By David Dorf on Oct 04, 2011
The big announcements at OpenWorld this year revolved around engineered systems, where software and hardware are optimized to work together. Larry Ellison kicked off the conference by explaining that Oracle's engineered systems use a "parallel everywhere" approach to squeezing greater performance out of commodity parts.
Below are six such systems and my thoughts on their applicability to retail:
There are several retail customers that are running Exadata with great results. When lots of data is involved, be it analytical or transactional, Exadata increases database performance, reduces storage and electricity costs, and simplifies through server consolidation. Exadata is perfect for running merchandising, supply chain, and data warehouse/BI applications.
Exadata's brother is Exalogic, the middleware machine. It shares much of the Exadata architecture but instead of focusing on data, its focused on fast application execution. It does particularly well with Java-based applications like ATG Web Commerce.
3. SPARC SuperCluster
While Exadata specializes in database workloads and Exalogic specializes in middleware workloads, the SPARC SuperCluster is a general-purchase system that handle both well. Think of the SuperCluster as Exadata and Exalogic merged together in single lower-cost rack running the new SPARC T4 and Solaris 11. This might be the best solution for mid-tier retailers that can't invest in Exadata and Exalogic separately.
When dealing with data, memory is faster than flash which is faster than disk. Exadata uses a cost-effective approach by using all three types of storage to maximize performance. Exalytics, on the other hand, is simply focused on speed and therefore memory. It's a specialized BI Machine that uses the TimesTen in-memory database to render very fast analytics "at the speed of thought." Essbase is also a database option. This is where real-time analytics and visualization shine. Retailers running Hyperion and/or OBIEE will benefit from Exalytics.
5. Big Data Appliance
With Twitter generating a billion tweets per week, we are being overwhelmed with data, and more specifically unstructured data. To help deal with this deluge of data, the Oracle Big Data Appliance includes the necessary hardware and software to acquire, manage, and analyze huge volumes of data. It includes open-source versions of Hadoop and R, popular choices for big data problems. Retailers that are serious about collecting social data about their brands, products, and customers will benefit from this solution.
6. Database Appliance
At the low-end, if your database needs don't require Exadata or Exalytics, then the Oracle Database Appliance is the perfect combination of hardware and database in a simple, cost-effective package. This data center simplification play will likely resonate with smaller mid-tier retailers.
Part 2 will discuss cloud computing.