Citrix vs. VMware, who comes out on top?
There is an ongoing battle for dominance in the virtualization field, and lines are being drawn. Whose side are you on? Myself, being a VMware evangelist, I obviously side with VMware. That being said, I don’t feel as if I am offering a biased opinion. I’ve been to Citrix’s Maryland headquarters in Bethesda, and they gave me a well thought out demo and really impressed me with some of their features. I was specifically going on a work assignment to evaluate their VDI solution offering, XenDesktop. While I was there, they also gave me a thorough education on their hypervisor offering, XenServer. Both products are very pleasing to the eye, with rich graphics, given their use of the HDX technology that delivers amazing crisp desktop images. We actually are in the process of deploying a 10 seat VDI test environment using their XenDesktop solution, and thus far there have been no problems. I don’t want to get into brand bashing, I’m not into that, even as much as I love VMware and the software they provide, I still respect Citrix and their forge into the virtualization market. I want to present this argument with the highest level of objectivity I can offer it. With that in mind, please read this post with an objective mind, and judge for yourself. Which solution is better? VMware may not be suitable for your needs, while Citrix fits, and vice versa. Let’s be objective and logical in our reasoning for choosing a solution, I assure you with that approach you will never make the wrong decision based on your goals.
Back in 2009, Citrix CTO Simon Crosby and VMware Technical Marketing Manager Scott Drummonds debated virtualization/hypervisor performance. This event was held at the Burton Group Catalyst Conference North America 2009 and was moderated by Chris Wolf. This is a year old but I think the arguments remain the same. Have a look at the video and see who you think comes away with the most compelling argument. Some of the main points of debate were focused on technical nuances that offer the greatest performance benefit/detriment to VMware and Xen hypervisors, effective methods for measuring hypervisor performance and the potential barriers to perfomance scalability.
What does each company bring to the table?:
VMware, the global leader in virtualization and cloud infrastructure, delivers customer-proven solutions that significantly reduce IT complexity and enable more flexible, agile service delivery. VMware accelerates an organization’s transition to cloud computing, while preserving existing IT investments and enabling more efficient, agile service delivery without compromising control. With more than 190,000 customers and 25,000 partners, VMware helps organizations of all sizes lower costs, preserve freedom of choice and energize business through IT while saving energy—financial, human and the Earth’s. VMware delivers virtualization and cloud infrastructure solutions that enable IT organizations to energize businesses of all sizes. With the industry leading virtualization platform – VMware vSphere™ – customers rely on VMware to reduce capital and operating expenses, improve agility, ensure business continuity, strengthen security and go green. With 2009 revenues of $2 billion, more than 190,000 customers and 25,000 partners, VMware is the leader in virtualization which consistently ranks as a top priority among CIOs. VMware is headquartered in Silicon Valley with offices throughout the world and can be found online at http://www.vmware.com (taken from vmware.com)
Citrix Systems was born from the idea of unlocking applications from datacenters, and employees from the office – changing the way IT and people work. Today, this is the promise of virtual computing, and Citrix is at the epicenter. In the world ahead, the better way to work is virtual – completely untethered from the office, devices, networks and datacenters. This virtual workstyle creates new possibilities for personal efficiency and business speed. The world’s smartest organizations are discovering that “going virtual” is how to stay ahead in a business world that’s changing faster than ever. Virtualization goes far beyond just consolidating servers. It’s also about powering the cloud, enabling online collaboration from anywhere, and taking desktop computing to an entirely new level. It’s about elevating IT from managing equipment and systems, to enabling them to deliver computing as an on-demand service. For our customers, it’s about unleashing the productivity and creativity of people, allowing them to seize business opportunity ahead of competitors. By applying our virtual computing vision to everything from meetings to desktops, networks and clouds, computing can be far simpler. Our belief in the power of simplicity becomes more market-relevant every day and is the basis for how Citrix will continue to drive long-term value for shareholders, employees, customers and partners.
Both VMware and Citrix present valid arguments as to why they should be chosen based completely on the information, products, and statements shown above. Let’s dig a little deeper though. With the recent release of VMware View 4.5, and the flood of entrants into the desktop virtualization field, I’ve chosen to focus on each company’s desktop virtualization solution, VMware View 4.0 and Citrix XenDesktop 4.0 (with all fairness to Citrix, I’ll be comparing View 4.0 to XenDesktop 4.0).
First up to bat, Citrix. In February of 2010, Citrix commissioned Miercom to evaluate the effectiveness of two similarly configured VDI solutions using XenDesktop 4 and Vmware View 4. Let’s first take a look at the configuration Miercom used to base their validation on:
- *2 HPDL360 servers w/ 2 Intel Xeon Quad Core processors
*Each server had 12GB of RAM & 3x72GB disk drives for installing XenDesktop 4 and VMware View 4 suite for testing
*XenDesktop 4 server loaded w/ XenServer v.184.108.40.206, Service Console 3.0, Dekstop Delivery Controller v. 4.0, Virtual Desktop Agent 4.0.4522, Virtual Desktop Client 220.127.116.11560
*VMware View 4 server loaded w/ ViewConnection Server 4.0.0-210399, ViewAgent 4.0.0-210939 and ViewClient w/ offline 4.0.0-210939
How was it done? This is an excerpt from the report on the research was conducted given the equipment listed above:
We connected two servers to the Apposite WAN emulator through a switch.We used an Apposite WAN emulator http://www.apposite-tech.com for bandwidth restriction and to introduce packet loss and latency to the network to simulate remote user connections and constricted WAN environments. We also used Login Consultants VSI script http://www.loginconsultants.com that helped automate
the launch of Microsoft applications including Excel, Word and Outlook in a consistent manner while measuring CPU and bandwidth utilization. We used the ClearSight Network Analyzer
http://www.clearsightnetworks.com in our test bed to analyze the efficiency of the application protocols used: ICA with HDX for Citrix and PCoIP for VMware. We utilized Video Clarity ClearView systems
http://www.videoclarity.com to determine VidMOS scores for quantitative analysis. Each server was equipped with a domain controller, DNS, connection server and broker as the part of the
virtualization platform. We measured bandwidth utilization and assessed the Quality of Experience (QoE) while playing Flash movies accessed through Internet Explorer 8 from websites including http://www.hulu.com and http://www.youtube.com, using the virtual desktop software installed on each client. We used an Ixia XM2 chassis http://www.ixiacom.com to apply IMIX background traffic while capturing the Flash
video content for analysis. Real-world traffic was also used in testing as generated by Ixia’s test platform and test applications, principally IxNetwork for Layer 2-3 routing and switching traffic and IxLoad for Layer 4-7 application traffic.
The results were actually pretty shocking to me. While accessing applications like MS Word, Excel and Powerpoint, Citrix XenDesktop used only .377Mbps while VMware View used 1.029Mbps. That is 64% less bandwidth usage by XenDesktop! That is pretty significant. When streaming flash video XenDesktop consumed 3Mbps compared to View’s consumption of 29Mbps. The conclusion that Miercom came to was that “Citrix XenDesktop 4 enables businesses to deploy VDI in larger scale and supports Flash video far better than VMware View 4. This report to me seemed heavily scored on bandwidth consumption alone. The conclusion that I would come to is that, yes, Xen consumes far less bandwidth…period. To me this is a very one sided effort by Citrix, to make their product seem superior to VMware’s. I can’t say that you can objectively get the whole picture based off of this report. Click here to read the full report with stats and diagrams.
Next up, VMware. VMware has not participated in one of these reports. Remember now, the Miercom report was commissioned and paid for by Citrix, so you can’t consider it completely unbiased in it’s research. We only know very “birds-eye-view” details in the configuration of the VMware components, if you read the report you can clearly see the difference in detail that Citrix goes into about the configuration of their components vs. VMware’s. That being said, there was a rebuttal to the Miercom report. Here are some excerpts from a blog post by Mike Coleman which appeared on the VMware View blog back in March 2010 (one month after the report surfaced):
Recently Miercom published a report that was commissioned and paid for by Citrix. VMware did not participate in these tests they, and we have no knowledge around how Miercom/Citrix set those tests up, how they configured VMware View 4, or how they arrived at some of their conclusions. All that being said, we would gladly welcome the opportunity to take part in unbiased, apples to apples comparison with Citrix as we’re confident that VMware View can stand on its own.
In a comparison of Virtual Desktop Infrastructure (VDI) implementations, Citrix XenDesktop 4 provided better overall performance when compared to VMware View 4
It appears that the Miercom report did not consider the image quality of office applications. XenDesktop 4 uses lossy client side image caching compared to VMware View 4’s lossless techniques, which favor the View 4 user experience. Many customers insist on lossless image delivery – image compression artifacts are totally unacceptable to them. The PCoIP protocol provides network efficient progressive image refinement that builds the desktop image to a lossless image. As a result, View is consistently chosen to deliver a high-quality desktop experience.
In our testing utilizing the Sysmark benchmark View 4 used 40% less bandwidth when run on a 1Mbps network compared to running the same test on a 100Mbs network. However, XenDesktop 4 used essentially the same bandwidth in both cases – pointing out the weakness of a fixed quality system – Citrix cannot take advantage of additional network resources when available. Miercom used 100Mbps as their ‘WAN’ for the single user test which skews the results and is not a real deployment scenario. To reiterate – the PCoIP protocol dynamically adapts the image quality based on the available network resources – thusly View 4 favors the best user experience when operating on an unconstrained network. So using a single user session to measure bandwidth performance is not a valid testing scenario. A more realistic scenario would be to compare View 4 to XenDesktop 4 on a constrained network supporting multiple users – this would result in significantly less bandwidth used by each View 4 session, while giving the best experience possible for the concurrent users.
Flash video was delivered with an average of 65% less CPU usage, 89% less bandwidth, and excellent Quality of Experience by XenDesktop 4 compared to View 4
Part of the strength of PCoIP is the fact that it adapts to the network. So, if you were to test a single desktop, as an example running multimedia, then you would observe that PCoIP consumed whatever bandwidth was available at the time. If bandwidth is available, it will use it to deliver the highest quality experience. If bandwidth is restricted, as is recommended in constrained environments, PCoIP “throttles down” to consume less bandwidth and is a “good citizen” in the network. Our customers report that PCoIP has been game changing for them and enables the delivery of a desktop environment that is secure, scalable, and easy to manage. With regards to the flash video test, this is not an apples to apples comparison – since the tests were conducted running Flash redirection with an unrealistically high packet loss of 0.5 to 5%. A correctly provisioned network will have packet loss of 0.1% or less which would result in great video performance.
Overall, XenDesktop 4 uses system resources more efficiently and is capable of scaling more effectively
Interesting that Miercom/Citrix does not call out the management costs of HDX Flash optimization/redirection. In addition to requiring a more powerful (and costly) client, the client media CODECs must be constantly updated as Adobe, Microsoft etc update their media protocols. Only VMware View 4 supports a true zero client on the desktop to provide the lowest operation cost, highest security and future-proof scalability (supports any media or graphics). In closing, we really do believe that VMware View is the only desktop virtualization product that has been designed from the ground up to deliver and manage virtualized desktop environments as a service. And, we do that with less cost and complexity than any other solution on the market.
A great argument made by VMware. No, VMware didn’t have numbers or diagrams or an outside consultant come in to validate their product, did they need to? I feel like the product speaks for itself. Again, maybe Citrix XenDesktop doesn’t work for your infrastructure, the same could apply to VMware View. In conclusion, it all comes back to making the decision that fits best with your infrastructure’s needs/wants/goals. Marketing, reports, or numbers shouldn’t sway you. Get an evaluation copy from both vendors and do a test run, this is the only true way to decide on who comes out on top, Citrix or VMware.