Cloud forecast: how to prevent a storm

The cloud offers the promise of flexibility and simplicity. But for companies considering a move to the cloud, evaluating and comparing options can be ‘cloudy’ and time consuming as there is no industry framework for comparing services and measuring quality.

The CIO or IT decision maker needs to be sure that his cloud service is up to scratch and will deliver the quality needed during critical times, but what quality controls are available to enterprises embracing the cloud?

Reliability: not every cloud is the same

The reliability provided depends on the architecture of the cloud provider and should reflect how business critical the cloud services are – or in other words, what is the cost of redundancy versus any potential downtime.

At the lower end of the scale, with services at a lower level entry price, the reliability of services might not be any better than a local single installation via a local server. At the higher end, there would typically be a level of redundancy embedded in the architecture and a duplication level that would guarantee zero downtime.

Insurance against bad weather: SLAs

Service level agreements are the best protection to ensure that cloud services are up to standard. Any business embracing cloud services should ensure that there are strong guarantees written into contracts with strict SLAs in place, according to their requirements. Any investment in getting a true understanding of your cloud provider and its levels of reliability early on, will save money spent on testing services in the future.

Typically, providers will not offer high SLA breach payments if they aren’t running a reliable cloud service and don’t have high levels of redundancy built into their architecture. That means that the best insurance against ‘bad weather’ is to agree on strong SLAs and understand the guarantees that your provider has in place, along with any penalties or potential refunds if services are not meeting agreed standards.

Tests: investigating your cloud forecast

There are different approaches available depending on which element of cloud services are to be investigated. On a lower level, what the kind of resources are being made available by the cloud provider can be evaluated. On a higher level, tests will be run for application performance and availability.

In some countries, it is a legal requirement for companies to store their data in specific geographies, so it’s worth asking the cloud provider that question, but you can also run a trace on the servers being used in the cloud.

Preparing for the storm: danger points

In an industry with strict regulatory requirements, you need to know that your cloud infrastructure complies with the rules. Does the supplier guarantee the location in which the data is stored or whether strictly private data is secure from being accessed by other parties? If the company needs to adhere to certain timings for retrieving data, the CIO also needs to know that the service provider selected can guarantee the service levels needed to comply with these requirements.

Another point of consideration is the durability of the provider. A storm can be created if an unstable or start-up cloud provider goes out of business, which could lead to the manufacturer’s data being sold onto another provider or, even worse, no longer being accessible.

Perhaps the most important danger area is the lack of a framework for comparing service providers, which makes it difficult for enterprises to determine what they are getting. Preparing for a storm is therefore the safest approach: it is essential to invest the time to understand a potential cloud provider and agree on SLAs upfront.