Dell Desktop Virtualization Solutions Stack With Teradici .

1y ago
671.05 KB
9 Pages
Last View : 1y ago
Last Download : 4m ago
Upload by : Rafael Ruffin

Dell Desktop VirtualizationSolutions Stack withTeradici APEX 2800 serveroffload card Performance ValidationA joint Teradici / Dell white paper

Contents1.Executive overview.22.Introduction.32.1The APEX 2800 server offload card.32.2Dell positioning in the VDI eco-system.33.Validation testing as a commitment to quality.43.1Test methodology.43.2Test load scenarios.43.3Test measures around three elements.54.Test results.65.Benefits demonstrated by testing.76.Summary.77.Appendix.81. Executive overviewFollowing the successful launch of the Teradici APEX 2800 server offload cardin February 2012, Dell decided to test the card on their Desktop VirtualizationSolutions (DVS) Stack to better understand its value proposition and possiblebenefits it could bring to the solution.The testing completed by the Dell DVS Engineering team was able to confirmthe effectiveness of the card in offloading, from the server CPU, the PCoIPencoding tasks of the displays of the 64 most active virtual machines, reducingserver CPU utilization by up to 40% (video workload).This result confirms Teradici’s positioning of this card in providing an extralevel of insurance to maintain performance expectations in VMware Viewenvironments that support end users with increased workload complexity.Teradici APEX 2800 Server Offload Card – Dell DVS Stack validationpage 2

2. Introduction2.1Teradici APEX 2800 server offload cardThe Teradici APEX 2800 card offloads the tasks related to the remote displayprotocol (PCoIP) from the server CPU, reducing server CPU utilization by up to50%. However, as opposed to a GPU offload card, it does not directly improvethe end-user experience but rather ensures CPU cycles are available for the“normal” tasks to be carried out by the CPU. This results in two benefits: The card protects and ensures a consistent end user experience asworkloads change, by monitoring the activity of all the displays onthe server and dynamically offloading the most active 64, seamlesslyswitching between soft encoding and APEX 2800 hardware encoding. The card increases the server density, by allowing IT managers to putmore Virtual Machines (VM) on the server. Depending on the workloadconsidered, this benefit ranges from 10% to 100% increase.The first point is particularly important as VDI begins to be deployed for moredemanding users with less predictable workloads, using more multimediaapplications in terms of pixel changes.For more information, please refer to the Teradici whitepaper “TeradiciAPEX 2800: Addressing the Next VDI Bottleneck” available for download atwww.teradici.com2.2Dell positioning in the VDI eco-systemDell has defined Desktop Virtualization as a pillar of its strategy moving forward.Its recent strategic acquisitions in this market emphasize how much Dellbelieves in this industry.Dell’s DVS Enterprise offering is a comprehensive solution portfolio designedto enable customers to experience the benefits of virtual end user computing.The solution is based on VMware View 5.1 and VMware vSphere 5 ESXi with DellPowerEdge rack or blade servers, EqualLogic SAN for iSCSI or Compellentfor FC, PowerConnect or Force10 network switches and Wyse thin clients.This allows Dell to be the only vendor offering a true end-to-end solution fordesktop virtualization.Based on extensive engineering, architectural design and scalabilitycharacterization work, the Dell Desktop Virtualization Solutions can bequickly and confidently deployed into production environments, therebyhelping to eliminate much of the costly and time consuming trial anderror often encountered during complex custom infrastructure design andimplementations. This solution provides scalability and manageability for upto 50,000 users.Dell Wyse zero and thin client devices and software provide superior security,reliability and energy efficiency when compared to a traditional PC. Wysedesktop devices and software help streamline the delivery of VMwareinfrastructure to millions of users around the world. Zero clients create a moresecure environment that minimizes or eliminates exposure to data loss, virusesand malware. By utilizing zero or thin clients as the access device for end users,deployments can benefit from centralized management and complete control ofall endpoints. Since zero clients eliminate components with high failure rates,deployments can expect reduced costs and improved reliability over the life ofa desktop virtualization deployment.Teradici APEX 2800 Server Offload Card – Dell DVS Stack validationpage 3

While the Teradici APEX 2800 has already been added to the Dell online e-storein USA and Europe, Dell wanted to better understand the value propositionof the card and decided to run the tests on its DVS Enterprise stack to gainvisibility of the following: When does the card provide the most value to customers? What is the benefit of the card under the Dell office worker workload? Considering different workloads, does the card improve user experience?For further information on Dell’s Desktop Virtualization Solutions go -client3. Validation testing as a commitment to qualityBoth Teradici and Dell are committed to providing quality technological advancesto their customers. As part of this commitment to quality, the companiesconducted validation testing with the APEX 2800 on the Dell DVS stack.3.1Test methodologyIn the test environment Login VSI (developed by Login Consultants) wasused to generate the desktop workloads. Login VSI is a widely-used tool forvalidation of VDI environments and server-based computing / terminal servicesenvironments. It installs a standard collection of desktop application software(e.g. Microsoft Office, Adobe Acrobat Reader etc.) on each VDI desktop. It thenuses launcher systems to connect a specified number of users to availabledesktops within the environment.The tests were monitored using esxtop on vSphere which samples CPUutilization every 5 seconds, instead of using vCenter that samples only every20 seconds and could therefore miss some CPU peaks and related activities.The same sets of tests were run on an identically configured PowerEdge serverwithout the Teradici APEX 2800 server offload card.3.2Test load scenariosTesting was performed using various workloads that simulated the varied CPUusage environments associated with businesses. These different test casesprovide evidence of the overall cost and time saving possibilities of the APEX2800. The workloads, generated through Login VSI, are described below: Premium Workload ( Login VSI heavy workload) Considered to be a conservative “office worker” workload withoffice productivity applications and 30-second video into a 10minute loop Users with a desktop resolution of 1600 x 1200 The workload continued for 30 minutes after all sessions werelogged on to exclude the affect of the login boot storm whichthe APEX 2800 does not impact Continuous Video Playback ( Login VSI custom workload) Considered to be a true multimedia workload, or a “peakworkload”, where the APEX 2800 is expected to show greatbenefits Continuous video playback at 480pTeradici APEX 2800 Server Offload Card user experience, when the server CPU, or vCPU, starts to exceed recommendedthresholds (over 75% for server CPU, over 80% for vCPU). By freeing up valuableCPU cycles, the card was able to indirectly improve the end user experience,enabling smoother video and crisp audio playback, for example.Teradici APEX 2800 Server Offload Card – Dell DVS Stack validationpage 6

5. Benefits demonstrated by the testingThe testing was able to answer the elements put forward prior to the testing:Is there a reduction in CPU usage? Yes, in all cases. We were able to demonstrate a reduction in CPUutilization, which ranged from 6% to 40%. However, 6% reductionof an already low server CPU utilization number can be quite small,especially when the server is not at maximum capacity. The highestreductions in host CPU usage were achieved with workloads thathad high amounts of video and graphically intensive applications.Does the APEX 2800 smooth CPU spikes? Yes, for two reasons: 1) Hardware encoding is more stable than softencoding, by design, and 2) CPU spikes due to pixel changes on thedisplays are reduced by 50% instantaneously, without any lag. Thisvalidates the “insurance” value proposition of the card: “protect andinsure a consistent user experience at all times as load changes”.Are there differences with higher rates of pixel changes? Yes, the PCoIP remote display protocol, used in VMware View, is smartenough to send over the network only the pixels that have changed onthe display; little activity on the display means a small amount of PCoIPtasks and therefore low impact on CPU utilization. On the other hand,with a lot of pixel changes, a lot of PCoIP tasks are required to updatethe display at the end point, hence creating higher load on the CPU, sothat the offload onto the APEX 2800 card will result in an important CPUutilization reduction.6. SummaryDell was able to confirm the value of the card as an “insurance” policy in VMwareView deployments and a means to de-risk VDI roll-out as VDI moves beyondtask workers, with predictable and low multimedia activities in terms of pixelchanges, to office workers and power users who represent a more demandingset of end users, with unpredictable workloads and more multimedia activities.By monitoring the activities of all the displays on the server, and dynamicallyoffloading the most active 64, the Teradici APEX 2800 is able to reduce CPUutilization peaks by up to 40%, greatly reducing the impact of load variation onuser experience. This makes the APEX 2800 an ideal choice in virtual desktopenvironments that need to support heavy workloads.In market verticals with some multimedia display activities in the averageworkload, the benefits of the card are readily apparent and can be expressed as: Increased server density and reduced cost per VM. Improved user experience should some of these multimedia activitiesrequire more CPU cycles.Teradici APEX 2800 Server Offload Card – Dell DVS Stack validationpage 7

7. AppendixSpecifications of the server used for the tests were: 2x Intel Xeon X5660 processor 96 GB memory 8 x 146 GB SAS 6 Gbps 15k disks – RAID 10 PERC 6i integrated RAID controller Embedded Broadcom 5709 GbE LOM with TCP/IP Engine (4 port)Teradici APEX 2800 Server Offload Card – Dell DVS Stack validationpage 8

Dell Inc. (NASDAQ: DELL) listens tocustomers and delivers worldwideinnovative technology, business solutionsand services they trust and value. For moreinformation, visit drives innovation to fundamentallychange the way people use and deploycomputers by developing technology andsolutions that deliver a true, uncompromisedPC user experience over IP networks. Ourfocused approach in designing advancedimage processing algorithms enables thephysical separation of the computer andthe user, and ultimately will change the wayenterprises compute.Teradici Corporation 1.604.628.1200 2004 – 2012 Teradici Corporation. All rights reserved. Teradici, PCoIP, PC-over-IP and APEX are trademarks of Teradici Corporationand may be registered in the United States and/or other countries. All other trademarks are property of their respective owners.Specifications subject to change without notice.

TERADICI APEX 2800 SERVER oFFLoAD CARD – DELL DVS STAC VALIDA TIon page_4 While the Teradici APEX 2800 has already been added to the Dell online e-store in USA and Europe, Dell wanted to better understand the value proposition of the card and decided to run the tests on its DVS Enterprise stack to gain visibility of the following:

Related Documents:

Desktop virtualization is driven by a combination of cost savings, increased ability to comply with myriad regulations, and an improvement in data and application security. The two fundamental forms of desktop virtualization are: Server-side application/desktop virtualization Client-side application/desktop virtualization !

desktop virtualization called Dell Virtual Remote Desktop (VRD). This solution combines Dell's client, server, storage, networking hardware, and services with Citrix XenDesktop desktop virtualization technology and Microsoft's server virtualization and management infrastructure.

The Dell Desktop Virtualization Solutions (DVS) Simplified VDI appliance is designed to remove these adoption barriers. DVS Simplified makes VDI cost-effective and easy to deploy for smaller, simpler implementations. DVS Simplified opens up the benefits of desktop virtualization to deployments with as few as 25 users.

In this section, we give an overview of virtualization and describe virtio, the virtualization standard for I/O devices. In addition, we discuss the state-of-the-art for network I/O virtualization. 2.1 Overview of Virtualization and virtio The virtualization technology is generally classi ed into full-virtualization and paravirtualization.

desktop virtualization can be a powerful incentive. 9. Scales up as needed Many desktop virtualization initiatives start out as pilot projects. It gives the organization - IT, management and users - an opportunity to experience the benefits of desktop virtualization, fine-tune their infrastructure, and report back to

Before you can use a stack you have to initialize the SP to point to one value higher than the highest memory location in the stack. For the HC12 use a block of memory from about 3B00 to 3BFF for the stack. For this region of memory, initialize the stack pointer to 3C00. Use LDS (Load Stack Pointer) to initialize the stack pointer.

TOSYM.123 Dell 1320C-593-10258 HY BK DELL 1320C/2130CN/2135CN 2.000 9,50 TOSYM.124 Dell 1320C-593-10259 HY C DELL 1320C/2130CN/2135CN 2.000 9,80 TOSYM.125 Dell 1320C-593-10260 HY Y DELL 1320C/2130CN/2135CN 2.000 9,80 TOSYM.126 Dell 1320C-593-10261 HY M DELL 1320C/2130CN/2135CN 2.000 9,80 TOSYM.251 Dell 593-10493 BK Dell 1230C Color Laser/1235CN Multifunction Colour Laser 1.500 .

Agile Software Development with Scrum An Iterative, Empirical and Incremental Framework for Completing Complex Projects (Slides by Prof. Dr. Matthias Hölzl, based on material from Dr. Philip Mayer with input from Dr. Andreas Schroeder and Dr. Annabelle Klarl) CHAOS Report 2009 Completion of projects: 32% success 44% challenged 24% impaired Some of the reasons for failure: Incomplete .