The SLA should contain not only a description of the services to be provided and their expected service levels, but also metrics that measure the services, the obligations and responsibilities of each party, the corrective measures or penalties applicable to violations, and a protocol for adding and removing metrics. The service elements include the particularities of the services provided (and what is excluded if there is reason to doubt), the conditions of availability of the service, standards such as the window of opportunity for each level of service (for example, prime time and non-prime time may have different levels of service), the responsibilities of each party, escalation procedures and cost/service trade-offs. Since the late 1980s, SLAs have been used by fixed telecommunications operators. Today, SLAs are so prevalent that large organizations have many different SLAs in the company itself. Two different units in an organization write an SLA, one being the customer and the other the service provider. This approach helps maintain the same quality of service across different units of the organization and across multiple locations in the organization. This internal SLA scripting also makes it possible to compare the quality of service between an internal department and an external service provider. [4] Overview of contracts – This first section defines the basis of the agreement, including the parties involved, the start date and the generalisation of the services provided. Typically, these processes and methods are left to the outsourcing company to ensure that such processes and methods can support the SLA agreement. However, it is recommended that the client and the outsourcing company collaborate during the negotiation of the SLA in order to eliminate misunderstandings about the process and method of support, as well as the management and reporting methods.

FP7 IRMOS also looked at aspects of translating application-level SLAs into resource-based attributes to bridge the gap between customer expectations and cloud provider resource management mechanisms. [14] [15] The European Commission presented a summary of the results of different research projects in the field of ASAs (from specifications to control, management and implementation). [16] Measures should be designed in such a way as not to reward bad behaviour on both sides. For example, if a service level is violated because the customer does not provide information on time, the provider should not be punished. . . .