A Dell EMC Technical White Paper Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by Nutanix Software Dell EMC Engineering January 2017
Apr 14, 2018
A Dell EMC Technical White Paper
Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by Nutanix Software Dell EMC Engineering January 2017
2 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Revisions
Date Description
November 2016 Initial release
January 2017 Branding update
The information in this publication is provided “as is.” Dell Inc. makes no representations or warranties of any kind with respect to the information in this
publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose.
Use, copying, and distribution of any software described in this publication requires an applicable software license.
Copyright © 2017 Dell Inc. or its subsidiaries. All Rights Reserved. Dell, EMC, and other trademarks are trademarks of Dell Inc. or its subsidiaries. Other
trademarks may be the property of their respective owners. Published in the USA. [2/6/2017] [A Dell Technical White Paper] [1078-BP]
Dell EMC believes the information in this document is accurate as of its publication date. The information is subject to change without notice.
3 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Table of contents Revisions............................................................................................................................................................................. 2
Acknowledgements ............................................................................................................................................................. 5
Feedback ............................................................................................................................................................................ 5
Executive summary ............................................................................................................................................................. 5
1 Introduction ................................................................................................................................................................... 6
1.1 Objectives ........................................................................................................................................................... 6
1.1.1 Audience ............................................................................................................................................................. 6
1.2 Terminology ........................................................................................................................................................ 7
2 Product Overview ......................................................................................................................................................... 8
2.1 Dell EMC XC Series Overview ........................................................................................................................... 8
2.2 Benchmark Factory Overview ............................................................................................................................ 9
3 Solution infrastructure ................................................................................................................................................ 10
3.1 Physical system test configuration ................................................................................................................... 10
3.2 Dell EMC XC storage and cluster configuration ............................................................................................... 11
3.3 VM layout .......................................................................................................................................................... 12
3.4 Database layout ................................................................................................................................................ 13
3.5 Network configuration ....................................................................................................................................... 13
4 Test methodology ....................................................................................................................................................... 14
4.1 Test criteria ....................................................................................................................................................... 14
4.2 Measurement and monitoring ........................................................................................................................... 14
4.3 Test approach ................................................................................................................................................... 15
4.3.1 Baseline configuration ...................................................................................................................................... 15
4.3.2 Four-node configuration ................................................................................................................................... 15
4.3.3 Five-node configuration .................................................................................................................................... 16
4.3.4 Six-node configuration ...................................................................................................................................... 16
5 Test results ................................................................................................................................................................. 18
5.1 Baseline performance ....................................................................................................................................... 18
5.2 Performance scalability .................................................................................................................................... 19
5.2.1 Four-node test .................................................................................................................................................. 19
5.2.2 Five-node test ................................................................................................................................................... 20
5.2.3 Six-node test ..................................................................................................................................................... 21
5.3 Summary .......................................................................................................................................................... 22
4 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
6 Conclusion .................................................................................................................................................................. 25
A Configuration details ................................................................................................................................................... 26
B Additional resources ................................................................................................................................................... 28
5 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Acknowledgements
This white paper was produced by the following members of the Dell Storage team:
Engineering: Chuck Armstrong and Doug Bernhardt
Feedback
We encourage readers of this publication to provide feedback on the quality and usefulness of this information
by sending an email to [email protected].
Executive summary
Online transaction processing (OLTP) applications range from web-based e-commerce sites, to accounting
systems, to customer support programs and are at the heart of every business function. As the foundation for
a wide range of mission-critical applications, OLTP operations (particularly those utilizing Microsoft® SQL
Server® database platforms) depend on exceptional performance and reliability.
SQL Server is the most widely deployed database in the world. Moreover, data generated by SQL Server
applications can grow exponentially over time. Therefore, designing and managing a storage and compute
environment that can effectively accommodate performance, capacity and future growth requirements is a key
challenge for customers today. The architecture of the Dell EMC™ XC Series appliances provides an
outstanding way for customers to meet performance and growth requirements.
This white paper demonstrates the linear scalability of converged Dell EMC XC powered by Nutanix for critical
SQL Server workloads. Adding nodes to an existing Dell EMC XC Nutanix cluster without system downtime,
enables growth to existing compute and storage resources in a linear fashion. This allows you to provide the
right size for existing applications and rely on proven linear scale for future needs.
6 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
1 Introduction Different types of database applications have varying performance and capacity needs. Understanding the
models for common database application workloads can be useful in predicting the possible application
behavior in a given infrastructure environment. The most common database-application workload models are
OLTP and Data Warehousing (DW). This paper focuses on OLTP database workloads.
Storage and compute infrastructure supporting OLTP platforms like SQL Server needs to not only non-
disruptively scale in both performance and capacity, but also deliver the required high performance in an
optimal, cost-effective way. The Dell EMC XC Series platform makes it possible to run SQL alongside other
virtual machine (VM) workloads without compromising performance.
OLTP database applications are optimal for managing rapidly changing data. These applications typically
have many users who are performing transactions simultaneously. Although individual data requests by users
usually reference few records, many of these requests are being made at the same time.
This whitepaper describes various executed tests that were based on the above factors and key guidelines
and best practices based on the test results. These best practices facilitate achieving optimal performance of
Microsoft SQL Server OLTP database applications using Dell EMC XC Series clusters.
1.1 Objectives The testing presented in this paper:
Establishes baseline I/O performance characteristics of a SQL Server OLTP database by simulating a
TPC-E like workload on a three-Node Dell EMC XC cluster
Characterizes performance using Benchmark Factory™ for Databases, simulating Microsoft SQL
Server OLTP transactions on the three-Node Dell EMC XC cluster by simulating TPC-E like workload
Validates that the SQL Server database deployment can scale linearly by adding additional nodes to
the XC cluster
1.1.1 Audience This paper is intended for system administrators, database administrators and storage architects interested in
deploying MS SQL Server OLTP database solutions using Dell EMC XC series appliance. It is assumed that
the readers of this document have familiarity with Dell EMC XC Nutanix administration and Microsoft SQL
Server database installation and administration tasks.
7 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
1.2 Terminology The following terms are used throughout this document.
Benchmark Factory: A Dell database performance tool that can simulate actual database transactions such
as TPC-C and TPC-E using the TPC specifications. More details on this tool and the download location can
be found at http://software.dell.com/products/benchmark-factory.
TPC Benchmark™ E (TPC-E): A benchmark developed by TPC to simulate OLTP database workloads. The
TPC-E benchmark simulates the OLTP workload of a brokerage firm. Although the underlying business model
of TPC-E is a brokerage firm, the database schema, data population, transaction mix and implementation
rules broadly represent a modern OLTP database system. More information on TPC-E can be found at
http://www.tpc.org/tpce/default.asp.
Dell EMC XC Series: The Dell XC Web-scale Converged Appliance integrates the proven Dell x86 server
platform and Nutanix web-scale software to provide an enterprise-class converged appliance for virtualized
environments. Backed by the Dell Global Service and Support organization, the XC appliance consolidates
compute and storage into a single platform. This platform enables application and virtualization teams to
quickly deploy new workloads to their private cloud. Appliance details and configuration information are at
http://www.dell.com/us/business/p/dell-xc-series/pd.
Virtual Machine (VM): A software computer that runs an operating system and applications. For the
purposes of this paper, it is run within the two supported hypervisors, either VMware ESX or Microsoft Hyper-
V.
8 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
2 Product Overview
2.1 Dell EMC XC Series Overview The Dell EMC XC Web-Scale Hyper-converged infrastructure is a hyper-converged solution that combines
storage, compute, networking, and virtualization into an industry-proven x86 Dell PowerEdge Server running
Nutanix web-scale software. By combing the hardware resources from each server node into a shared-
everything model for simplified operations, improved agility and greater flexibility, Dell EMC and Nutanix can
deliver simple, cost-effective solutions for virtualized enterprise workloads.
The Dell EMC Web-Scale Hyper-converged infrastructure is a scale-out cluster of high-performance nodes, or
servers, each running a standard hypervisor and containing processors, memory, and local storage
(consisting of SSD Flash and high-capacity SATA disk drives). Each node runs virtual machines just like a
standard hypervisor host.
Nutanix node architecture
In addition, the Nutanix Distributed Storage Fabric (NDSF) virtualizes local storage from all nodes into a
unified pool. In effect, NDSF acts like an advanced NAS that uses local SSDs and disks from all nodes to
store virtual machine data. Virtual machines running on the cluster write data to NDSF as if they were writing
to shared storage.
9 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Nutanix platform architecture
NDSF is VM-centric and provides advanced data management features. It brings data closer to virtual
machines by storing the data locally on the system, resulting in higher performance at a lower cost. The
Nutanix Virtual Computing Platform has a horizontal minimum scale of three nodes and has no maximum
technical limit, enabling organizations to scale their infrastructure as needs grow.
NDSF delivers a unified pool of storage from all nodes across the cluster that leverages techniques including
striping, replication, auto-tiering, error detection, failover, and automatic recovery. This pool appears as
shared storage resources to VMs, enabling seamless support of features like vMotion, high availability (HA),
and distributed resource scheduler (DRS), along with industry-leading data management features. This high-
performance, scale-out architecture allows plug-and-play nodes on demand for building a cluster that easily
grows along with the enterprise.
2.2 Benchmark Factory Overview Benchmark Factory for Databases is a Dell database performance-testing tool that enables database
workload replay, industry-standard benchmark testing, and scalability testing. Using the incorporated load
testing tools, you can make changes to the database environment, while mitigating the risks of unavoidable
database changes such as patches and upgrades, operating system migrations, and adjustments to virtual
machine configurations. In addition, its proprietary workload capture and replay supports Oracle and SQL
Server databases. This workload replay and scheduling software helps you eliminate slow SQL database
performance and dramatically simplifies high performance database management. More information on
Benchmark Factory can be found at http://software.dell.com/products/benchmark-factory.
10 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
3 Solution infrastructure The test configuration and other solution components used as part of this solution are described in this
section.
3.1 Physical system test configuration The physical test configuration diagram for this environment starts with the basic three-node Dell EMC
XC Series cluster as shown in Figure 3.
Three-node Dell EMC XC Series cluster representation
Each subsequent run of testing included an additional Dell EMC XC Series cluster node and an additional
Microsoft SQL workload VM. The final run of testing included six Dell EMC XC Series cluster nodes and six
Microsoft SQL workload VMs, as shown in Figure 4.
Six-node Dell EMC XC Series cluster representation
11 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
3.2 Dell EMC XC storage and cluster configuration Each Dell EMC XC 630 Series node used for this testing is comprised of the following hardware components:
Two, 400GB SATA SSDs
Eight, 1TB 7.2K RPM SATA Disks
Two, 16-Core Intel E5-2698 v3 2.30 GHz CPUs
16, 32GB DDR-4 QR 2133MHz RAM Modules (512GB Total)
NDSF simplifies the storage layout of Dell EMC XC architecture. The following diagram illustrates the
simplified standard storage layout in a Dell EMC XC environment.
The local storage controller on each host ensures that storage performance as well as storage capacity
increases when additional nodes are added to the Dell EMC XC. Each Controller VM is directly connected to
the local storage controller and its associated disks. By using local storage controllers on each ESXi host,
access to data through NDSF is localized. It does not require data to be transferred over the network, thereby
improving latency. NDSF ensures that writes are replicated, distributing data within the platform for resiliency.
The minimum number of Dell EMC XC Series nodes in a cluster is three. When clustered together, create a
single storage pool and a single storage container within it. The storage container is presented to all nodes
within the cluster and is spread across all of the node disks.
The cluster, populated with all six nodes used for testing scalability, is shown from a hardware resource
perspective in Figure 5. The cluster provides resources required by VM workloads and arranges those
workloads in the most effective manner: placing VM workloads on nodes with available CPU and RAM
resources, as well as the best fit for local disk resources.
12 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Figure 5 Six-node XC 630 Series cluster resources
The XC 630 Series Nutanix-powered cluster used for this testing was designed and configured using best
practices from Nutanix available at http://go.nutanix.com/microsoft-sql-server-converged-infrastructure.html
Every attempt is made by the cluster to keep virtual machines and their associated storage on the same
cluster node for performance consistency. However, each cluster node is connected to, and communicates
with, the other nodes on a 10Gb network. This communication allows virtual machines and their associated
storage to reside on different cluster nodes. In such a case, the cluster will begin to move the data on that
virtual machine to the local cluster node using the same 10Gb network.
3.3 VM layout The baseline for this scaling study was executed using three XC 630 series cluster nodes (Minimum number
of nodes in XC cluster). The workload applied to the cluster was in the form of three virtual machines running
Microsoft Windows Server 2012 and Microsoft SQL Server 2014; one virtual machine per node.
Each virtual machine consisted of 24 virtual CPUs, 64 GB RAM, and thin-provisioned disks in the layout
described in Table 1.
Microsoft SQL Server Disk Layout
13 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
100GB Drive C: Windows OS Disk
80 GB Drive D: Windows Pagefile Disk
50 GB Drive E: Database Disk
50 GB Drive F: Database Disk
50 GB Drive G: Database Disk
50 GB Drive H: Database Disk
25 GB Drive T: TempDB Disk
33 GB Drive L: SQL Logs
VM settings followed the Nutanix best practices found at http://go.nutanix.com/microsoft-sql-server-
converged-infrastructure.html.
Scaling the test environment consisted of adding one XC 630 series cluster node and one Microsoft Windows
and Microsoft SQL Server virtual machine workload at a time. The same tests were executed following each
addition to the cluster.
3.4 Database layout The database used for this testing was created by Benchmark Factory for a TPCE-like workload. For this
workload, the size of the database is based on the benchmark scale specified for the test. A benchmark scale
of 10 was used for this test and resulted in approximately a 100GB database. The database consisted of four
40GB physical data files and a single 25GB log file. The files were sized with adequate free space to
accommodate the data generated during the test. Each data file was placed on a separate volume. The log
file was placed on its own volume as well. An additional drive was allocated for TempDB as a best practice
even though TempDB utilization was virtually irrelevant during the test.
3.5 Network configuration For this study of scale, the use of VM to VM communication was not required. Therefore, this environment did
not utilize the network configuration. However, networking best practices can be found in the Nutanix best
practices guide found at: http://go.nutanix.com/microsoft-sql-server-converged-infrastructure.html.
14 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
4 Test methodology A series of tests were conducted using Benchmark Factory to simulate a TPC-E like workload to understand
the performance and scalability characteristics of SQL server databases deployed on Dell EMC XC Nutanix
clusters. The test was run with a user load of 30, the default transaction mix and no delay in the transaction
latency. Typically, in TPC-E like tests there is a certain amount of latency included in the transactions to
simulate actual users. However, this test was not run to simulate actual users but rather to generate an
acceptable workload on the system. The duration of an entire test run was one hour. The first 15 minutes of
that hour were used as ramp-up time for the test to achieve a steady state. Once a steady state was
achieved, statistics were calculated based on the performance of the final 45 minutes of the run.
4.1 Test criteria The test study presented in this paper was executed to develop best practices for:
Maintaining the database server CPU utilization below a 70% average
Maintaining application response times below two seconds (the industry standard acceptable latency
for MS SQL server OLTP database applications)
Eliminating bottlenecks (such as database locking, user and system I/O wait, or others) while
executing the peak I/O load
Ensuring that workload performance scales in a linear fashion as additional cluster nodes and
application workload is added
4.2 Measurement and monitoring Detailed performance metrics were captured from various layers within the XC cluster.
Benchmark Factory
The following metrics were captured from Benchmark Factory while executing TPC-E database
transactions:
- TPS: Transactions per second, based on the simulated load
- Average Response time: The time from when SQL is sent until a response is received
- Average Transaction Time: The time from when a query is submitted until the result is returned
Windows Performance Monitor
The following metrics were captured from Windows Performance Monitor (perfmon) within the guest
during the tests:
- Disk Transfers/sec: Rate of read and write operations to disk
- % Processor Time: Percentage of elapsed time the processor spends on execution
XC Prism
15 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Prism, the GUI management utility, was used to monitor cluster activity and health during test runs.
Real time data during test runs indicated that the cluster remained healthy during the significant
workload testing.
4.3 Test approach A TPC-E like workload was simulated on SQL Server databases running on a three-node XC Nutanix cluster
to establish a performance baseline. After establishing the baseline, the number of nodes in the cluster was
incrementally increased to six to validate the linear scalability. The details of the different tests and the
configuration are described in the following section.
4.3.1 Baseline configuration The baseline performance characterization was performed on a three-node XC cluster, running three VMs,
each containing a SQL Server instance, a single database and an instance of Benchmark Factory.
Base test configuration
4.3.2 Four-node configuration Once baseline performance metrics were captured for three nodes, an additional XC node was added and the
tests were performed for a four-node configuration. The additional XC node contained the same hardware
configuration as the original three. The additional VM contained the same workload, software components
and configuration.
16 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Test 2: Four-node, four VM configuration
4.3.3 Five-node configuration After performance metrics were captured for four nodes, another XC node was added and the tests were
performed for a five-node configuration. Following the same test methodology, the additional XC node
contained the same hardware configuration as the original three. The additional VM contained the same
workload, software components and configuration.
Test 3: five-node, five VM configuration
4.3.4 Six-node configuration The final scalability test was performed with six nodes using the same test methodology as the previous tests.
17 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Test 4: six-node, six VM configuration
18 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
5 Test results The TPC-E-like tests were run on four separate configurations. First, a three-node configuration was tested to
establish a performance baseline. Then, a test was run with an additional Dell EMC XC630 node for the
cluster and an additional VM to test a four- node configuration. This process was repeated for the five-node
and six-node configurations. The results from these two tests are described in Section 5.2.
5.1 Baseline performance To establish the performance baseline, the OLTP workload was run on a three-node XC cluster with three
VMs. Each VM contained a single instance of SQL Server, a single database, and Benchmark Factory. The
number of users was adjusted within Benchmark Factory to produce a realistic production workload in terms
of CPU and I/O. The performance in terms of Transactions per Second, CPU, and I/O are shown below. In
addition, all instances reported an application response time of 15ms. (Note – All performance measurements
are from the workload VM / application perspective)
Transactions per second and transaction time per VM for baseline (three-node) test
1879 1945 1921
15 15 15
SQL01 SQL02 SQL03
Trans Per Sec
Trans Time (ms)
19 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
CPU utilization and disk requests/sec (IOPS) per VM on baseline (three-node) test
As you can see from the results, consistent performance was observed on all three VMs hosting SQL Server
databases with very low latency when a realistic TPC-E like production database workload was simulated on
a three-node XC cluster.
5.2 Performance scalability
5.2.1 Four-node test This test was performed by adding a Dell EMC XC630 node to the cluster and a VM to the three-node XC
cluster configuration.
Transactions per second and transaction time per VM on a four-node test
3102 3202 3192
65% 64% 63%
SQL01 SQL02 SQL03
IOPS
% CPU Time
1879 1885 18641967
15 15 15 15
SQL01 SQL02 SQL03 SQL04
Trans Per Sec
Trans Time (ms)
20 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
CPU utilization and disk requests/sec (IOPS) per VM on a four-node test
When the fourth node was added, linear scalability in performance was observed. The average latency
remained very low on all VMs.
5.2.2 Five-node test
Transactions per second and transaction time per VM on a five-node test
3157 3122 30843289
64% 64% 65% 62%
SQL01 SQL02 SQL03 SQL04
IOPS
% CPU Time
1906 1887 1928 1979 1961
15 15 15 15 15
SQL01 SQL02 SQL03 SQL04 SQL05
Trans Per Sec
Trans Time (ms)
21 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
CPU utilization and disk requests/sec (IOPS) per VM on a five-node run
When the fifth node was added, linear scalability was observed.
5.2.3 Six-node test
Transactions per second and transaction time per VM on a six-node test
3172 30823202 3293 3260
64% 64% 65%62% 62%
SQL01 SQL02 SQL03 SQL04 SQL05
IOPS
% CPU Time
18721969
18701953 1951 1938
15 15 15 15 15 15
SQL01 SQL02 SQL03 SQL04 SQL05 SQL06
Trans Per Sec
Trans Time (ms)
22 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
CPU utilization and disk requests/sec (IOPS) per VM on a six-node run
Adding a sixth node further improved the performance in a linear fashion without increasing latency. This
test clearly demonstrates the fact that as nodes are added, Dell EMC XC appliances are able to provide
linear scalability in performance without compromising latency.
5.3 Summary The tests results in Sections 5.1 and 5.2 clearly show the linear scalability of MS SQL Server database
deployed on Dell EMC XC Nutanix clusters. In all tests, throughput and system performance was linear when
scaling from a base configuration of three-nodes up to six-nodes.
The following chart shows the scale of this test environment. The blue line indicates the average TPS for all
SQL Server VMs included in the test run. The colored points on the chart show the TPS for each SQL Server
VM in the test run as a percentage above or below the average.
As shown in the chart, the linear scalability of the Dell EMC XC Nutanix cluster was exceptional. The
individual throughput measured in TPS for all virtual machines only varied 4% from the average.
30683259
30803268 3242 3217
63% 61% 62% 61% 61% 62%
SQL01 SQL02 SQL03 SQL04 SQL05 SQL06
IOPS
% CPU Time
23 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
TPS Scale: Average and differentials for each test run
Additionally, the CPU and Disk I/O utilization from Prism, the Nutanix graphical user interface, for the whole
cluster during the final round of testing (the six-Node test), can be seen in Figure 18.
XC cluster real time performance data during the six-node test
Figure 19 shows CPU utilization of each of the VM workload SQL servers during the final round of testing.
24 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
XC cluster real time performance data during the six-Node test
25 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
6 Conclusion The Dell EMC XC Nutanix platform is well equipped to handle the throughput and transaction requirements of
a demanding Microsoft SQL (MSSQL) server, given NDSF localized I/O and server-attached flash.
Nutanix offers platforms with a range of form factors and performance characteristics that enable you to
match the workload with the most effective platform. This modular approach enables easy scaling for MSSQL
deployments. Start with a small environment and linearly scale up as demands increase, adding both
compute and storage at the same time. A Nutanix powered data center provides scalable, high performance
MSSQL servers while also providing the ability to run many types of virtual workloads.
The test results in this paper have illustrated the linear scale capability of the Dell EMC XC Nutanix platform
for SQL Server workloads. Both database throughput and system resource utilization were proven to scale in
a linear fashion by simply adding nodes to the cluster as required.
To achieve optimal database performance, adhere to the applicable best practices presented in this paper. It
must be ensured that the entire ecosystem, including server resources and network switches, are sized and
configured appropriately to meet the workload performance requirements. Also ensure that the operating
system and databases are configured with the optimal settings and design recommendations mentioned in
this paper.
26 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
A Configuration details
This section contains an overview of the configurations used for the testing described in this document.
Test configuration - Hardware components
Solution Configuration - Hardware Components: Description
XC Series Nodes (Compute and Storage) Nutanix Operating System version (NOS) 4.1.2.1
XC 630 Series Web-scale Converged Appliance (2) 16-Core Intel E5-2698 v3 2.30 GHz CPUs (16) 32GB DDR-4 QR 2133MHz RAM (512GB Total) (2) 400GB SATA SSDs (8) 1TB 7.2K SATA Disks (2) Intel I350 Gigabit NICs (1) Intel 82599 2-Port 10 Gigabit NIC (1) Intel X520 2-Port 10 Gigabit NIC
Testing utilized 3 nodes for the baseline. Node count was incremented up by one for each successive test run. Intel I350 NICs are used for Hypervisor management. Intel 82599 NIC ports are used for Cluster operations between cluster nodes and CVMs. Intel X520 NIC ports are used for “Production” network traffic, – Not used for this testing.
Management server Nutanix Control Virtual Machine (CVM) (8) vCPUs (32) GB RAM
The cluster is managed via a CVM on each cluster node.
Network Dell Force10 S4810 switch Firmware: 8.3.12.0
This 10Gb switch was used for Cluster operations; connecting all Cluster nodes and all CVMs.
27 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
Test configuration - Software components
Test configuration - Software components
Database Server SQL Server 2014 SP1 Enterprise Edition Build 12.0.4100.1
Testing Application Dell Software Benchmark Factory for Databases Version 7.2
Operating System Windows Server 2012 R2 Standard Edition
28 Virtualized SQL Server Performance and Scaling on Dell EMC XC Series Web-Scale Hyper-converged Appliances Powered by
Nutanix Software | 1078-BP
B Additional resources
Support.dell.com is focused on meeting your needs with proven services and support.
DellTechCenter.com is an IT Community where you can connect with Dell Customers and Dell employees for
the purpose of sharing knowledge, best practices, and information about Dell products and your installations.
Referenced or recommended publications:
Dell EMC XC Series Appliances powered by Nutanix - Documentation:
http://en.community.dell.com/techcenter/storage/w/wiki/11457.dell-xc-series-appliances-powered-by-
nutanix-documentation
Dell EMC XC Series Web-Scale Converged Appliance:
http://en.community.dell.com/techcenter/storage/w/wiki/11454.dell-xc-series-web-scale-converged-
appliance