Cloud Services And The Hidden Cost Of Downtime update from March 2014

As any networking professional knows, downtime costs money. However, few know exactly how much money downtime costs. Estimates, calculations, and incidentals are all open to interpretation. This creates a lot of uncertainty.

Cloud computing is a good tool to use here. Many IT pros are turning to cloud-based technologies to mitigate the cost of downtime. However, is the viability of a cloud migration backed by facts or based on suppositions?

The assumption that cloud services can reduce downtime is founded on the belief that third-party providers deploy all sorts of continuity technology that all but guarantees uptime. That belief, coupled with service-level agreements (SLAs) that make promises about limiting unscheduled interruptions in service, can give you a sense of security. The real question becomes whether that sense of security is false or justified -- and, more importantly, whether a value can be assigned to it.

To determine that, one has to delve into some complex calculations to reveal the true operational costs of the cloud -- how much it costs when it works and how much it costs when it doesn't. Those costs have to compared to an equal scenario for non-cloud services. Of course, operational costs and intangibles also must be taken into account.

For example, IT operations have to take into account the costs of a facility, the staff, and everything surrounding it and then correlate the value the facility offers to the enterprise. That's a very complex calculation. On the other hand, a cloud service can be calculated down to the fees charged and any related incidentals: internal support staff, connectivity, etc. Though it seems simpler to calculate the cost of the cloud, the numbers can be easily skewed by overlooked elements, contractual price changes, changes in scale, and so on.

The biggest unknown that can affect a strategy's cost is the downtime demon -- how much something costs when it doesn't work. Many IT managers overlook the cost of downtime, simply because they believe it will be mitigated by a cloud strategy. However, this can skew the numbers and make a cloud service seem more affordable than it actually is.

One only has to look at recent history to see that cloud services are not impervious to interruptions. Take Amazon Web Services, one of the most popular providers of cloud services. During the past few years, AWS outages have affected businesses small and large. Its biggest customer, the Amazon.com retail site, suffered downtime in 2013 due to cloud services failures, as well as consumer services such as Instagram and Flipboard,

[Security also is a key consideration when evaluating a cloud service. Read about a project to develop a framework for secure cloud connectivity "Cloud Security Alliance Launches Secure Network Effort."]

What exactly did downtime to Amazon.com cost the company, one that reportedly makes about $120 million per day from its Web presence? A quick and dirty calculation shows that it lost $5 million in revenue for every hour the site was down. However, the cost is probably actually higher, since that figure takes into account only sales from the site. What about the cost of all the ancillary business tasks that could not be performed? How many employees sat idly in their cubicles while things went awry?

Of course, this example is based on the largest of scales. But the lesson remains true: Before making that leap into the cloud, IT managers need to know how much downtime would cost their businesses.

That raises another question: Can IT managers protect their businesses from excessive downtime in the age of the cloud, especially when they have little control over technology tucked away in some virtual datacenter? The trick here is to understand the true costs and use that to negotiate effective SLAs that protect the business and shift the liability back to the service provider. Those costs need to be factored into the legal agreements struck with a public cloud provider, and they have to be explained to upper management, so they can understand the likely costs of downtime if it occurs.

Most IT managers know they will have to deal with downtime at one time or another, regardless of the infrastructure in place (cloud or local). The real litmus test for value is in preventing downtime and measuring the cost before signing on the dotted line.