It is not farfetched to look at scientific applications when seeking to trace the eventual trajectory of cloud computing. Grid, the closest thing to a precursor of cloud, pretty much began and ended without getting much out of the sciences. So we might keep half an eye on science and the cloud.
Out of curiosity as much as anything I recently looked at some of Microsoft Research's scientific Azure cloud projects and found it was a very various world. As one example, there are many ways to approach data handling in the cloud. The science folk are hacking at Azure to uncover clues in genetic code, watch the climate, figure the motions of the oceans and more.
For science, it seems, the semi-SQL or non-SQL data architectures - like those Amazon and Google pushed when cloud computing first emerged - are where the cloud action is for now. But there is a lot of existing SQL under the sky right?
It was news last year when Microsoft added a SQL-style data architecture to join the more file-system-like Azure Storage Services that comprised the Azure data architecture to that point. But they are not alone in looking for some ''SQLness'' to cloud computing.
Microsoft's ''storage services'' are various too. They include Table Services, Blob Services and Queue Services.
It is a lot to sort through and it all suggests that cloud computing may be a developer's chore - and not a CIO's button-push - for more than a short time.
The varied data architectures and other 'cloud enhancements' further suggest that cloud is going through growing pains where customers (just like home buyers or car buyers) say: ''I really like it but it would be great if you could just add ... '' Fill in the blank.
This is a familiar formula in technology adoption. Something comes along that is better, because it simplifies something that has become complex. But at least some of the complexity was there for a reason. Web services simplified things that were cumbersome in, say, CORBA. But once people added SOAP RPCs, authentication, guaranteed delivery, and what have you - well - the new technology was a bit different. It was a little more complex.
It will be interesting to see how this plays out in cloud computing because much of its promise was as a mass market entity. Some of the projections about cloud were based on ''cloud as a cookie cutter'' approach.
Let's add that Microsoft Azure was not cited here as an anomaly - although the fact that Microsoft marches to a slightly different beat than others tends to make it appear a bit anomalous.
If anything, Microsoft's experience could hold it in good stead on the cloud. It knows the role of 'good-enough' provider. In some ways it is an 80/20 house, meaning here one that gives people 80 percent of what they need at 20 percent of what the other provider is offering. This could position it to do well in cloud computing, where providers will have to carefully measure cookie cutting profit margins versus customization costs.
Leave a comment