Service level agreements (SLAs) are, in effect, a business’s contract with a cloud provider. The SLA will cover the provider’s responsibilities in addition to provisions for handling incidents affecting the service.
While some companies will offer standard “off-the-shelf” SLAs, good providers will strive to provide flexible, amendable contracts to accommodate your business needs and the scalable nature of cloud. The latter type of SLA has become the preferred option, with a lack of customisable SLAs listed as one of the Top 5 inhibitors to cloud adoption, according to the Longhaus Cloud Pulse report, in which Datacom was recently ranked the #1 Local Cloud Provider and overall No.3 overall provider of Infrastructure as a Service in part for our robust SLAs.
Datacom’s customised SLAs will typically cover server availability, which includes both production and development and network availability, and adhere to ITIL® best practices. For instance, a cloud service provider (CSP) might promise between 99.5 and 99.99 per cent server availability. Good SLAs will also include Change, Incident, Problem and Capacity Management to minimise the number of “Priority 1” incidents that would have a severe impact on the business. Datacom will also continually review your SLAs regularly to ensure they continue to perform to your standards and potentially provide service credits should we fail to meet one of the provisions in your agreement.
A service availability clause is an increasingly sought after component in SLAs and an area more concerned with the “end service” rather than the individual components that make up a service (for example, one component of a high availability cluster may be offline, but the email service remains available). The challenge with service availability is the end-to-end nature of measurement and the number of third parties involved in delivering a given service. Datacom includes a guarantee of 99.9% uptime on all servers hosted in our cloud and can increase this percentage for specific systems if necessary.
There are other important areas upon which a good SLA like Datacom’s will touch, such as:
- Security and privacy in the form of adherence to the Privacy Act, data sovereignty and data encryption
- Response and restoration, specifically response times for “Priority 1” incidents
- Accessibility of data if you want to retrieve, analyse or move it out of the cloud
- Location of the data, whether it is onshore or offshore and complies with regulations there
- Both engagement and disengagement provisions, the latter for when you want to remove your data from the cloud; Longhaus’ cloud research actually advises enterprises to agree on disengagement before engagement
- Disaster recovery and business continuance plans built upon recovery point objective (RPO) and recovery time objective (RTO) in case of an outage, natural disaster or other major disruption
- Assessment of readiness for cloud adoption by professional services
Additionally, an SLA should include metrics for measuring how well the provider delivers the cloud service, such as elasticity, and recourse for both the consumer and the provider if SLA terms aren’t met. Datacom covers this area by offering service credits for any part of the SLA that is not met.