As cloud computing matures there’s no doubt that much more emphasis will be placed on the elasticity of the data deployed across global networks of cloud computing platforms.
Today, what we have are multiple instances of separate databases running in different data centers. For example, Xeround, a provider of a MySQL-compatible database service in the cloud, just announced that it has deployed a version of its database-as-a-service in Europe.
Xeround CEO Razi Sharir says the next step will be to actually deploy distributed databases in the cloud, which will allow these databases to truly support the elastic requirements of cloud computing applications.
We’re not quite there yet in terms of actual customer demand. But it’s pretty clear this is the direction that things are heading in, especially when you consider that elasticity is supposed to be one of the fundamental tenets of cloud computing.
For many databases, however, supporting this level of elasticity might require some fundamental reworking of the underlying architecture. Databases have been distributed in terms of instances running on servers that have access to a high-speed network for a while. And Oracle’s grid architecture is probably the best known instance of a scalable distributed database computing environment.
But now we’re asking distributed databases to scale over wide area networks that introduce a lot of latency into the equation. Solving this challenge will require a lot of reliance on caching and more advanced management capabilities for resolving potential data conflicts.
But in time, distributed databases will meet the next generation of cloud computing requirements, which should remove one of the remaining hurdles to widespread adoption of cloud computing in the enterprise.