Cloud Computing Reliability

Separating the Enterprise Cloud from Its Development Reputation

Some CIOs and IT pros disqualify cloud computing out of hand because of their notions of its unreliability or fear of their applications not being as available and accessible as their dedicated environment. Interestingly enough, virtualization has matured in the areas of reliability and uptime to the point where nay-sayers should reconsider.

In addition, decision makers can now distinguish between development cloud computing resources and enterprise level cloud computing, which offers guaranteed uptime, enhanced security, reliability and customer support. These enhanced benefits of enterprise virtualization options make it an increasingly popular choice for businesses that are seeking the dynamic and fluid benefits of the cloud without suffering from unplanned downtime or the dreaded outages.

Big box cloud computing players such as Amazon and Google seem to straddle the fence somewhere in between — while they offer some level of reliability, they can also suffer from surprising downtime due to inelasticity or sudden burst connectivity which makes them more fitting for the dev environment versus enterprise usage.

Not all cloud computing environments are like Google and Amazon. They can be very elastic, meaning you only pay for what you need, but availability is not guaranteed which doesn’t work for most enterprises that can’t have mission critical applications and data going offline.

Because enterprise customers need assurances on uptime and reliability, top virtualization players guarantee availability plus factor in burstability to account for and accommodate sudden increases in activity.

Before companies consider the enticing offers of the big box commodity cloud, they need to reconcile what happens when the data servers that store your applications have major technical snafus, experience physical trauma or lose power entirely. Despite redundancies and precautionary measures, the data centers that host major sites such as Authorize.net and DailyMotion experienced significant downtime this past July.

Even after the promise of 99.9% uptime in October 2008, the Google App Engine had performance issues in July 2009 that resulted in high latency and, in some cases, data loss. The takeaway is that all cloud computing solutions are not created equal. A true enterprise solution will offer a 99.99% SLA and make the necessary provisions for reliable, available cloud computing.