Data is the single most valuable commodity in business today.  The rapid global adoption of mobile devices has led to a deluge of social media content, full-motion media, text, images and videos, creating massive opportunities to convert that information into analytics. As a result, there has been a great deal of discussion in the industry about database performance and in-memory databases, the goldmines of business data.

Most databases were optimally designed for storing predictable data types and sizes. Now, Dell is building a new appliance that can work for structured and unstructured databases: the Dell Database Acceleration Appliance for Oracle. Customers can quickly deploy these databases in flexible, high performance, low latency and highly scalable environments, avoiding the cost and performance concerns associated with traditional data storage. This enables our customers to quickly scale to handle the exponential growth in data demands.

The benefits of this appliance go far beyond simply managing data demands. Since it is deployed on standards-based Dell hardware, the appliance can work hand in hand with existing IT, including Dell Fluid Cache for SAN. And in benchmarking, we have found that we can offer substantial savings to cost, power, cooling and rack space, creating a more scalable and efficient database environment.

Under the Hood

The working set sizes of databases often can push server capacity to the limit and can be an obstacle to efficient scaling. While many database solutions reduce the working set by introducing complex configurations, this can bring performance degradation to that database. To address this, the Dell Acceleration Appliances for Databases recognize and map the database to the available resources within the appliance.  Overall performance can be improved by 10-40 percent, depending on the workload. The appliance also uses I/O acceleration to lever the DRAM in our Dell PowerEdge servers to boost performance. The I/O acceleration works to optimize server requests quickly from the DRAMs on the appliance. Finally, the solid state storage memory used in this appliance is as much as ten times the capacity of most servers using a PCI express slot.  It is also as much as one tenth of the cost, with the average appliance delivering this higher capacity at under $3/GB. The appliance will support MongoDB and Apache Cassandra, as well as Microsoft SQL Server 2014, MySQL, Sybase, Oracle and others.