Bob Zurek

Everyday I take time out to read about what’s currently going on in the database market. Sources include blogs, analyst perspective like Curt Monash and his DBMS2.com <http://DBMS2.com>  website, emails from peers in the industry and press articles and releases, just to name a few. Everyone has an opinion about what is transpiring in the market. As you know, columnar databases, key/value stores, NoSQL solutions (I prefer to call them NotSQL instead), Map/Reduce and even internally invented data stores have all become quite visible in the market and there are certainly lots of opinions about them all. One thing is for sure, the data management landscape is really changing fast. Frankly, I love it.  It shows that innovation is alive and kicking and this is certainly the case at Infobright where we are embarking on our next adventure in the database market as Mark Burton pointed out in his most recently CEO blog on our website.

Another widely discussed topic in this market is performance and database scalability as data growth continues full steam. At Infobright our core focus has always been to optimize our solution in order to extract as much performance as possible out of SMP hardware without requiring our customers to go to a multi-node deployment. Our Knowledge Grid combined with our columnar store has been a great innovation for delivering high performance without hardware complexity or high cost. Our job has always been to reduce the complexity of deploying a database that scales from 100's of gigabytes to tens of terabytes, which is the mainstream of database deployments these days. In a recent blog post by Tony Bain from ReadWriteWeb.com <http://ReadWriteWeb.com>  entitled "Is the Relational Database Doomed?" he states:

Relational databases scale well, but usually only when that scaling happens on a single server node. When the capacity of that single node is reached, you need to scale out and distribute that load across multiple server nodes. This is when the complexity of relational databases starts to rub against their potential to scale. Try scaling to hundreds or thousands of nodes, rather than a few, and the complexities become overwhelming, and the characteristics that make RDBMS so appealing drastically reduce their viability as platforms for large distributed systems.

Our customers tell us that we do scale well without requiring them to buy and deploy lots of hardware, just like Tony points out in his post. The big question is what is "large"? Industry research shows that well over 75% of the data warehouses deployed in the market are in the range of 150 gigabytes to 10 terabytes. If you have very high-end requirements, then get yourself a forklift and bring in Teradata, as they handle that well but at significant cost and complexity.

Our job has always been to extract as much out of an SMP deployment as possible with our solution. To further extend what we can do well in a SMP environment, you will see us continue to reach new performance heights with the "scale up" work we are doing in the next release of our software, coded named Cypress. In this release you will see significant performance enhancement through our core next generation workload and memory management infrastructure. These features combined with a set of new unique features in our Knowledge Grid (patent pending efforts under way) will deliver even more performance without complex hardware configurations. This work will clearly be very beneficial as we move into offering a simplified multi-node option in the future.