A Principled Technologies test report 2
Database performance and resiliency in the Intel processor-powered
Dell PowerEdge R630 running Microsoft SQL Server 2014
SAVE SPACE, GAIN PERFORMANCE + RELIABILITY
The PowerEdge R630 is a space-saving, 1U server designed by Dell to be capable
of supporting complex and demanding enterprise applications. With two or more of
these servers in a cluster running database software, such as our solution that ran SQL
Server 2014, the Dell solution can bring database resiliency and high availability to your
datacenter. We set out to test database performance with these features in place and
with different storage configurations.
To configure the servers for high availability, we set up a resilient two-node
cluster running Windows Server 2012 R2 and SQL Server 2014, using the AlwaysOn
technology from Microsoft SQL Server. AlwaysOn is a high-availability and disaster-
recovery feature available in SQL Server 2014.
To learn about the application performance the Dell PowerEdge R630, powered
by Intel Xeon E5-2660 v3 processors, can deliver on the cluster with the servers
configured with HDDs, we created and ran an online transaction processing (OLTP)
database workload with the benchmark DVD Store 2.1. We then repeated same test
with two additional Dell PowerEdge R630 servers, configured with 1.8-inch SATA SSDs.
The results of our test follow. For information on the server, software, and benchmark,
see Appendix A. For detailed system configuration information, see Appendix B. For
specifics of our testing, see Appendix C. For a cost comparison of our configurations, see
Appendix D.
WHAT WE FOUND
HDD configuration
We tested the HDD configuration of the server at medium and heavy OLTP
workload levels. First, we used a think time of 100 milliseconds, to represent a
moderate database workload. We configured a single SQL Server instance to host four
databases and then found that the Intel-powered Dell PowerEdge R630 HDD
configuration delivered a solid output of 20,736 orders per minute. Next, we increased
the intensity of the workload by changing the parameters to reflect an application with
higher utilization. As Figure 1 shows, decreasing the think time to 20 milliseconds,
representing a more difficult workload, boosted performance a modest 16.2 percent.
Think time refers to
the length of time
the virtual users
pause between
transactions,
simulating user
pauses during real
transactions.
(19 paginas)
Manymanuals.com
Manymanuals.de
Manymanuals.fr
Manymanuals.it
Manymanuals.pl
Manymanuals.cz
Manymanuals.es
Manymanuals-pt.com
Comentarios a estos manuales