SLA requirements for cloud
One of the most important requirements for cloud delivery is the SLA (Service Level Agreement). With applications and data in the cloud, the SLA requirements for cloud have to reflect the change of business process and the risk to customers of not being able to access applications or data.
Realism: Expecting 100% uptime is unrealistic. Only an extremely small number of organizations get close to this figure with their on-premise solutions. For cloud, where there are many more variables that will impact performance, this is a wholly unrealistic target. Rather than use unobtainable numbers, it is important to identify key metrics for the SLA and how these will be priced.
Latency: If the application is latency sensitive, it is critical that the data and application are located in close proximity, ideally within the same data centre.
Set a performance baseline: Metrics only work if there is something to compare them with. Part of the cloud migration planning should include application profiling. This will provide an indication of how applications use resources and what levels of performance are currently being achieved through the on-premise installations. This data is ideal for setting base metrics and then deciding on what additional levels of performance are required.
Certification: The cloud platform provider has the opportunity to ensure that they conform to a range of international standards for data centre operations. These range from power efficiency to security standards. Organizations looking to take on cloud services should ensure that their partner is, at the very least, meeting the same standards as they do. Those that meet the most stringent standards will attract a premium price for access to their facilities and services.
Resolution: With any service-based agreement there is always the potential for failure. There should be a very clear approach to how any dispute will be resolved. This should start with the SLA which will detail response times and how an issue is notified. The more business-critical the system, the shorter the response time, and therefore the more important it is that there is a clear, unambiguous and effective process to escalate any issues.Watch Full Movie Online Streaming Online and Download
Compensation and cost: Compensation levels should be clear and at the front of the SLA. They will identify what can be expected should the SLA be breached and how any compensation will be applied. Transparency is important and a clear process will improve customer relationships.
Always on: Disaster recovery and business continuity are still challenges for enterprise IT. Part of a premium SLA will show what the process is for ensuring business continuity, and how the disaster recovery process can be invoked.
My paper “How to move to the cloud” is a useful guide to a successful migration to the cloud. It’s available to download here.