Dell POWEREDGE R720 - Manual (Page 4)

Consolidating Legacy OLTP Application Servers with the Dell 12 th Generation PowerEdge R720
Server
4
Executive Summary
With the introduction of the Dell™ 12 th generation server product line, customers have an
opportunity to lower total cost of ownership (TCO) by consolidating their legacy database
application servers. Reasons to consolidate database servers include: system underutilization,
hardware/software end of life (EOL), reducing costs associated with power and cooling, or
improving overall performance.
Dell strives to reduce the complexity of migrating from legacy database production
environments to the latest generation hardware and software while maintaining application
performance needs. The purpose of this document is to evaluate and highlight the performance
benefits of consolidating legacy online transaction processing (OLTP) database applications
running on Dell’s 9
th generation PowerEdge™ 2950 server onto Dell’s 12 th generation PowerEdge
R720 server.
Introduction
OLTP database applications are characterized by processing a large numbers of concurrent
transactions at a given time. For example, today’s banking institutions provide online access to
their customers where they can access account information, transfer funds, and pay bills.
Online banking applications rely on processing thousands of transactions in a time-sensitive
manner. The database server hosts where these transactions are processed require sufficient
resources in terms of CPU, memory, and fast data access for results to be generated instantly.
Legacy servers with lower performing CPUs and slower storage are no longer capable of
servicing the increased demands within reasonable response time.
Dell’s Solutions Performance Analysis team simulated an OLTP database workload to measure
the performance of legacy and current generation PowerEdge server, storage, and software
stack. This white paper provides customers with a performance comparison of a Microsoft SQL
Server
® database solution running on a legacy 9 th generation Dell PowerEdge 2950 and
Microsoft SQL 2005 to the new 12 th generation Dell PowerEdge R720 server and Microsoft SQL
2012 RC1.
Test methodology
Quest’s Benchmark Factory for Databases (BMF) is a database build and workload simulation
tool that measures database performance. To test the performance, a 200GB database
modeled after the industry-standard TPC-C benchmark was built on each of the solutions.
To characterize the legacy environment, we configured a PowerEdge 2950 connected to 2 Dell
PowerVault™ MD1120 storage arrays, running Microsoft Windows® 2003 Server R2 x64, and
Microsoft SQL Server 2005 Enterprise x64. Using the Quest’s Benchmark Factory TPC-C workload
profile, we populated the database with a scale factor of 3000. BMF simulates users issuing
TPC-C like SQL transactions that exercise each database’s storage sub-system with a random
read and write I/O data access pattern in user increments of 250. During each database user
simulation, BMF collected the total number of I/Os per second (IOPS), Average Query Response
Time (AQRT), and CPU utilization. The test was stopped after the average query response time
reached 2 seconds.
To characterize the R720 Solution environment, we configured a PowerEdge R720 server
connected to 2 Dell PowerVault MD1220 storage arrays, running Microsoft Windows 2008 Server