Brett Winterford at ITNews.com.au recently reported on a group of digeridoo-gooders at the Universiity of New South Wales, who asked a question that many of us were wondering – how do cloud apps perform under stress tests, and are they up to the tasks that modern enterprises typically throw at their in-house networks?

Testing Amazon, Google, and Microsoft’s cloud offering, they simulated 2000 concurrent users connecting to each of the cloud services, “measuring response times and other performance networks.”

What they found was that the Web-based cloud services could scale up to meet the demand… but not reliably.  Response times varied depending on the time of day, what features were added and dropped, and… well…

"Using Google AppEngine, none of your data processing tasks can last any longer than thirty seconds, or it throws an exception back at you," [Researcher Anna Liu] said. "This is very consistent with the Google business model - they want to enable simple web applications to thrive on the Internet. AppEngine is there to enable the rapid development of simple web applications that don't include intense computing at the back end."

According to Liu, the company most poised to accept migration of in-house enterprise apps is Microsoft.  Still…

"None of the platforms have the kind of monitoring required to have a reasonable conversation about performance," she said. "They provide some level of monitoring, but what little there is caters for developers, not business users. And while Amazon provides a dashboard of how much it is costing you so far, for example, there is nothing in terms of forecasts about what it will cost you in the future.”

Cloud computing applications may eventually replace many of the traditional IT applications, but the limitations inherent to Web based computing will mean that it is not an acceptable fit for many enterprise apps.  The opportunity to save money lies in knowing which apps will work best with cloud computing, if moving to the cloud will save you money in the long term, and how well those applications will perform when they’ve been moved to the cloud.

Perhaps the solutions that show the most promise (and this is idle speculation here) is an application that is based in the in-house datacenter but, if it should need additional processing beyond what in-house resources can provide, dynamically requisition additional computing power from the cloud.  Such an application would probably require cloud computing interoperability standards that are still in their nascent stages however – and you still need to monitor performance to know when more resources are needed, and if adding resources from the cloud actually improves the performance of the application.