MT Expert: When clouds burst , NGD

MT Expert: When clouds burst

By devstars 04 | July | 2013

Simon Taylor says it's not a case of if the cloud fails - it's a case of when.

Clouds of the digital variety are great for increasing competitive edge and making cost savings. But they don’t offer much in the way of silver linings when your vital applications and emails suddenly stop working. Out of nowhere, chaos reigns all around - and while a few minutes of unplanned downtime may be okay, several hours or days will be catastrophic.

Looking at the statistics it’s a case of when, not if, your cloud fails. Even big-name providers such as Amazon, Google and Microsoft aren't immune. Think of all the havoc caused by the likes of Hurricane Sandy, with floods and power outages leading to IT systems knocked out not just locally but globally.

You may believe you have everything covered. No doubt you have pinned your cloud providers down to tight service level agreements and they have impressed you with all their remote 24/7 systems monitoring they can do so that any IT problems can be quickly nipped in the bud.

Fair enough, but there are still the fundamental risks associated with cloud computing to consider. Ultimately, your cloud resides in data centres which could be just down the road, on the other side of the country, in another country altogether. Some aren’t particularly secure or efficient and even the most diligent cloud provider can’t predict or prevent the consequences of data centre downtime caused by power outages, storm damage or security breaches.

On top of this, the majority of cloud providers are often reliant on third party data centres they don’t actually own. But at the end of the day it will still be your company’s reputation and profitability on the line following a ‘cloud burst’.

To mitigate these risks, you need to know what’s behind your cloud - not just who’s providing it: who owns the data centres, where they are located, which of your applications they are hosting.
Ask your cloud provider for a data centre healthcheck, including the following areas:

Resilience and Disaster Recovery

Are back-up systems and fail-safe measures in place? ‘Tier 3’ category data centres should be used as a minimum.

To mitigate risks of downtime from natural disasters or terrorism, see they are located well away from flood plains, large urban areas, flight paths.

Supply should come directly from the grid to ensure reliable delivery and reduce outages caused by spikes and surges.

A choice of carrier networks ensures continuity in the event of failure. These must offer sufficient bandwidth for your provider to efficiently deliver the applications concerned

Data centre location impacts the cloud provider’s price and is influenced by cost of real estate, labour, power supply. Compare prices against similar providers using other data centre locations

Access to data
Check there’s a legal agreement in place should your cloud provider or their data centre partner go into administration. An escrow agreement will ensure you have legal access to retrieving your data

-  Simon Taylor is the chairman and co-founder of data centre provider Next Generation Data