Inevitably, cloud announcements, like Amazon's introduction of Elastic Beanstalk, sound good, but there are limits.
The possibilities of cloud computing are nearly limitless. At the same time, the opportunity to step in a hole while sprinting toward cloud computing still remains extremely high. I'm thinking, for example, of Amazon's recent announcement of its Elastic Beanstalk scaling service.
Elastic Beanstalk is an automated service that can sub for the former manual auto-scaling designations set by an Amazon customer. If traffic builds on your application, Elastic Beanstalk will clone your workload instance and run it alongside the original until the traffic challenges the two of them, when it repeats the process. You determine what type of virtual server you initially want, in terms of CPU and memory. Then if the need arises, Beanstalk propagates identical copies without you doing anything further.
All of this is a good start, especially for a free service, but only a start. As I noted in Wednesday's InformationWeek story, Amazon's Beanstalk service is limited to Java applications, and I could have added, only those Java applications that use Apache Tomcat as the application server -- a major subset of a subset of cloud applications.
Some of today's more weighty cloud applications requiring automated scaling are built on the LAMP stack, not Tomcat. Or they are production systems cranking away in conjunction with Red Hat's JBoss application server or a WebSphere or WebLogic server from IBM or Oracle. I'm going to say that Tomcat applications are those least likely to tax the scaling management capabilities of Beanstalk, which of course may be why Amazon is starting there.
This announcement reminds me of Google's early statements about App Engine, which was a valuable, managed cloud environment for enterprising developers, with Gadgets and connectors to Google Docs and other apps, provided you wanted to program in Python, specifically Python release 1.2. That was it. Much of Google's programming was done in Python, so why wouldn't the rest of the world be happy with that option? Google of course added Java in April 2009. (Beware of the cloud maker's mindset.)
SaaS As Innovation Driver?Software as a service is the clear No. 1 way enterprises consume cloud. InformationWeek's SaaS Innovation Survey reveals three tips to get the most from SaaS: Make it a popularity contest. Have an escape plan. And remember that identity is the new perimeter.