While moving to the cloud-based solutions, Cloud Service Level Agreements (SLA) play a crucial role. An SLA comprises of entire service level management strategy and is seen as an important and unmissable component. A Cloud SLA helps in filling all the existing gaps between customers’ and Cloud providers’ expectations. Also, Cloud SLAs can emerge as a great communication enabler between the involved parties.
SLA, as the name suggests, is an agreement, not any form of contract. It is an agreement that takes place between the internal and external customers and Cloud providers. It usually contains the complete documentation regarding services that the service provider will be providing. I
n crisp language, it is the promise that has to be mandatorily followed by the service provider. SLAs comprise most IT services where providers offer Cloud-related assistance like Cloud services, IaaS, PaaS, SaaS, and DRaaS, to name a few. Besides being an agreed service level, SLA also helps in defining a complete overview of what can be done if the Cloud provider fails to provide desired levels of availability. A clearly defined SLA assures enhanced communications between the provider and the customer.
However, for a well-defined SLA, several important factors have to be considered. These crucial factors assure customers have complete transparency and understand all the terms that providers need to follow at all times of services compulsorily.
Key Features of Cloud SLA
Any reputed cloud service provider defines a clear and transparent SLA that promises to offer all the services mentioned in the Cloud SLA. Such Cloud providers will provide all these services, but such promises will be backed with penalties & incentives. The mandatory features of any Cloud SLA include the following-
- System Availability- Committing to offering a system availability of 99% or even higher
- Disaster Recovery- Backup will be performed within 24 hours if there is a disaster at data centers
- Data Ownership and Integrity- User must be able to get his data out of the service provider’s system if he’s not looking to go use avail of the services
- Response Time- The service provider must be able to categorize issues and accordingly respond
- Escalation Path- The user needs to have a well-defined escalation path if he feels some issues need to be escalated
- Maintenance- The service provider should define major maintenance activities at regular intervals by timely notifying the user
- Product Notifications- The service provider must update the user about new products and releases & upgrades
Service Availability Levels
Before approaching the SLA, one must consider the promises and guarantees mentioned in the SLA, such as service availability. Availability defines how frequently might go down. Availability is usually expressed in terms of three, four or five nines. If the Cloud provider expanses beyond nines depicting lower ranks of nines, increasing the costs. Besides high availability, business functions that are offered should are offered must also be taken into account.
Service availability levels can be defined under two major categories: the services and their reliability? The service may vary depending on the Cloud model, i.e., IaaS, SaaS and PaaS. When we talk about IaaS, the service provider is responsible for offering the data center infrastructure as a service. Talking about the PaaS Cloud model, the provider is responsible for providing the platform’s functionalities. Availability in this model includes the reachability and usability of the forum. For SaaS, the service provider offers enhanced availability of data as well as applications. With SaaS, the availability in terms of uptime.
The SLA should also discuss the formation and expiry. The agreement’s time frame is such an element present across different technical agreements and contracts. The starting date of SLA allows the user to track his IT performance. In the case of new services being introduced or renewal of existing services must be reciprocated to the user with accurate details. The SLA should also clearly define the meantime to repair as well as mean time to respond.
Exceptions and Exemptions
While going through the SLA, the user must analyze the limitations as well as exemptions. The common exemptions may differ as per Cloud deployment models.
Some of the common exemptions with IaaS include-
- The service provider will not be responsible for what users are going to do with servers
- The service provider will not be responsible for any unsubstantiated OS-based connections
- The service provider will not be responsible for any externally deployed network
When we talk about PaaS, the exemptions might be similar to IaaS. The service provider will be responsible for offering the platform and not what functionalities that the user will run on it.
For SaaS, the service provider will be responsible for offering the service as a whole, making Cloud service providers more accountable than IaaS and PaaS.
SLA Implementation Reports
There need to be well-defined criteria on which the performance and its evaluation can be performed. The performance needs regular evaluations, and any loopholes in the performance also must be reported. The Cloud service provider must present an SLA Implementation Report that contains the visible figures of the conducted activities. Certain KPIs are also marked in the report, and these have to be inevitably fulfilled during the agreement period. These KPIs have been implemented after having mutual consent between the provider and the user. The user must carefully define the indicators or metrics for recording the performance.
Also, it is essential to note that SLAs must be measured appropriately against the defining metrics. Some mechanisms can be deployed for recording any breaches in the SLA, and using these mechanisms, various SLA reports can be generated that serve as the discussion base between the user and the service provider. It is better to have automated SLA reports whenever possible.
All the SLA reports are prepared along with the necessary activities being carried out. So, it is important for the user even to measure his SLA’s performance. The most-effective way to calculate the service provider’s overall SLA achievement is by creating a form of Availability Index.
Availability Index can be calculated using the below formulae-
Availability Index = (Multiple of actual availabilities/Multiple of target availabilities) * 1000
After determining all the reports and determining the performance, the user must also be aware of the implications of an SLA. Using SLAs, users can construct the system for fulfilling the SLA, and in case the SLA is not fulfilled, it is clear that there is data loss and businesses are going to suffer. The SLA metrics for measuring SLA are essential for determining long-run success. A good SLA is essential as it helps define customer expectations and commitments, assuring KPIs for customer services and internal organizations.
Hence, SLAs have now become an important component of any IT strategy. SLAs become important when determining the performance of a Cloud provider. SLAs also give a strong scenario for creating systems as well as optimal resource utilization. It is important that users clearly read the terms mentioned in the SLA, and Cloud providers adhere to the commitments made in the SLA. They must be accountable for any losses/issues that might arise during the unavailability of services.