but if the response time was too slow, companies could incur lost productivity and--before the Internet became widely used--increased phone charges.
As the Internet became an integral part of many company's operations--both internally and externally--the demands on the networks that ran those companies increased exponentially. Response time, throughput, available bandwidth and many other criteria became important business considerations in addition to simple network availability. Yet, the tendency in service level agreements continued to be based on quantifiable network metrics rather than business-level metrics. In addition, the thrust of service level agreements tended to be punitive for the service provider, resulting in overall negative associations for the agreement (Wilson, 1999).
THE NEED FOR SERVICE LEVEL AGREEMENTS
To some degree, the need for service level agreements stretches back to an earlier time in computer processing where companies could not afford to own or lease their own computing resources and so used time-sharing arrang...