Service Level Agreement (SLA) Explained

In today's service-driven economy, anyone who avails of any service would naturally want that service to be exceptional. But the question is, how does one measure service quality? How can one definitively say whether service standards are being met or not?

Enter the service level agreement. A Service Level Agreement, or simply SLA, is a formal contract or agreement between a service provider—be it an IT company, a cloud service, or a managed service provider—and its customer. It specifies measurable metrics by which the quality of service rendered by the service provider is assessed to determine if agreed-upon criteria are being met.

Without a clear SLA in place, both providers and customers can face uncertainty. The provider may be unsure of the expected level of service, and the customer may have no way of knowing if they are receiving the service they are paying for.

This agreement, in particular, enumerates in writing the responsibilities of both parties—the exact services the provider will deliver (e.g., cloud services, data storage, IT repair and troubleshooting) and the things the customer must do (e.g., reporting incidents promptly, ensuring that set guidelines are being followed).

Together with these specific responsibilities are the various metrics by which the provider’s performance will be measured. These metrics vary depending on the service provider or the service being availed, but among the common ones are as follows:

• Uptime. This refers to the percentage of time a service is available. An Internet service provider may, for instance, commit to providing a 24/7 Internet connection to its clients.
• Response Time. This refers to how quickly the provider responds to incidents or requests. A managed services provider, for instance, may promise within-the-hour response times from receipt of an incident report.  
• Resolution Time. This refers to the time taken to resolve issues. An IT company, for example, may guarantee next-day repairs for troubleshooting services or incident resolution 24 hours after receipt of the incident report.
• Throughput. This refers to the rate at which data or transactions are processed. A data centre, for instance, may commit to analysing 1 zettabyte of data per day for its client.
• Capacity. This refers to the resources allocated to the customer. A cloud service, for example, may provide 30GB of free storage monthly to its customers.

The expectation, of course, is that the service provider will be able to meet whatever is indicated in the metrics outlined in the SLA. If not, there will be penalties and remedies, and these are also contained in the agreement. So, if the provider fails to meet SLA targets, these penalties or remedies, which may include service credits, refunds, and additional support, will thus be triggered.

Consider as an example a cloud hosting provider, which we will call “CloudHigh,” and an e-commerce company, which we will refer to as “Commerce Anywhere,” doing business. Let’s say CloudHigh will be providing cloud services to Commerce Anywhere. Before any service is even rendered, they will agree upon an SLA and determine the metrics of quality service. These metrics may include the following:

• Uptime: CloudHigh guarantees 99.9% uptime.
• Response Time: CloudHigh commits to responding to issues within 30 minutes.
• Resolution Time: CloudHigh guarantees that critical issues will be resolved within four hours.

The penalties will be outlined as well, and they may include:

• If uptime falls below 99.9% in a month, Commerce Anywhere receives a 10% service credit.
• If the response time exceeds 30 minutes, CloudHigh provides an additional month of service free.
• If the agreed-upon conditions are not meant for two consecutive months, Commerce Anywhere can terminate its contract with CloudHigh.

Similarly, the SLA contains Commerce Anywhere’s responsibilities as the contracting party, and these responsibilities may include:

• Distill data to be stored in the cloud.
• Report critical incidents within an hour after discovery.
• Secure its endpoints.
• Pay the monthly bill 10 days after the last day of the preceding month.

In the same way, Commerce Anywhere has remedies in case CloudHigh fails to do its part in the deal, the latter also has remedies should the former be remiss in its responsibilities—and these will also be outlined in the SLA. So, in this case, possible remedies may include early termination options or monetary penalties for non-compliance. 

As the example shows, the SLA serves a critical function as it not only enumerates the expected services and the quality expected of them but also outlines ways to even out the agreement in case either party does not meet expectations.

The specificity of the SLA, therefore, manages risks by defining performance standards, thus helping providers reduce the risk of service disruptions or poor quality—and saving customers from bad service in the process. The SLA also guarantees accountability, not only on the end of providers but also on the clients.

As a final note, keep in mind that SLAs vary based on the type of service and industry. But whatever the metrics and responsibilities outlined in them, these agreements are undoubtedly essential to maintaining a healthy provider-customer relationship and ensuring reliable services.