Cloud Services : Service Level Agreements (SLAs) and Availability

Cloud Services, in the context of service level agreements (SLAs), providers often make statements about availability that are not technically verified. This article show specific examples of how availability of Cloud Services can be calculated. Previously, we wrote about availability in Cloud Computing Services under the article title Availability in Cloud Computing, second important article in this context is Service Level Agreement for Cloud Computing Services. Another couple of important articles are :

1. Understanding Scalability and Service Continuity are not equivalent terminologies in Cloud Computing
2. Clarity of Cloud Services in the context of pricing

During the customer attracting  phase, often small and medium-sized hosting companies and resellers arises the question of the availability of services themselves. But many manages to have only a vague idea of ??the problems of availability. Most often, the availability of a service is described by a percentage. But it is frequently overlooked that in calculating this percentage, then in the service level agreement (SLA) will be written to lurk some pitfalls.

Cloud Services : Example of Server Availability

Such pitfalls are, for example, the reference point for the calculation of down time (year or month), the definition of the underlying services (server, firewall or network connection) or the potential link between availability. Vendors often get customer with the statement “We would like to offer a server with the availability of 99.5 percent. ” In such cases, this requirement will be included in an offer, it is assumed that the desired Availability refers to the particular year and only on the specific server or node. The customer probably do not realize that the server may be down a whole day at a time and availability is still adhered to – because 0.5% includes not less than the promised or the server availability is not affected if the firewall or the network connection is gone, the server is still running, but is not accessible from the outside. Legally, the Server is up.

Cloud Services : Calculation of Downtime

The public favorite hosting services, usually gives a Dedicated server at $200 / month whereas Rackspace, Media Template charges nearly 10 times, what is the difference that makes the 10 times pricing? This is predominantly the firewall or network connection, which are used from another bigger hosting service – they are not dedicated. This is fact that, there are only few Hosting companies with full own setup. Resellers are better to avoided for complexity. So, the fact that was never apparent – a dedicated server provider ultimately uses some form of Cloud Services, which in turn can have issues remain fully blank to the end user.

A good overview of how the percentage availability of a service affects in practice, is obtained with a table of annualized potential downtime. A chart is here :

Alternatively, specifying the relative availability of critical projects on the month is the best way. For example, 99.9 percent availability is expected in the specific month is expected to have a maximum outages of 43:48 minutes. When we are talking about the availability of services and percentages, you have to consider whether you can live with the maximum possible downtime. Another point is the maximum outage duration. You should not be within a downtime longer than half an hours at a time, at maximum.