SLA in a cloud environment

December 03 2015 By Posted in Expertise

For reasons of economy and efficiency, the cloud has become an increasingly indispensable part of the application chain. The IT department is responsible for the quality of the digital services it provides to the company’s internal and external clients. And yet, when those services depend on the cloud, IT finds itself in the position of a client with respect to cloud providers. How can you check the quality of the services delivered by cloud providers and track their conformity with SLAs in order to enforce them? It all rests on the idea that “if you can’t measure it, you can’t manage it”.


The cloud is getting bigger

With services ranging all the way from infrastructures, to platforms, down to software, the cloud offers solutions to many of the issues that companies are grappling with. The strength of the cloud lies in its ability to enable companies to adjust to new ways of working, with solutions addressing issues of mobility and itinerant security, resource scalability, content distribution, data centralization, and more.


Being dependent means being demanding

For companies that outsource their applications, infrastructures or platforms to the cloud, there are also drawbacks. The quality of the transmission network once again takes on major importance in a private cloud, but most especially in a public cloud. In this respect, one noted weakness of the cloud is that, exceptions apart, it does not guarantee very high availability. And when a malfunction occurs – whether at the level of the outsourced application or in the network or the platform – internal IT production teams cannot intervene directly. Such a loss of control means IT managers are forced to rely on service level agreements (SLAs) to ensure quality for their own users.

And yet, generally speaking, SLAs only partially meet the needs of the companies that depend on cloud service providers. Among the clauses to look carefully at in an SLA are the ones dealing with availability, performance, security (redundancy, segregation between clients, location with respect to legislation on data protection), traceability (actions and data), and the penalties applicable to the provider in the event of non-conformity.


Monitoring: measure to manage better

Service Level Management is a challenge to IT managers. They need to be especially vigilant about the metrics cited in the SLA because:

  • downtime of less than a few minutes and maintenance downtime are usually not counted
  • SLA indicators are measured just outside the data center (so there aren’t any metrics about user-perceived quality)
  • and lastly, notification and proof of any shortcoming on the part of the cloud provider must be supplied by the client

Fortunately, measuring the quality of the service provided is not a complex matter because there are tools for performing this task. Datametrie makes it possible to associate SLA measurements with business actions, and brings together QoE (quality of experience) and QoS (fine technical diagnostics). Monitoring generates alerts when anomalies are encountered.

To go another step farther and conduct individualized performance measurements, RUM BI measures the service provided to each user, and takes into account the user’s context (device, OS, versions thereof, etc.). In this way you have all the information you need to build new business SLAs.

For further information, please contact our team.


Leave a Reply

Your email address will not be published