Nice work Xeround
Great to see Xeround launch their GA this week. The emergence of NewSQL as the fulfillment of the academic observation that transactions and SQL can theoretically scale horizontally is underlined by the existence proof of a vendor with a GA product that makes those claims.
Xeround are highlighting two key elements of the NuoDB (fka NimbusDB) world view. Firstly, that data management in the cloud needs to scale-out, not scale up. This is not a minor distinction, or one that can be papered over by putting steam-belching big iron “in the cloud”. Larry Ellison in his “What is the cloud anyway? It’s just a bunch of computers on networks” phase was making a case for scale-up by focusing on the demand side of cloud computing. It’s certainly true to say that in cloud computing the user does not know or care how the services are delivered. Why should the user care whether services are delivered by one monster machine or a gazillion cheap blades?
Of course the user cares and the datacenter owner cares too, in both cases because they care about a) service level guarantees, b) control, and c) costs. The bottom line is that scale-up economics cannot win that battle in the long run. If scale-out can be delivered for SQL/transactional databases then the user and CIO both get to a radically better SLA/control/cost envelope. It can be delivered, is being delivered, and clearly this is where users will be going over time.
The second piece of what Xeround are doing that is directionally significant is their delivery of their product as a service. Everything is moving towards the utility model. Power stations in your back yard are no longer the model. The question of what exactly data management will become in a cloud computing world is a big one, DBaaS being one part of the picture. But you can take to the bank the idea that applications will become less tightly coupled to their data services, and those will increasingly be delivered by some kind of utility model.
We have different ideas about how to do all of this technically (it’s extremely hard to do), how to package it (for clarity NuoDB/NimbusDB is a downloadable product, not principally a service), and how to build a robust business around it, but Xeround are to be congratulated on getting to this point.
We wish them well!