A Dell Technical White Paper
Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak WAN optimization testing for a DaaS environment
Dell Wyse Solutions Engineering June 2014
2 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Revisions
Date Description
June 2014 Initial release v.1.0
3 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
THIS WHITE PAPER IS FOR INFORMATIONAL PURPOSES ONLY, AND MAY CONTAIN TYPOGRAPHICAL ERRORS AND
TECHNICAL INACCURACIES. THE CONTENT IS PROVIDED AS IS, WITHOUT EXPRESS OR IMPLIED WARRANTIES OF
ANY KIND.
© 2014 Dell Inc. All rights reserved. Reproduction of this material in any manner whatsoever without the express
written permission of Dell Inc. is strictly forbidden. For more information, contact Dell.
PRODUCT WARRANTIES APPLICABLE TO THE DELL PRODUCTS DESCRIBED IN THIS DOCUMENT MAY BE FOUND
AT: http://www.dell.com/learn/us/en/19/terms-of-sale-commercial-and-public-sector Performance of network
reference architectures discussed in this document may vary with differing deployment conditions, network loads, and
the like. Third party products may be included in reference architectures for the convenience of the reader. Inclusion
of such third party products does not necessarily constitute Dell’s recommendation of those products. Please consult
your Dell representative for additional information.
Trademarks used in this text:
Dell™, the Dell logo, Dell Boomi™, Dell Precision™ ,OptiPlex™, Latitude™, PowerEdge™, PowerVault™,
PowerConnect™, OpenManage™, EqualLogic™, Compellent™, KACE™, FlexAddress™, Force10™ and Vostro™ are
trademarks of Dell Inc. Other Dell trademarks may be used in this document. Cisco Nexus®, Cisco MDS®
, Cisco NX-
0S®
, and other Cisco Catalyst®
are registered trademarks of Cisco System Inc. EMC VNX®
, and EMC Unisphere®
are
registered trademarks of EMC Corporation. Intel®
, Pentium®
, Xeon®
, Core®
and Celeron®
are registered trademarks of
Intel Corporation in the U.S. and other countries. AMD®
is a registered trademark and AMD Opteron™, AMD
Phenom™ and AMD Sempron™ are trademarks of Advanced Micro Devices, Inc. Microsoft®
, Windows®
, Windows
Server®
, Internet Explorer®
, MS-DOS®
, Windows Vista®
and Active Directory®
are either trademarks or registered
trademarks of Microsoft Corporation in the United States and/or other countries. Red Hat®
and Red Hat®
Enterprise
Linux®
are registered trademarks of Red Hat, Inc. in the United States and/or other countries. Novell®
and SUSE®
are
registered trademarks of Novell Inc. in the United States and other countries. Oracle®
is a registered trademark of
Oracle Corporation and/or its affiliates. Citrix®
, Xen®
, XenServer®
and XenMotion®
are either registered trademarks or
trademarks of Citrix Systems, Inc. in the United States and/or other countries. VMware®
, Virtual SMP®
, vMotion®
,
vCenter®
and vSphere®
are registered trademarks or trademarks of VMware, Inc. in the United States or other
countries. IBM®
is a registered trademark of International Business Machines Corporation. Broadcom®
and
NetXtreme®
are registered trademarks of Broadcom Corporation. QLogic is a registered trademark of QLogic
Corporation. Other trademarks and trade names may be used in this document to refer to either the entities claiming
the marks and/or names or their products and are the property of their respective owners. Dell disclaims proprietary
interest in the marks and names of others.
4 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Table of contents Revisions ............................................................................................................................................................................................. 2
Executive Summary .......................................................................................................................................................................... 5
VMware Horizon DaaS ..................................................................................................................................................................... 5
1 Configuration of the test environment .................................................................................................................................. 6
1.1 Hardware component details ....................................................................................................................................... 6
1.2 Software component details ......................................................................................................................................... 7
1.3 PCoIP configuration details........................................................................................................................................... 7
1.4 Networking configuration diagram ............................................................................................................................. 8
2 Test results and analysis ........................................................................................................................................................... 9
2.1 Test case 1 ........................................................................................................................................................................ 9
2.1.1 Application bandwidth reduction ............................................................................................................................... 10
2.1.2 Peak appliance throughput .......................................................................................................................................... 11
2.2 Test case 2 ...................................................................................................................................................................... 12
2.2.1 Application bandwidth reduction ............................................................................................................................... 12
2.2.2 Peak appliance throughput ......................................................................................................................................... 13
2.3 Test case 3 ...................................................................................................................................................................... 15
2.4 Results summary ........................................................................................................................................................... 15
3 Conclusion ................................................................................................................................................................................ 16
About the authors ........................................................................................................................................................................... 16
5 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Executive Summary
This document contains the architecture and results for testing conducted to investigate the impact of
Windows 7 on VMware Horizon DaaS 6.0 using Ivy Bridge processors, and determine the functionality of
the Silver Peak wide area network (WAN) optimization appliances in an emulated WAN environment.
VMware Horizon DaaS
VMware Horizon DaaS enables companies to deliver virtual workspaces to their end users – including full
Windows Client desktops, shared desktops and applications – as a monthly subscription service
VMware enables a risk-free incremental evolution of the next-generation workspace. It provides a
complete virtual workspace from the cloud, delivering desktops and applications as an easily-managed,
integrated cloud service. Enterprises can rapidly provision desktops and applications to users on any
device, anywhere. And virtual desktops are transformed from capital expenditures inherent in onsite
enterprise desktop virtualization and physical PC refreshes, into a predictable operating expenditure.
6 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
1 Configuration of the test environment
1.1 Hardware component details
Hardware Components: Description
VMware Compute
Host
1 x Dell PowerEdge R720 Servers:
VMware ESXi 5.1 Update 2
2 x E5-2690 v2 @ 3.0 GHz 10c Ivy Bridge Processors
256GB Memory @ 1600MT/s
12 x 300GB 15K SAS internal disk drives (Raid 10)
Broadcom BCM5720 1GbE QP NIC
iSCSI: Broadcom 57810
PERC H710P RAID Controller
BIOS version: 2.0.18 (or higher)
This system hosts the Windows 7 VDI desktops.
VMware
Management Host
1 x Dell PowerEdge R720 Servers:
VMware ESXi 5.1 Update 2
2 x E5-2690 v2 @ 3.0 GHz 10c Ivy Bridge Processors
256GB Memory @ 1600MT/s
12 x 300GB 15K SAS internal disk drives (Raid 10)
Broadcom BCM5720 1GbE QP NIC
iSCSI: Broadcom 57810
PERC H710P RAID Controller
BIOS version: 2.0.18 (or higher)
For the ESXi environment, 12 x 146 GB drives were configured in a RAID 10 array.
Each VM hosted the following workloads on Windows server 2008 R2 SP1 and Server 2012 R2:
VMware vCenter
File Server
Domain Controllers
Stratusphere Hub (5.6.2)
VMware Horizon DaaS appliances
Test Launcher Server
(VDI Workload
Generator)
2x Dell PowerEdge R710 servers:
VMware ESXi 5.5
Intel-(R)-Xeon- (R) CPU X5670 @ 2.9 GHz
96 GB RAM @ 1333 MHZ
10 x 146GB 15K SAS internal disk drives
Broadcom 5709 1GbE NIC
Login VSI version 4.1
VMware ESXi 5.5 installed on the Test Mgmt R710 server.
Login VSI launchers created on this server to initiate VMware View sessions from each VM to simulate VDI workload.
The VMs were loaded with Windows 8.1 Enterprise 64-bit operating systems.
Network 1 x Dell PowerConnect 6248 1Gb Ethernet Switch
2 x Dell Force10 S4810 switches
3 x WANem
Deployment Stack and Test Stacks configured on separate switches separated by physical switches and virtual WANem appliances.
7 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Performance
Monitoring
VMware Virtual Center 5.1
Silver Peak GMS
Performance data captured from VSphere client and GMS web interface using scheduled reports
WAN Optimization Silver Peak GX-V v6.2.4
Silver Peak VX-7000 v6.2.4.0_49732
The VX-7000 appliances will be used to optimize WAN traffic
1.2 Software component details
Software Components Description/Version
VMware vCenter Version 5.1
VMware Horizon DaaS Version 6.0
Hypervisor VMware ESXi 5.1 Update 2
Windows 7 Enterprise (32-bit) Virtual Desktops
Windows 8.1 Enterprise (64-bit) Login VSI launchers
Windows Server 2012 R2 Infrastructure VMs
1.3 PCoIP configuration details The gold image used for testing was configured using VMware’s PCoIP best practices. These settings are
shown on page 20 in the “VMware Horizon View Best Practices Performance Study” white paper found
here. The settings are intended for VMware Horizon View environments, but also apply to VMware Horizon
DaaS desktop images using PCoIP. The following PCoIP settings were applied to the gold image in the
test environment, but it should be noted that the optimal settings for any other deployment may vary
greatly depending on the operating environment, WAN bandwidth and quality of the link.
Setting Default Configured Description
Build to lossless On Off Enabled ability to enable or disable build to lossless
Maximum session bandwidth
90,000 Kbps 90,000 Kbps Specifies the maximum bandwidth in a PCoIP session
PCoIP session bandwidth floor
No minimum 0 Specified a lower limit, in kilobits per second, for the bandwidth that is reserved by the PCoIP session.
Session Audio BW limit 500 Kbps 100 Kbps Reduces bandwidth usage of audio with usable quality
Minimum Image Quality 50 50
Maximum Initial Image Quality
90 70
Maximum Frame rate 30 15 In WAN conditions, this will be helpful for video playback and fast graphics operations
8 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
1.4 Networking configuration diagram The following network diagram shows the logical layout of the test environment. There are geographically
separated service provider, data center, and client sites. Between each of the sites, are Silver Peak and
WAN emulation (WANem) appliances to simulate a WAN environment.
9 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
2 Test results and analysis The primary focus of the tests was to determine the effect of using Silver Peak WAN optimization
appliances with Windows 7 using VMware Horizon DaaS 6.0 on Ivy Bridge processors. Aspects that were
investigated were bandwidth reduction, error correction, latency, and throughput. In each test, 110
Enhanced users were used as this desktop density was well within the capabilities of the PowerEdge R720
compute host and SolidFire SF-3010 storage array at the VMware DaaS Service Provider site. Note that 110
Login VSI Enhanced user workloads were not chosen to test scale of server or WAN but rather to generate
an application WAN traffic payload across the back haul to the corporate data center from the desktop
infrastructure in the cloud.
Primary testing objectives:
Validate that the Silver Peak WAN optimization (WANOP) appliances function as expected
Determine what affect the WANOP devices have on user experience and performance
Determine what affect the WANOP devices have on application traffic
The following table lists the tests that were performed and the parameters that were variable between
tests. The Service Provider (SP), Customer Datacenter (DC), and Client Site (CS) latencies and loss were
modified as follows:
Test # SP to DC Latency CS to SP Latency
1 40 ms @ 1% loss 38 ms
2 80 ms @ 1% loss 38 ms
3 40 ms @ 1% loss 120 ms @ 2 % loss
Note: Test 3 is quantified by LoginVSI client-side test tools (or other client-side test tools)
In all tests, the Silver Peak maximum bandwidth was set to 155,000 Kbps. Also, forward error correction
(FEC) setting was set to “enable” with the default ratio of 1:10. The FEC setting is described in the following
screenshot:
Each test was conducted twice. Once with the Silver Peak appliance optimizations disabled and once with
the optimizations enabled to determine the effect of the optimizations on network traffic and desktop
experience.
2.1 Test case 1 The following validation was done for 110 Enhanced users using the standard Performance Analysis and
Characterization (PAAC) testing methodology called out in the Reference Architecture document.
Test # 1 SP to DC Latency - 40 ms @ 1% loss CS to SP Latency - 38 ms
10 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
2.1.1 Application bandwidth reduction As seen in the following diagrams, LAN (light blue) and WAN (dark blue) bandwidth are the same in the
diagram on left. This indicates that all data transferred on the LAN also had to traverse the WAN to access
the file or web services.
In the diagram on the right, WAN traffic is greatly reduced while LAN traffic remains very high. This
confirms that the Silver Peak WAN appliances served the CIFS and HTTP resources out of cache and
reduced the WAN traffic by 98%.
PCoIP traffic remains unchanged as a major portion of the packets are encrypted and look unique from
the perspective of the Silver Peak WAN optimization appliances and thus are not served from cache.
The diagrams below depict the application reduction by time for the duration of the test. As seen in the
diagram on the left, there is no reduction of data and the LAN and WAN traffic are nearly identical. In the
diagram on the right, we can see that the data that traversed the WAN was drastically reduced while the
LAN traffic remained very high.
11 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
2.1.2 Peak appliance throughput In the diagrams below, the peak throughput can be seen on both the LAN and WAN for all appliances. In
the diagram on the left, peak throughput for the WAN is 140 Mbps while the LAN is approximately 153. The
peak throughput is very close in this scenario as the data is not cached on the Silver Peak appliances.
In the diagram on the right, the peak WAN throughput is less than 10 Mbps while the LAN is actually
slightly faster than the gigabit Ethernet that was used for the test. This is most likely due to the fact that the
Silver Peak VX-7000 was caching the CIFS and HTTP traffic and was connected to the same VMware
vSwitch as the tenant desktops.
The following diagrams, the throughput over the WAN and LAN are shown for the duration of the test.
Figure 1 Silver Peak WAN optimization disabled
12 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Figure 2 Silver Peak WAN optimization enabled
2.2 Test case 2 The following validation was done for 110 Enhanced users using the standard PAAC testing methodology
called out in the Reference Architecture document.
Test # 2 SP to DC Latency - 80 ms @ 1% loss CS to SP Latency - 38 ms
2.2.1 Application bandwidth reduction As seen in the following diagrams, LAN (light blue) and WAN (dark blue) bandwidth are the same in the
diagram on left. This indicates that all data transferred on the LAN also had to traverse the WAN to access
the file or web services.
In the diagram on the right, WAN traffic is greatly reduced while LAN traffic remains very high. This
confirms that the Silver Peak WAN appliances served the CIFS and HTTP resources out of cache and
reduced the WAN traffic by 80%.
PCoIP traffic remains unchanged as a major portion of the packets are encrypted and look unique from
the perspective of the Silver Peak WAN optimization appliances and thus are not served from cache.
13 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
The diagrams below depict the application reduction by time for the duration of the test. As seen in the
diagram on the left, there is no reduction of data and the LAN and WAN traffic are nearly identical. In the
diagram on the right, we can see that the data that traversed the WAN was drastically reduced while the
LAN traffic remained very high.
2.2.2 Peak appliance throughput In the diagrams below, the peak throughput can be seen on both the LAN and WAN for all appliances. In
the diagram on the left, peak throughput for the WAN is 93 Mbps while the LAN is approximately 90. The
peak throughput is very close in this scenario as the data is not cached on the Silver Peak appliances.
In the diagram on the right, the peak WAN throughput is less than 10 Mbps while the LAN is near speeds
native to the gigabit Ethernet that was used for the test. This is most likely due to the fact that the Silver
Peak VX-7000 was caching the CIFS and HTTP traffic and was connected to the same VMware vSwitch as
the tenant desktops.
14 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
The following diagrams, the throughput over the WAN and LAN are shown for the duration of the test.
Figure 3 Silver Peak WAN optimization disabled
15 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
Figure 4 Silver Peak WAN optimization enabled
2.3 Test case 3 The following validation was done for 110 Enhanced users using the standard PAAC testing methodology
called out in the Reference Architecture document.
Test # 3 SP to DC Latency - 40 ms @ 1% loss CS to SP Latency - 120 ms @ 2% loss
This test was very similar to test case 1, but the latency and loss setting between the service provider and
customer sites were modified to determine the effect that the Silver Peak appliance had on PCoIP
presentation traffic.
The following table details the findings that have been averaged over all of the desktops:
SP Disabled SP Enabled % Improvement (round)
Desktop Login Delays 117.38 51.54 ~128%
2.4 Results summary
Test Case
CS to SP Latency
CS to SP Loss
SP to DC Latency
SP to DC Loss
Silver Peak Data Reduction
Max. Throughput (Not Optimized)
Max. Throughput (Optimized)
1 38 0 40 1 98% 153 Mbps 1120 Mbps
2 38 0 80 1 80% 93 Mbps 926 Mbps
3 120 2 40 1 78% 145 Mbps 899 Mbps
16 Dell Wyse Datacenter for VMware Horizon DaaS and Silver Peak | v.1.0
3 Conclusion The Silver Peak VX-7000 WANOP appliances greatly improves the quality of line and the resulting desktop
experience for customers who wish to utilize the Dell Wyse Datacenter for VMware Horizon DaaS solution
within a service provider scenario. The appliances were shown to reduce WAN traffic by up to 98% and
decrease user login delay by up to 56%.
About the authors
Nicholas Busick is a Senior Solutions Engineer with Dell Wyse Solutions Engineering building, testing,
validating and optimizing enterprise VDI stacks.
Gus Chavira is the Senior Principal Engineering Architect for VMware Horizon based solutions at Dell. Gus
has extensive experience and expertise on the VMware solutions software stacks as well as in Enterprise
virtualization, storage and enterprise data center design. Gus has worked in capacities of Sys Admin, DBA,
Network and Storage Admin, Virtualization Practice Architect, Enterprise and Solutions Architect. In
addition, Gus carries a B.S. in Computer Science
David Hulama is the Senior Technical Marketing Advisor for VMware Horizon View solutions at Dell. David
has a broad technical background in a variety of technical areas and expertise in enterprise-class
virtualization solutions.