“The Great Cloud Computing Deception” is what I wanted to call the webinar that we (IBM) will be doing on July 28 jointly with xkoto. But since we wanted to be civil we settled on a more politically correct “Scalability in the Cloud: Fact or Fiction?” for a title.
Cloud Computing has a lot of promise. If you listen to the cheerleaders (at times myself included) it will cure cancer, eliminate hunger and global warming. I am starting to think that the hype around Cloud Computing has gotten out of hand and it actually threatens the success of this very promising direction in IT. Don’t take me wrong, I am not disillusioned in the Cloud. Quite the opposite; I am as bullish on it as I ever was but I think we need to inject a dose of reality and we need to acknowledge that to be successful with Cloud Computing you need to do more than just sign up for an account with Amazon Web Services and pay 10 cents per hour for your server.
One of the most hyped and the least understood attributes of the Cloud is “elasticity” or in other words ability to get compute capacity when you need it on a moments notice and to pay only for the capacity that you actually used. This is indeed a fantastic feature and for some it will result in thousands or even millions in cost savings and more agile IT. That is – if it can be achieved – and it is one big “IF” and it is precisely the subject of this webinar. So come join us for an hour on July 28 in the conversation on this topic. We will have Ariff Kassam, co-founder and CTO of xkoto, Rav Ahuja our IBM DB2 Cloud Computing Product Manager, Paul Lapointe, Solutions Architect from xkoto and me (Leon Katsnelson) share our opinions, do a demonstration and answer questions or engage in a debate if time permits. You can register free of charge here.
xkoto is a solid technology, and I wish we could implement it in our shop. It offers the holy grail for a DBA, where he is able to perform maintenance during the week and can take an active server out of the cluster for maintenance. He can also scale the database wide, on the fly instead of having to just get a beefier server and scale up.
This is not some hokey active-passive failover cluster hack that just crashes the database and filesystem and mounts it on another server. It’s a true active-active setup.
Regarding xkoto and cloud computing, IBM is in a unique position here to do something truly revolutionary. Or, get ahead of the coming change in the landscape on be an early pioneer.
What we have the opportunity here to do is to create a true clustered WAN solution, offering high availability, on demand resources, which can scale up or down as needed, and reduced network latency, by delivering improved application performance due to moving data sources closer to remote user offices.
This concept is trying to get branded by Cisco as the “intercloud.” Basically it’s a true active-active, peer to peer architecture, where the database and the app live in an environment spread over the WAN, with no true “primary” server. Failure on a single node does not bring down the application or the data source, as it has an insane amount of redundancy built into it, to the nth degree.
xkoto has provided the resiliency at the database layer. XtreemOS Linux (http://www.xtreemos.org/) now provides the OS. The app is the easiest one of all to add into the mix. Now you just need the network hardware, architecture, heavy security, and best practices to join it all together with. You need a Cisco to help you out here.
Only IBM can do this kind of vertical integration. Oracle thinks they can after they bought Sun, but the only thing they are good at vertically integrating is bloated license compliance models. Microsoft is the only other player that has access to all the pieces. Cisco is moving there but they have a steep learning curve on the application and DB side.
10 years ago IBM pushed Linux over the top and into the mainstream, and now an opportunity has arisen for the next generation of technology. Can they do it again ???