Cloud computing architecture: Relational memories are made of this? |
 |
By Jack Vaughan
03 Nov 2009 | SearchSOA.com |
|
|
Curt Monash has followed the relational database world more closely than anyone over the years. SearchSOA.com reporter Rob Barry caught up with Monash for a story Barry was writing on data and the cloud…a sometimes contentious topic for sure.
Will the world rewrite its mass of apps for the new architecture? At one point, Google would have you believe they would. Given their success in wicked fast search engine technology, their arguments were heard. Their arguments - that relational architectures were overused, and that the speed of parallelism on a many-clustered cloud obviated the need for the relational store and called for embrace again of clever file structures, albeit wicked fast ones.
Given his track record, it is worthwhile to see what Curt Monash would have to say. In effect, he told Barry, there won't be much difference between the ways enterprises structure data in and out of public cloud.
"I don't think there will be much difference because of the work that's been put into parallelizing relational database use," said Monash. "Everybody should have data off premises that can be gotten in a hurry. And that has to be managed by similar technology."
This issue came to light again last week when Amazaon said it would add MySQL to an arsenal that heretofore was noted for reliance on SimpleDB.
People have written about the imminent death of the RDB before; some may recall that the object database was going to undo the RDB. The day may come when they are right. Still, the day seems far away when most apps are considered best served by non-relational means. To top it off, the skills in place and the work done to date must be considered. What do you think?
|
|
|
|
