The Quest for an Independent Real-World Virtualization Benchmark

As we explained in our Xeon Nehalem review, comprehensive real-world server benchmarks that cover the market are beyond what one man can perform. Virtualization benchmarking needs much more manpower, and it is always good to understand the motivation of the group doing the testing. Large OEMs want to show off their latest server platforms, and virtualization software vendors want to show how efficient their hypervisor is. So why did we undertake this odyssey?

This virtualization benchmark was developed by an academic research group called the Sizing Server Lab. (I am also part of this research group.) Part of the lab is academic work; the other part is research that is immediately applied in the field, in this case software developers. The main motivation and objective of the applied research is to tell developers how their own software behaves and performs in a virtual environment. Therefore, the focus of our efforts was to develop a very flexible stress test that tells us how any real-world application behaves in a virtualized environment. A side effect of all this work is that we came up with a virtualization server benchmark, which we think will be very interesting for the readers of AnandTech.

Although the benchmark was a result of research by an academic lab, the most important objectives in designing our own virtualization benchmarks are that they be:

  • Repeatable
  • Relevant
  • Comparable
  • Heavy

Repeatable is the hardest one. Server benchmarks tend to run into all kinds of non-hardware related limits such as not enough connections, locking contention, and driver latency. This results in a benchmark that rarely runs at 100% CPU utilization and the CPU percentage load changes for different CPUs. In "Native OS" conditions, this is still quite okay; you can still get a decent idea of how two CPUs perform if one runs at 78% and the other runs at 83% CPU load. However, in virtualization this becomes a complete mess, especially when you have more virtual than physical CPUs. Some VMs will report significantly lower CPU load and others will report significantly higher CPU load when you are comparing two servers. As each VM is reporting different numbers (for example queries per second, transactions per second, and URL/s), average CPU load does not tell you the whole story either. To remedy this, we went through a careful selection of our applications and decided to keep only those benchmarks that allowed us to push the system close to 95-99% load. Note that this was only possible after a lot of tuning.

Comparable: our virtualization benchmark can run on Xen, Hyper-V and ESX.

Heavy: While VMmark and others go for the scenario of running many very light virtual machines with extremely small workloads, we go for a scenario with four or eight VMs. The objective is to find out how the CPUs handle "hard to consolidate" applications such as complex dynamic websites, OnLine Transaction Processing (OLTP), and OnLine Analytical Processing (OLAP) databases.

Most importantly: Relevant. We have been working towards benchmarks using applications that people run every day. In this article we had to make one compromise: as we are comparing the virtualization capabilities of different CPUs, we had to push CPU utilization close to 100%. Few virtualized servers will run close to 100% all the time, but it allows us to be sure that the CPU is the bottleneck. We are using real-world applications instead of benchmarks, but the other side of coin is that this virtualization benchmark is not easily reproducible by third parties. We cannot release the benchmark to third parties, as some of the software used is the intellectual property of other companies. However, we are prepared to fully disclose the details of how we perform the benchmarks to every interested and involved company.

The Virtualization Benchmarking Chaos vApus Mark I: the choices we made
POST A COMMENT

66 Comments

View All Comments

  • binaryguru - Monday, June 1, 2009 - link

    It seems to me, x86-based virutalization software is getting more and more complicated. Not only is x86 virtualization getting more complicated, it is getting more and more difficult to get reliable performance from it.

    Let me explain my point.

    The industry is clearly trying to do more with less hardware these days. Getting raw VM performance on commodity hardware is getting to a point where there is no predictable way to plan for an efficient VM environment.

    Current VM technology is trying to simulate the flexibility and performance of mainframes. To me, this is clearly an impossible goal to achieve with the current or future x86 platform model.

    All of the problems the industry is experiencing with VM consolidation does not exist on the mainframe. Running 4 'large' VMs for 'raw' performance. How about running 40 'large' VMs for 'raw' performance. Clearly, we all know that is impossible to achieve with current VM setups.

    Now I'm not saying that virtuallization is a bad idea, it clearly is the ONLY solution for the future of computing. However, I think that the industry is going about it the wrong way. Server farms are becoming increasingly more difficult to manage, never mind the challenge of getting 100s of blade servers to play nice with each other while providing good processing throughput.

    This problem has been solved about 20 years ago; and yet, here we are, struggling again with the "how can I get MORE from my technology investment" scenario.

    In conclusion, I think we need to go back to utilizing huge monolithic computing designs; not computing clusters.
    Reply
  • mikidutzaa2 - Friday, May 29, 2009 - link

    Hello,

    It would be useful (if possible) to have latency numbers/response times on the tests as well because rarely we are interested in throughput on our servers. What we usually care more is how long it takes the server to respond to user actions.

    What is your opinion?
    Reply
  • JohanAnandtech - Friday, May 29, 2009 - link

    I agree. I admit it is easier for us or any benchmark person to use throughput as immediately comparable (X is 10% faster than Y) and you have only one datapoint. That is why almost

    Responsetime however can only be understood by drawing curves relative to the current throughtput / User concurrency. So yes, we are taking this excellent suggestion into consideration. The trade off might that articles get harder to read :-).
    Reply
  • mikidutzaa2 - Friday, May 29, 2009 - link

    Looking forward to your new articles then, glad to hear :).

    The articles don't necessarily have to be harder to read, you could put the detailed graphs on a separate page and maybe show only one response time for a "decent"/medium user concurrency.

    Also, I would find interesting (if you have time) to have the same benchmarks with 2vcpu machines, I think this is a more common setup for virtualization. Very few people I think virtualize their most critical/highly used platforms - at least that's how we do it. We need virtualization for lightly used platforms (i.e. not very many users) but we are still very much interested in response time because the users perceive latency, not throughput.

    So the important question is: if you have a virtual server (as opposed to a physical one) will the users notice? If so, by how much is it slower?

    Thank you.
    Reply
  • RobAm - Tuesday, May 26, 2009 - link

    It's good to see some unbiased analysis with respect to virtualization. It's also especially interesting that your workloads (which look much more like real world apps my company runs as opposed to SPECjbb, vmark, vconsolidate) shows a much more competitive landscape than vmware and Intel portray. Also, doesn't vmware prohibit benchmarking without their permission. Did they give you permission? Has VMware called offering to re-educate you? :-) Reply
  • Brovane - Tuesday, May 26, 2009 - link

    I was hoping for a some benchmarks on the Xeon x7xxx CPU for the Quad Socket Intel boxes. We are currently have Dell R900's and we where looking at adding to our ESX cluster. We where debating between the R900 with Hex cores our Xeon x55xx series CPU's in the R710. I see the x55xx series where bench marked but nothing on the Xeon MP series unless I am missing that part of the article. Reply
  • JohanAnandtech - Tuesday, May 26, 2009 - link

    Expect a 24-core CPU comparison soon :-). Reply
  • Brovane - Tuesday, May 26, 2009 - link

    You also might want to a 12-core comparison also. We have found that with a 4-socket box that you usually run out of memory before you run out CPU power. With the R900 having 32-Dimm Sockets, the R900's we purchased last year have 64GB of RAM and just use 2x2.93Ghz CPU's we max memory before CPU easily in our environment. Since Vmware licensing and Data Center licensing is done per Socket we only populate 2 of the sockets with CPU's and this seems to do great for us. You basically double your licensing costs if you go with all 4 sockets occupied. Just a thought as to how sometimes virtualization is done in the real world. There is such a price premium for 8GB memory Dimm's it isn't worth it to put 256GB in one box with all 4 sockets occupied. The 4GB Dimm's did reach price parity this year so we were looking at going for 128GB of memory on our new R900's however Intel also released Hex-core so we still don't see much reason to occupy all 4 sockets. Reply
  • yasbane - Tuesday, May 26, 2009 - link

    I know positive feedback is always appreciated for the hard work put in but it seems very rare that we see any non-microsoft benchmarks for server stuff these days on Anandtech. Is there any particular reason for this...? I don't mean to carp but I recall the days when non-microsoft technologies actually got a mention on Anandtech. Sadly, we don't seem to see that anymore :(

    Cheers
    Reply
  • JohanAnandtech - Tuesday, May 26, 2009 - link

    Yasbane, my first server testing articles (DB2, MySQL) were all pure Linux benches. However, we have moved on to a new kind of realworld benchmarks and it takes a while to master the new benchmarks we have introduced. Running Calling Circle and Dell DVD store posed more problems on Linux than on Windows: we have lower performance, a few weird error messages and so on. In our lab, about 50% of the servers are running linux (and odd machines is running OS-X and another Solaris :-) and we definitely would love to see some serious linux benchmarking again. But it will take time.

    Xen benchmarks are happening as I write this BTW.
    Reply

Log in

Don't have an account? Sign up now