with VMware EUC Principled Technologies recently released what is clearly a biased and inaccurate report comparing the cost and scalability of XenApp Horizon 5.2.
Because some people can read this report and not understand it contains inaccuracies and prejudices behind it, I think the answer is required. The community deserves to hear the truth.
It is also important to understand that this blog is not directed to VMware as a company, rather I wrote this as a response to ensure that you, perhaps a Citrix and / or VMware customer how VMware EUC sponsored reports spread misunderstanding and mislead the community. I consider VMware high technology company, and I have great respect for most of the technical and commercial people to EUC (@ Andreleibovici, @DStafford to name a couple).
This less-than-precise sponsored report also compares non-persistent VDI with RDSH. Here is my blog myth-busting some of these claims. In this post, I want to break down and analyze some of VMware sponsored report claims specifically around scalability XenApp 5.2 vs Horizon
Scalability claim :.
for the purposes of this blog, I will assume that the number of scalability in the report is accurate for Horizon. However, I do take umbrage figures submitted for XenApp and I would start by looking at the claim that XenApp 6.5 can scale to 151 users to lossy compression and 145 users with lossless compression vs 174 users with VMware VDI carries on hardware identical. The test was conducted using the material G8 HP with 256 GB RAM and Dell EqualLogic PS-6110X using RAID 10
The XenApp numbers represented in the report are significantly lower than Frank Anderson digits (Citrix architect solution) tested in the Citrix laboratory on a very similar hardware using local storage (not SAN) some time ago. Frank numbers indicate that you can get 203 users sessions with a medium workload Login VSI on Hyper-V2. Does VMware vSphere hinting that has scalability issues for these poor results? Or are they simply distort the results by using the poor voluntarily testing methodology? Note that Frank used a similar configuration VM (4vCPU and 16 GB of RAM), identical processors (Intel Xeon E5-2680 processor and 128 GB of RAM used, 50% less memory than it reports the configuration). The configuration of storage it uses featured HP Smart Array Controller PCIe 3.0 SAS 6 Gb s / eight and 10000 RPM SAS disks set up in RAID 0 + 1 volumes, but the scalability is limited by the processor not storage .
now, you should not take the floor on Citrix scalability tests, HP released an unsubsidized architecture and objective XenDesktop reference by HP (see page 23) on a similar material demonstrating that they can get the light workload 284 Login VSI XenApp users.
This is another point of evidence and another lab report by Citrix Solutions Lab on how to design a XenApp solution for mobilizing applications for 500 users. You can get around ~ 205 users using HP hardware G8 and Intel Xeon E5-2670 CPU, local SSDs and 50% less memory than paper claims to VMware. Again, the only difference is that these figures are on Hyper-V2 (which by the way is clear).
In fact, some customers may want to install XenApp on the bare metal, which again tested and proven Frank can get ~ 245 users on similar hardware.
The point I am making is that single server scalability on the same hardware for XenApp / RDS will exceed both XenDesktop VDI and VMware View VDI . You can take a look at other alliance partners architectures such as Dell DVS reference architecture for XenDesktop and IBM smart desktop cloud infrastructure for XenDesktop
There is no way the numbers for the load average job would be to 151 users with a lower density of VMware VDI, so it is a shame that there are still so many points of existing XenApp data such misleading document sponsored by VMware. There are several other misleading elements clearly in this report:
- When taking the equipment on the account, the VMware sponsored report easily skips about IOPS and expensive storage requirement for VDI users (disk array Dell Equal Logic PS6110X) to support 175 users. XenApp 6.5 can produce a large number of scale, even without a SAN or NAS
-. 145 users for XenApp is the max Login VSI in the report, but the entire report focuses on that number. VMware can you publish VSIMax for testing
- At 51 minutes of the CPU test View increases to 100% and flat lines. This does not occur with the XenApp test meaning user sessions would be much more sensitive to XenApp
- Check the user response time. For XenApp, all response times are below the 4000ms brand, whereas this is not true with View
-. There is no mention of the total RAM consumed by XenApp. Our experience and independent reports show that it would be expected to be 50% of what a session consume VDI
- VMware partner jumped report the total amount of storage required for VDI. The basic rule is that each linked clone takes about 15% of the size of the golden image. Assuming a 40GB or image, the clone storage calculation linked to 174 workstations pooled would be about ~ 1 Terra Byte
If you are really interested in learning about scalability XenApp then I recommend you read the following article CTXS
Bottom line: RDS VDI scalability and numbers are close, but there are still at least 15 to 20% gap where RDS offers the highest user density without special storage or IOPS requirement and 50% less RAM
in part 2 of my blog, I will talk about the comparison of XenApp costs to cost VDI VMware Horizon.
@vishalg
0 Komentar