The upsurge in cloud computing presents development teams with new challenges and opportunities. There is a school of thought maintaining that the best bet is to upload applications as-is to the cloud, doing as little rework as possible. Yet another school contends that applications should be dramatically recast if they are to exploit cloud architecture to full advantage.
Can people learn from past experiences in parallelism? That is hard to say. Parallel and multiprocessor systems of the 1980s were highly specialized, succeeding at scientific tasks but not at broader applications. That was somewhat true for grid computing too.
Out of the parallel system ferment came Tuple spaces and then JavaSpaces as means to gain scalability in parallel processing using distributed object caching. Somewhat under the radar, technologists have worked to commercialize these highly parallel schemes in recent years, especially in applications aimed at traders on Wall Street. To get a view on how development issues will pan out in the cloud, we recently spoke with Nati Shalom, CTO, GigaSpaces, which offers a commercial version of JavaSpaces.
“If an application was originally built in a static way, moving that application unchanged to the cloud means you have simply placed it in another hosting environment,” said Shalom. Fair enough. For some people, that may be worthwhile. But by Shalom’s measure, that is a bit of a waste.
Still, changing an application to run well in the cloud is an adventure. You may fine tune your static app for a flexible treatment, take a cluster, and add machines, but no linear scale-up will occur.
Nothing will happen, or, just as likely, something will break, said Shalom. The natural stack is not built to take advantage of the new environment.
The bet here is that you have to re-think your problem, if you are looking for game-changing cloud benefit. But it is early in the game.
Ever faster processors and every cheaper blade computers have gotten us where we are as the cloud era begins. These underlying technologies have been the building blocks of several paradigm shifts. Mainstream developers have been relatively insulated.
Will that remain the case? It is hard to say, but it should prove to be interesting to watch.
Stay tuned for regular updates on this topic and let us know if there are particular issues around cloud computing for developers that you would like us to address.
Azure data architecture reworked as release date nears
Microsoft Azure’s data schemes have been altered on its way to a wide beta, which means better support for established SQL schemes.
Salesforce unfazed by Oracle competition in cloud computing
Oracle’s colorful CEO Larry Ellison took shots at upstart Salesforce.com, but the company remains undeterred by competition. Salesforce has enabled business-line managers to quickly build apps, without relying wholly on IT development resources.
VMLogix adds support for testing in the cloud
VMLogix Inc. jumped into the cloud with the beta release of LabManager – Cloud Edition for running virtual labs within the Amazon Elastic Compute Cloud (EC2).
Google App Engine plus Amazon AWS: Best of both worlds
Developers writing enterprise applications need both flexible resources and simplicity of use, according to a debate at JavaOne.