hpe simplivity hyperconverged infrastructure for vdi ... · for vdi specifically, hpe simplivity...

This ESG Lab Review was commissioned by HPE SimpliVity and is distributed under license from ESG.

© 2017 by The Enterprise Strategy Group, Inc. All Rights Reserved.

Abstract

This ESG Lab Review presents results from the analysis of Virtual Desktop Infrastructure (VDI) testing on an HPE HPE

SimpliVity hyperconverged infrastructure, with a goal of highlighting high levels of performance and operational efficiency.

Using Login VSI, testing included evaluating the process of rapidly provisioning up to 2,000 linked-clone VMs across two five-

node clusters and simulating a login storm and node failure to measure virtual desktop performance and VM density

potential.

The Challenges

The use of server virtualization to consolidate server infrastructure and improve operational efficiency has seen

phenomenal growth over the past decade and hyperconverged infrastructures have further accelerated virtualization

adoption by offering data-center-in-a-box simplicity for quick on-boarding. The success of server virtualization is

unparalleled and a large amount of its success can be attributed to the continued use of desktop virtualization. According to

ESG’s annual IT spending intentions survey, server virtualization has been among the top IT priorities reported by

respondents for the last four years. In fact, one in four respondents listed the increased use of server virtualization as one of

their most important IT priorities over the last year. Further, desktop virtualization also landed in the top five most-cited IT

priorities with one in four specifying it specifically (see Figure 1).1

Figure 1. Top Ten Most Important IT Priorities for 2015

Source: Enterprise Strategy Group, 2015

With server virtualization and desktop virtualization being top priorities, it is not surprising that organizations expect a

large amount of their IT budgets to be spent on these initiatives. When asked to indicate what IT initiatives would have a

significant impact on their organizations’ storage spending over the next 12-18 months, 24% cited supporting server

virtualization implementations, while 23% specified desktop virtualization/thin client initiatives like VDI. 2

1 Source: ESG Research Report, 2015 IT Spending Intentions Survey, February 2015. 2 Source: ESG Research Report, 2015 Data Storage Market Trends, October 2015.

22%

22%

23%

24%

25%

25%

25%

26%

26%

34%

0% 5% 10% 15% 20% 25% 30% 35% 40%

Improving collaboration capabilities

Building a “private cloud” infrastructure

Business continuity/disaster recovery programs

Regulatory compliance initiatives

Using cloud infrastructure services

Desktop virtualization

Increasing use of server virtualization

Managing data growth

Improving data backup and recovery

Information security initiatives

Top 10 most important IT priorities over the next 12 months. (Percent of respondents, N=601, ten responses accepted)

ESG Technical Review

HPE SimpliVity Hyperconverged Infrastructure for VDI Environments Date: June 2017 Author: Mike Leone, Senior Validation Analyst

Enterprise Strategy Group | Getting to the bigger truth.™

http://www.esg-global.com/research-reports/2015-it-spending-intentions-survey/

http://www.esg-global.com/research-reports/2015-data-storage-market-trends/

Lab Review: HPE SimpliVity Hyperconverged Infrastructure for VDI Environments 2


The Solution: HPE SimpliVity Hyperconverged Infrastructure for VDI Environments

HPE SimpliVity architecture is a proven hyperconverged infrastructure solution purpose-built from the ground up to meet

the demanding needs of virtualized environments. Packaged in easy-to-deploy 2U building blocks, HPE SimpliVity

infrastructure has combined enterprise functionality and capabilities with the operational and cost efficiency of the cloud to

deliver improved VM density, performance, and scalability.

HPE SimpliVity hyperconverged infrastructure includes all IT services below the hypervisor, including compute, storage (HDD

and SDD tiers), deduplication, network switching, backup, replication, WAN optimization, cloud gateway, and numerous

other functions that eliminate the need to buy/deploy disparate products. As HPE SimpliVity hyperconverged infrastructure

nodes are added to an IT infrastructure, they become a multi-node cluster or—as HPE calls them—a Federation. A

Federation is a self-healing and self-learning global “cluster of clusters” that can be managed from a single VM-centric user

interface. The Federation offers all of the functional and operational advantages of a HPE SimpliVity cluster at a global scale.

The HPE SimpliVity Data Virtualization Platform spans the entire Federation, allowing all VM movement (including backups)

between sites/clusters to leverage the HPE SimpliVity global deduplication/compression/ optimization functionality.

As shown in Figure 2, HPE SimpliVity software incorporates three unique core innovations:

• Accelerated Data Efficiency: HPE SimpliVity Data Virtualization Platform, The underlying data architecture, performs global deduplication, compression, and optimization of all data at inception, inline, with no impact to performance. This is done through the HPE OmniStack Accelerator Card.

• Global Unified Management: An intelligent network of collaborative systems that provides massive scale-out capabilities as well as VM-centric policies that are abstracted from the infrastructure and managed through a single unified interface for the entire global infrastructure.

• Built-in Data Protection: Automated backup/restore, clone, and replication capabilities are built in and leverage always-on data efficiency, allowing for granular RTOs/RPOs with minimal overhead.

Figure 2. HPE SimpliVity hyperconverged Infrastructure for VDI Environments


For VDI specifically, HPE SimpliVity has made significant strides in optimizing the provisioning, deployment, and protection

of dynamic VDI environments. With HPE SimpliVity technology, traditional VDI pain points related to deployment,

management, performance, protection, and cost are greatly reduced. Industry-proven HPE SimpliVity technology provides

an enhanced end-user experience in VDI environments by delivering up to 250 low-latency, highly-available virtual desktops

on a single node. And it drastically simplifies the customer transition from pilot deployments to working production VDI

environments by offering an easy-to-deploy, easy-to-scale VDI solution.



Provisioning and Logon Performance Predictability

ESG Lab audited the performance of a ten node HPE SimpliVity Federation that simulated a growing VDI deployment for a

global organization. First, a local VDI deployment of 1,000 users was provisioned and tested on a five-node cluster. Then an

additional 1,000 virtual desktops were provisioned on a separate five-node cluster in a different location, increasing the

total number of VDI users to 2,000, all of which could be centrally managed. All performance characteristics leveraged

VMware Horizon View executing in the HPE SimpliVity environment with key areas of focus put on provisioning times, login

performance, and sustainability of performance through node failures.

The industry-standard VDI benchmarking tool Login VSI was used for testing. LoginVSI validates application performance

and response times for various predefined VDI workloads with an ultimate goal of showing desktop density potential for a

given set of hardware and software components. Final results are presented using easily understandable Login VSI metrics

that represent the number of concurrent sessions running when the VDI environment reaches a saturation point.

Provisioning Times

The speed at which a large VDI environment can be deployed can have a significant impact not only on IT staff productivity

and availability, but the whole workforce. Whether deploying for the first time or applying a new gold image with updates,

the goals are the same—get it done quickly and correctly to minimize downtime and improve workforce productivity.

Historically, provisioning desktops is a time-consuming task that can place a heavy burden on storage resources. Starting

with the local five-node cluster, ESG Lab measured the time it took to provision 1,000 office worker linked-clone virtual

desktops with VMware Horizon View 6. The configuration of the virtual desktop consisted of 1 vCPU and 1.5 GB of vRAM

running a 64-bit version of Windows 7 on a 25GB virtual disk. As shown in Figure 3, provisioning and deploying the 1,000

virtual desktop environment to a login-ready state with VMware Horizon View 6 was impressively quick, completing in just

70 minutes.

Figure 3. Deploying 1,000 Linked Clone Virtual Desktops with HPE SimpliVity Architecture


0

200

400

600

800

1,000

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70

Nu

mb

er o

f D

eskt

op

s D

eplo

yed

Elapsed Time (minutes)

Deploying 1,000 Linked Clones with HPE SimpliVity (Office Worker Desktops and VMware Horizon View 6)

Deploying 1,000 Linked Clonesin 70 minutes



Logon Storm

A logon storm represents a large number of end-users launching their virtual desktops simultaneously during a common

time period such as the beginning of the workday. This can often lead to high response times and degradation of services

due to the overwhelming number of data requests that need to be handled by the infrastructure.

A logon storm was simulated with Login VSI with a goal of understanding how the five-node cluster could handle the 1,000

virtual desktops. Login VSI reported two metrics, the VSI Index Average and the VSImax v4 total. The VSI Index Average

serves as an average response time measurement, but leverages built-in rules to help factor in spikes while not offsetting

the overall average. VSImax v4 shows the amount of concurrent sessions that can be active before the infrastructure is

saturated. The VSImax v4 calculation is primarily used to provide insight into the potential scalability that an environment

can sustain while meeting acceptable performance requirements.

The results as provided by Login VSI are shown in Figure 4. A VSIbase of 760 ms was determined during the baseline phase,

which takes 30 VSI samples, eliminates outliers, and averages the remaining times. The VSImax Average is represented by

the slightly upward trending blue line—as more sessions are started, the VSImax Average gradually increases. The VSImax

v4 score is the point where the VSImax Average crosses the VSI Threshold (straight red line). This signifies the point at which

performance becomes unacceptable. In this particular test case, the VSImax Average did not cross the VSI Threshold,

meaning that the five-node cluster could easily support 1,000 virtual desktops. It also meant that no VSImax v4 score would

be reported because the threshold was never hit.3 The last VSImax average measurement as the 1,000th virtual desktop

logged on was 1,029 ms, while the VSI Threshold was 1,761ms.

Figure 4. Logon Storm Results of 1,000 Users


The next phase of testing focused on showing performance predictability by simulating the growth of an organization that

requires a second set of virtual desktops for a branch office. A second cluster was identically configured and deployed

within the same Federation—five nodes with 200 virtual desktops per node. Unlike a typical VDI deployment, even though

half of the desktops were located in a remote location, with the HPE SimpliVity Federation, data could still be shared

between clusters if necessary. Additionally, administration and management of both clusters within the Federation could be

centralized.

3 VSImax calculation documentation

http://www.loginvsi.com/documentation/Login_VSI_Analyzing_Results



A second logon storm was simulated across the entire Federation, which now consisted of 2,000 total simulated users. As

expected, results were similar to that of the 1,000 logon storm result (shown in Figure 5). Similar to the 1,000-session test,

the 2,000-session test yielded a VSIbase of 757ms, a VSImax was never hit, and the VSImax Average at the end of the test

was 1038ms.

Figure 5. Logon Storm Results of 2,000 Users


Why This Matters

Performance is a top concern when it comes to virtual desktop infrastructures and ESG research has shown that IT managers view performance as a top challenge when it comes to implementing desktop virtualization. Predictable performance scalability is a critical concern when multiple users are running diverse applications on a shared VDI infrastructure. A burst of I/O activity from one desktop (e.g., a user logging on) can lead to poor response times and lost productivity for other users. To get the most of their VDI investments, IT managers are looking for a scalable VDI solution that’s easy to deploy and manage as it cost-effectively scales to the meet the performance needs of thousands of VDI users without impacting the end-user experience.

ESG Lab validated that HPE SimpliVity technology easily supported the demanding requirements of a VDI environment by delivering predictable performance for a 2,000 virtual desktop deployment across two separate clusters managed as a single Federation. HPE SimpliVity Data Virtualization Platform allowed the five-node cluster to easily handle the impact of the I/O bursts commonly associated with VDI logon storms, while provisioning times and logon storms both yielded impressive results. Deploying all the linked-clone virtual desktops took just 70 minutes, while the VSImax as reported by Login VSI was never reached, proving the cluster could comfortably support even more than the tested 200 users per node.



Resiliency and Performance Sustainability

HPE SimpliVity hyperconverged infrastructure is designed with enterprise reliability in mind. Deployment on highly available

servers provides n+1 fault tolerance for key hardware components along with RAID-protected data storage. HPE SimpliVity

hyperconverged infrastructure leverages all of the HA features built into VMware, including multipathing and VMware HA to

provide access to VMs even in the event of a node or guest OS failure.

For the next phase of testing, ESG

Lab analyzed VDI performance

results within the same Federation

while injecting simulated node

failures. As shown in Figure 6, the

original Federation consisted of two

five-node clusters, with each node

supporting 200 VDI users, totaling

2,000 users across the Federation.

For resiliency testing, one node

from each cluster was intentionally

failed and the same total number of

users was then spread across the

remaining nodes. This changed the

total number of supported VDI

users to 250 per node, a 50-user

increase per node.

Using Login VSI,a logon storm simulation was performed both before and after simulated node failures. The results of the

comparison are shown in Figure 7 and Table 1. Once again, a VSImax was not reached, meaning the additional 50 users per

node was easily handled by each cluster. A slight increase of 30 ms to the VSIbase was witnessed, while the VSImax Average

increased by 15% up to 1,195 ms, still well below the VSI Threshold of 1,787 ms.

Figure 7. Test Bed Before and After Simulated Node Failures


Figure 6. Test Bed Before and After Simulated Node Failures




Table 1. Comparing Logon Results for 2,000 Users Before and After Simulated Node Failures

Federation Status # of

Clusters # of Nodes per Cluster

# of Users per Node

VSIbase VSImax Average

VSImax Threshold

Fully Operational 2 5 200 757 1,038 1,757

Two Failed Nodes 2 4 250 787 1,195 1,787


Why This Matters

Along with increasing the use of server virtualization and desktop virtualization, business continuity and disaster recovery also appeared on the top ten most-cited IT priorities for organizations in 2015.4 This is particularly important in VDI environments, where end-user experience dictates acceptable levels of performance and sustainable VDI density. Adding just a few seconds to virtual desktop boot-up times can significantly impact the end-user experience for the worse.

ESG Lab confirmed that a HPE SimpliVity hyperconverged infrastructure for VDI environments sustained high levels of VDI performance, even through simulated node failures. Two clusters with five nodes each originally supported 200 users per node. Simulated node failures on each cluster forced each node to take on an additional 50 users each. Response time impact was minimal, proving that the increased level of density could easily be supported.

4 Source: ESG Research Report, 2015 IT Spending Intentions Survey, February 2015.

http://www.esg-global.com/research-reports/2015-it-spending-intentions-survey/


© 2017 by The Enterprise Strategy Group, Inc. All Rights Reserved. www.esg-global.com [email protected] P.508.482.0188


The Bigger Truth

While the benefits of server virtualization are widely known, the benefits of desktop virtualization are proving to have an

equal or greater impact in today’s business environment. Unfortunately, the increased number of client device types, the

mobilization of the workforce, and a continued desire to increase operational efficiency all combine to make current

desktop virtualization and management a daunting task. IT is expected to provide a high-performing, predictable, and

productive computing environment, while controlling operational costs and hardware expenses without impacting the end-

user experience.

HPE SimpliVity data architecture enables organizations to cost-effectively simplify the deployment and management of

virtualized environments. HPE SimpliVity core data efficiency technology, global unified management, and built-in data

protection capabilities help to facilitate a VDI end-user experience that begins to rival that of a physical desktop. Extending

these technologies and benefits to a purpose-built solution for VDI environments and workloads, organizations not only gain

improved VM density, operational efficiency, and predictably high levels of performance, but they also gain confidence

knowing that the end-user experience and infrastructure resilience should no longer be a concern.

ESG Lab witnessed high levels of performance on a 1,000 VDI user cluster, which eventually grew to a geographically

dispersed, two-cluster, 2,000-user VDI environment. Provisioning times of all the virtual desktops took just 70 minutes,

while 200 users per node yielded more than acceptable times during a logon storm. Simulated node failures in each cluster

had no impact on performance and more importantly, no impact on the overall end-user experience.

Despite the promising benefits, the complexities involved in traditional VDI deployments are no mystery: increased time and

cost, architectural and interoperability difficulties, and most importantly, the need to pass the end-user experience test. By

leveraging a proven hyperconverged solution from an HPE SimpliVity solution purpose-built for VDI, many of the

complexities can be quickly rendered an afterthought. With centralized management across globally dispersed

organizations, cost-effective building blocks to handle organizational growth, and the ability to deliver high levels of

required end-user performance, the HPE SimpliVity hyperconverged infrastructure for VDI environments delivers a cost-

effective, highly efficient solution for any sized organization.

All trademark names are property of their respective companies. Information contained in this publication has been obtained by sources The Enterprise Strategy Group (ESG) considers to be

reliable but is not warranted by ESG. This publication may contain opinions of ESG, which are subject to change. This publication is copyrighted by The Enterprise Strategy Group, Inc. Any

reproduction or redistribution of this publication, in whole or in part, whether in hard-copy format, electronically, or otherwise to persons not authorized to receive it, without the express consent

of The Enterprise Strategy Group, Inc., is in violation of U.S. copyright law and will be subject to an action for civil damages and, if applicable, criminal prosecution. Should you have any questions,

please contact ESG Client Relations at 508.482.0188.

The goal of ESG Lab reports is to educate IT professionals about data center technology products for companies of all types and sizes. ESG Lab reports are not meant to replace the evaluation

process that should be conducted before making purchasing decisions, but rather to provide insight into these emerging technologies. Our objective is to go over some of the more valuable

feature/functions of products, show how they can be used to solve real customer problems and identify any areas needing improvement. ESG Lab's expert third-party perspective is based on our

own hands-on testing as well as on interviews with customers who use these products in production environments.

hpe simplivity hyperconverged infrastructure for vdi ... · for vdi specifically, hpe simplivity...

Documents