Oracle Database Appliance as a Consolidation Platform

Posted in: Technical Track

I had the chance to talk to several Oracle Database Appliance users at the annual Collaborate 2012 conference last month in Las Vegas. A common theme in these discussions, as well as discussions with Pythian clients, is an interest in using the ODA as a large-scale consolidation platform. ODAs offer all the benefits of engineered systems: a fast, simple setup, a single-vendor support, and pre-validated configurations. At first glance, however, the fixed non-expandable capacity of an ODA would make it look like a poor candidate for consolidation, particularly the fixed 4TB of usable disk capacity and 24-core processing capacity.

But there’s no reason why you can’t buy more than one ODA to install next to each other. Sure it would be a completely separate database cluster with separate storage, but as long as any single database fits in 4TB, scaling can be done horizontally by adding more and more machines. Relocating databases between ODA machines is a bit of a chore, but can be accomplished with minimal downtime using well-known tools like Oracle Data Guard. And even though the process isn’t currently automated, there’s no reason it can’t be scripted. (And yes, I’ve already put in a request to have this simplified in a future version of the ODA software.)

(Readers may point out that external storage, say a ZFS appliance, is now supported, but I would argue that this erodes the benefits of having a single, self-contained, and pre-tested system.)

Who would be good candidates for this type of platform? It offers an entry level to Oracle’s engineered system portfolio. Small deployments will find ODA’s pay-as-you-grow CPU licensing attractive, with its ability to license as few as two processor cores. And for users doing database consolidation, it allows databases to be relocated one at a time and license expenses to be deferred until the processing capacity is actually required. ODA is also a great starting point for environments outgrowing existing single-instance infrastructure, offering a simple way to expand capacity through Oracle RAC clustering.

And no ODA discussion is complete without mentioning its much larger cousin, Exadata. ODA is no Exadata: no smart scans, no I/O resource management, no flash cache, and no hybrid columnar compression (ZFS appliance storage excepted). But many consolidation environments are OLTP-only and may rarely use smart scans, and lower cost may permit more storage to make up for the lack of columnar compression. The lack of flash cache and I/O resource management are tougher to work around and will likely manifest themselves with slower I/O performance and less-efficient I/O sharing at peak usage.

But the big upswing is lower cost. The ODA has a list price of $50k with 24 compute cores and 20 600G disk drives. The closest Exadata equivalent would be a quarter rack, also with 24 compute cores but with 36 600G disk drives, and more than 6 times the cost at $330k. Adding in additional storage server license costs under Exadata and the cost of ongoing support, ODA presents significant cost savings.



Want to talk with an expert? Schedule a call with our team to get the conversation started.

About the Author

Marc is a passionate and creative problem solver, drawing on deep understanding of the full enterprise application stack to identify the root cause of problems and to deploy sustainable solutions. Marc has a strong background in performance tuning and high availability, developing many of the tools and processes used to monitor and manage critical production databases at Pythian. He is proud to be the very first DataStax Platinum Certified Administrator for Apache Cassandra.

3 Comments. Leave new

Yury Velikanov
May 15, 2012 6:30 pm

+1. Loven the post Marc. Good summary.



I want to park multiple database at 1 place,whatis the best practice.

Marc Fielding
July 16, 2012 8:17 am

Hi Baqar,

To answer your question, I would need to know more about your requirements and your environment. For example, what do you mean by “park my databases in one place”? What benefits are you looking to achieve by doing this? How large are your databases in terms of disk space and workload? Is your environment large enough to consider Exadata a platform?

Cheers, and good luck on your project!



Leave a Reply

Your email address will not be published. Required fields are marked *