A Service Level Contract (SLA) defines the level of service a customer expects from a provider and defines the metrics on which that service is measured and corrective actions or penalties, if they exist, if agreed service levels are not met. As a general rule, SLAs are located between companies and external suppliers, but they can also be between two divisions within the same company. Chubby [Bur06] is Google`s locking service for bulk distributed systems. In the overall case, we distribute Chubby instances so that each replica is in a different geographic region. Over time, we found that errors made by Chubby`s global body consistently generated service losses, many of which were visible to end-users. It turns out that Chubby`s real global failures are so rare that service owners have begun to add dependencies to Chubby, assuming it would never go down. Its high reliability created a misguided sense of security, as services could not function properly when Chubby was not available, but this rarely happened. Since the late 1980s, SLAs have been used by fixed-line operators. Today, ALS is so widespread that large organizations have many different ALSs within the company itself. Two different units in an organization script an ALS, one unit being the customer and another the service provider. This helps maintain the same quality of service between different units of the organization and in several sites within the organization.
This internal ALS script also compares the quality of service between an internal service and an external service provider.  An SLO is a service level objective: a target value or a range of values for a level of service measured by an SLI. A natural structure for SLOs is therefore SLI ≤ target or lower limit ≤ SLI ≤ ceiling. For example, we might decide to make Shakespeare`s research results “quickly” and adopt an SLO that our average search requirement rate should be less than 100 milliseconds. We design our cloud specifically for reliability, performance and data protection. Learn more about this in our white paper, “Not all clouds are created in the same way.” In addition, organizations may want to invest in different ALS agreements for different types of workloads. A critical cloud infrastructure service may require the availability of “six nine” to ensure that the core application functionality is still operational, while low-priority workloads can be performed relatively well in terms of service availability with low ALS power. Ideally, the SLI directly measures a level of service of interest, but sometimes only a proxy is available, as the desired measurement can be difficult to obtain or interpret. For example, customer-side latency is often the most relevant metric for the user, but it is possible to measure latency on the server. Depending on the service, the metrics to watch may include: Suppose you have a 99.5% availability standard for network availability. This still leaves nearly 4 hours per month of authorized downtime.