High Performance Oracle Database with the Dell Acceleration Appliances for Databases   Leave a comment

The performance of many business critical applications like Oracle database is often limited by bottlenecks in the infrastructure components—CPU, memory, network, and storage. Although CPU, memory, and network performance has improved dramatically over time, storage performance improvements have lagged due to the limitations of latency and throughput.
The Dell Acceleration Appliances for Databases (DAAD) combines Dell servers with Sandisk’s flash storage technology to significantly improve storage latency and throughput. The appliance provides an adaptable, flexible, highly available, and scalable storage solution to various IT organizations’ database implementation. The Oracle RAC database on the DAAD delivers the highly available and high-performance database solution to business critical applications

Dell Acceleration Appliances for Databases (DAAD) Overview:
The Dell Acceleration Appliances for Databases (DAAD) is a pre-built, pre-integrated appliance designed to accelerate leading database environments such as Oracle Database. This appliance combines Dell PowerEdge R720 Server, Dell networking, and ioMemory cards and ION Accelerator software from Fusion-io to improve database performance.
DAAD Architecture:
High-level appliance architecture:
• The appliance is built on the industry leading Dell PowerEdge R720 server.
• The appliance supports high-speed Fibre channel or iSCSI networking between database servers and the
appliance nodes.
• Each appliance node is equipped with four cards, each of which has 3.0TB of usable, persistent flash storage
in the PCIe slots of the R720 server. These cards are also known as ioMemory cards.
• The ION Accelerator software installed on each appliance node enables management and configuration including
• For High Availability, the ION HA Clustering configuration consists of a pair of two appliance nodes
connected with a private 40GbE point-to-point Interconnection network. The appliances provide
scalability across multiple appliance nodes or multiple pairs of HA Clustered appliances.


Oracle RAC Database Reference architecture based on Dell Acceleration Appliances for Databases

To ensure the high availability of the infrastructure, it is recommended that network and storage IO paths are redundant. The following diagram shows the architecture of Oracle RAC database implementation based on DAAD as the shared storage.

The whitepaper that I co-authored with two other engineers in my team last July describes the reference architecture of an Oracle database and DAAD pairing, and showed the analysis of performance improvement of the Oracle database based on this architecture. The solution described in this whitepaper here is a high-availability-enabled, two-node appliance that uses the Fibre Channel storage protocol. Each appliance node is comprised of a Dell PowerEdge R720 Server, four 3.0TB ioMemory flash cards with the ION Accelerator software installed, and high availability ION clustering enabled. The highly available two-node appliance solution provides 12TB of redundant tier-1 storage capacity with a total of 24TB of raw storage capacity. The storage appliance delivers over 1 million random read IOPS with 0.5 millisecond latency for 4k random read which equates to 27 times the IOPS and 4% of the latency that 96x 15k conventional drives can provide. On HammerDB’ s TPC-C-like performance studies, the four-node Oracle RAC database on the two-node appliance can deliver 960,000 peak New Orders per Minute (NOPM) and over 2.5 million Transaction Per Minute (TPM).

To see my other technical white papers, please refer to the publication table of this Oracle blog:

Posted May 26, 2015 by kyuoracleblog in DAAD, Fusion IO, Oracle RAC, Sandisk

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: