VMTN Blog

New performance results are out that show ESX supports more XenApp users with lower CPU usage than XenServer. Click through for the whole thing, including the details of the user-centric workload. For perormance testing of desktop or terminal services virtualization, simulating the workload is very important. It can be tricky, and if you don't simulate how real humans use their desktops, you may not come to conclusions that can apply to real-world deployments.

VMware: VROOM!: Virtualizing XenApp on XenServer 5.0 and ESX 3.5.

There has always been interest in running Citrix XenApp (formerly Citrix Presentation Server) workloads on the VMware Virtual Infrastructure platform. With the advent of multi-core systems, purchasing decisions are driven towards systems with 4-16 cores. However, using this hardware effectively is difficult due to limited scaling of the XenApp application environment. In addition to the usual benefits of virtualization, these scaling issues make running XenApp environments on ESX even more compelling.

We recently ran some performance tests to understand what can be expected in terms of performance for a virtualized XenApp workload. The results show that ESX runs common desktop applications on XenApp with reasonable overhead compared to a native installation, and with significantly better performance than XenServer.  ...

XenApp and other products that virtualize applications are prime candidates to be run in a VM. These results show that ESX can do so efficiently compared to using a physical machine. This was shown with a benchmark that: represents a real desktop workload, uses a metric that includes latencies of all operations, and requires that all operations complete successfully. Furthermore, ESX supports about 13% more users than XenServer at a given latency while using less CPU.

Last week we repeated our topic, Virtualizing Exchange and other enterprise messaging applications. Our guests, VMware's Dave Korsunsky and Jon Dybik, returned. A good time was had by all. As always, listen via the widget on the right, by downloading the mp3 directly, or by going through iTunes. (Every week, you can download the raw audio recording immediately, and then later I put it through Levelator, which makes it much more listenable and gives a constant volume from all the callers. So if you're blowing out your eardrums on your iPod with our surprise volume changes, try downloading again.)

Join us today in 3 hours or so (noon PST / 3pm EST / 8pm GMT / 7am Sydney EDT) for an Open Topic Roundtable. Come by, join the chat, or call in. More info here. Some topics we may cover:

• VI Toolkit (for Windows) aka PowerShell Toolkit version 1.5 released
• Multi-vendor post on high-throughput iSCSI
• From the email inbox: How many ESX Servers do you have to install before automation is really worth it? (Prediction: the answer will be 1.)

We focused on Microsoft Exchange last week, so other than our main page for things Exchangey:

• Deploying Exchange on a Dynamic Platform aka Virtualizing Microsoft Exchange.

all I've got for you today is our brand spankin' new section on the VMware Communities. Drop by and ask your questions there! Dave, Jon, and our other enterprise application experts are standing by.

• Virtualizing Exchange,  Domino, and RIM Community

Chad (of EMC) and Vaughn (of NetApp) posted today a great collaborative blog article (with others from VMware, Dell/EqualLogic and  HP/Lefthand) that has a nice backgrounder on iSCSI, talks about some design considerations, links to lots of resources, and then talks about some little-known configuration  and performance considerations. If you are not an iSCSI guru, you should read this post:

Virtual Geek: A Multivendor Post to help our mutual iSCSI customers using VMware.

Today’s post is one you don’t often find in the blogosphere, see today’s post is a collaborative effort initiated by me, Chad Sakac (EMC), which includes contributions from Andy Banta (VMware), Vaughn Stewart (NetApp), Eric Schott (Dell/EqualLogic), and Adam Carter (HP/Lefthand), David Black (EMC) and various other folks at each of the companies.

Together, our companies make up the large majority of the iSCSI market, all make great iSCSI targets, and we (as individuals and companies) all want our customers to have iSCSI success.

I have to say, I see this one often - customer struggling to get high throughput out of iSCSI targets on ESX.   Sometimes they are OK with that, but often I hear this comment: "…My internal SAS controller can drive 4-5x the throughput of an iSCSI LUN…"   

Can you get high throughput with iSCSI with GbE on ESX?   The answer is YES.  But there are some complications, and some configuration steps that are not immediately apparent. You need to understanding some iSCSI fundamentals, some Link Aggregation fundamentals, and know some ESX internals – none of which are immediately obvious…

If you’re interested (and who wouldn’t be interested with a great topic and a bizzaro-world “multi-vendor collaboration”... I can feel the space-time continuum collapsing around me :-), read on...

Stephen Foskett gives us the take-home. Essential Reading for VMware ESX iSCSI Users! - Stephen Foskett, Pack Rat.

• Ethernet link aggregation doesn’t buy you anything in iSCSI environments
• iSCSI HBA’s don’t buy you much other than boot-from-SAN in ESX, either
• The most common configuration (ESX software iSCSI) is limited to about 160 MB/s per iSCSI target over one-gigabit Ethernet, but that’s probably fine for most applications
• Adding multiple iSCSI targets adds performance across the board, but configurations vary by array
• Maximum per-target performance comes from guest-side software iSCSI, which can make use of multiple Ethernet links to push each array as fast as it can go

More like this, please.

From our Team Fusion blog: Windows 7 on Mac with VMware Fusion: A Practical Guide

There’s been a lot of buzz flying around about Microsoft’s next big operating system, Windows 7, which just entered public beta.

As our readers know, a great way to start testing out a new operating system is in a virtual machine, where you can see how it works without having to dedicate a whole physical machine. 

Well guess what: this is certainly the case with Windows 7 and VMware Fusion, too.  So we wanted to take the time to share best practices on how to make this happen, with screenshots and video of the process to make it nice and easy for you.

The secret is to use the Easy Install but change the OS from Vista to Windows Server 2008. And in 20 minutes or so ... enjoy.

Addendum: Rich Brambley at VM /ETC seemed to have no trouble installing the Windows 7 beta on VMware Workstation 6.5 by setting Vista as the OS, so not sure which way is best if you're on that platform. Check the VMware Communities for Workstation or Fusion if you're having trouble.

VMworld Europe is right around the corner. (You have registered, yes?) In the meantime, here are a few new items available at the 24/7 online version of VMworld:

VMworld.com Virtualization Events

VMworld.com Subscription

VMworld.com Virtual Exhibitor Booths

I normally prefer to quote regular customers talking about their VMware successes -- who needs to hear us quote ourselves about how great we are? But this one is pretty good, with EMC as the customer. (Of course EMC is Not VMware, but they do have a vested interest in us, so use whatever filters you deem fit; there are plenty of other successful VMware customer stories here.) EMC's Chad Sakac points to this latest case study on EMC and their experience virtualizing their Documentum eRoom implementation on VMware.

Virtual Geek: VMware and Documentum eRoom - or how EMC saved $1.5M.

1) the usual CapEx reasons (more important these days than ever!)

"Eliminated additional $1.5 milliion from planned budget expenditures and reigned in ongoing hardware and associated costs"

2) the normal "hey look at all this unexpected operational flexibility upside" reasons

"...several fail-overs have already been handled by VMware - the entire process only took a few minutes and the eRoom users never noticed a disruption.   In the past, such a failover would have significantly impacted users and produced longer down times."

3) the always fun "wow - performance is good.   WOW - really good!".    Everyone always is worried about this, almost always incorrectly.   Higher utilization can also be expressed as higher efficiency, which for some workloads has an effect of getting MORE on the same hardware (i.e. in this case we could have more front-ends).

"the final performance results acheived were significantly greater than expected.  Users were both pleased and astounded with the increased performancel eRoom activities such as site access, opening stored content, and saving files, are now instantaneous.   Response delays have been eliminated, yet the virtual site configuration is nearly identical to the original physical hardware configuration."

And the money quote I'll retype for you here, since somebody at EMC decided to password-protect the case study PDF. (Hint, folks, you want people to cut-and-paste from your success stories.)

"We saved the company $1.5 million right off the top, plus cut ongoing costs for floor space, power, depreciation, and labor -- and we get all these extra capabilities like easy expansion, smooth and plentiful fail-over, and faster performance. Basically, there was no downside." --Axel Kehlenbeck, project manager

A million and a half here, better SLAs there, and pretty soon you're talking about something worth doing. We won't be talking about eRoom, but we will be talking about virtualizing your enterprise messaging apps this Wednesday at noon PST on our VMware Communities Roundtable podcast.

I've had a number of conversations today that have made it clear to me we have a bit of a misunderstanding. The vExpert awards are not a popularity contest. The awards will be made on the basis of your activities in 2008 for the community. Do not spam all your friends and colleagues asking for testimonials to be sent to the nomination site.

There is no voting. Multiple nominations will not help you. Instead, a committee will look at the hard work you've put in spreading the good word about virtualization in 2008, and will pick out the most worthy virtualization advocates. Having hundreds of people submit nominations will just make more work for me, and it's never a good strategy to piss off the chairman of the judging committee.

Why did we do it this way? My understanding is that Microsoft just makes the decisions internally and then sends emails out informing you that you've gotten an MVP award. Similarly, while we started with some internal lists, we chose to have a few weeks for public submissions.

Why make all this noise and fuss? First, we felt like we might not know everybody and everything that's going on all around the world. And second, even if I know you as a blogger or Jean knows you as a VMUG leader, we might not realize all the cool stuff you've done outside those limited views.

This strategy is already paying off. Already from the entries so far I've found out that a community regular is also active on some other online forums and answers virtualization and VMware questions there, which is a plus to me. Another blogger, a partner, holds regular meetings with his competitors in the region to compare notes and talk technology. This is also a plus to me. (Small fib: I actually found out both these facts via email, but that kind of thing should be in your essay, dammit!)

So, please don't be modest. Nominate yourself. You know better than anyone else what you were up to last year. If it makes you feel better, I can change the title of the form to "Application," which was its original title anyway. Really, don't be shy. It's my fault for using the word "nominate" and not clarifying how the process was going to work.

I realize there is value in being recognized by your peers, and maybe later we figure out some other way to capture that, like a Virtual Prom King or VMware People's Choice Award or something. But right now, for this award, I don't want to penalize people without big megaphones. I want to find the people who have earned it with their actions.

The last 24 hours did have one good effect from my perspective! Many blogs have posted about vExpert to spread the word. Thanks! I really do appreciate it! And no worries -- you won't be penalized if lots of your readers and colleagues have nominated you, but people are going to get tired of our blogs if we do much more of it. My fault. Please accept my apologies.

I do encourage you to submit nominations for other folks that I might not know about, or to tell me a story I haven't heard.

Thanks! The response so far has been amazingly great, both out there and also internally here at VMware. We're going to have a good time together this year.

Questions? Feel free to shoot them to vexpert@vmware.com. That goes to me, but it lets me organize my inbox a little better.

Have a great weekend!
John

VMware: Uptime (VMware and Business Continuity): How to exploit the test bubble for all its worth.

It is important during a test failover to test your virtual machines without them being visible to the production systems.  Things like IP and name conflicts can happen if you are not careful and they will ruin your day.  So isolating those VM’s is important but we need to do that in a way that we can still properly test.

If you have only one ESX server you can use the automatically generated test bubble network that connects the VM’s together.  This private virtual switch will allow VM’s to talk to each other without the traffic leaving that switch.  So we are preserving the important isolation.  This is the default for a recovery plan.  It is shown as Auto in the Test Network column when you edit the Network part of a recovery plan.  You can see this in the screen shot below.  You can learn more about this on page 62 of the Admin Guide – URL for it is below.

When you have multiple hosts, and you need VM’s on each of the hosts to communicate with each other we cannot use the private virtual switch method as it has too much isolation (it does not span hosts).   The solution to that is to use an isolated VLAN.

We had some technical difficulties with our podcast provier this week, so we held a short roundtable session. By the time we got on the air, I'd let our guests go for the day, but one, Jon Dybik, did stick around, so we got to talk about virtualizating Microsoft Exchange for a bit anyway.

Still, we did decide to reschedule for next week. #32 Roundtable will be next week, Jan 21, at noon PST, and the topic will be, once again, Virtualizing Exchange (and Domino and BlackBerry).  More info here. For this week, listen via the widget on your right or the mp3.  Links for our short podcast this week:

Microsoft virtualization licensing

• Microsoft Licensing FAQs
• Windows Server Virtualization Calculators
• Server Virtualisation Part 1 - Server Operating System Licensing - Windows 2008 with/without Hyper V from Microsoft's Emma Healy
• Hey Dad - Microsoft Virtual Machine Licensing Policies (fun video explanation)
• Chris Wolf on licensing: Interpreting Microsoft’s Revised Virtualization Licensing Policy

Edward Haletky's Virtualization Security Round Table Podcast - go NOW (Thursday @ 11:00 am PST / 2:30 pm EST)

Virtualizing Exchange

• Everything you need is right here on this page: Deploy Exchange on a Dynamic Platform.
• Skip right to the Resources tab if you must.
• This white paper is getting good reviews: Deploying Exchange Server 2007 on VMware Infrastructure: A VMware Internal Case Study

A few years ago, I asked one of the most prolific posters on the VMware Communities if he wanted to be recognized for all the effort he was putting into answering questions on the forums, and the obvious breadth of knowledge he displayed. "Oh no," was the answer I got. "I don't want my boss to know how much time I'm spending on the site. I've learned so much doing it, but I don't think he would understand." I hope your boss has a more enlightened viewpoint than his did, because it's time to recognize your contribution, and your expertise, in public.

VMware is very proud to introduce the VMware vExpert Awards. The VMware vExpert Awards will be given to individuals who have significantly contributed to the overall community of VMware users over the past year, either online or offline. You might be contributing online to blogs, forums, wikis, or other online sites. You might be organizing VMUG meetings or otherwise getting the word out to local IT professionals. You're helping spread the word about virtualization and making people successful in deploying this game-changing technology. We want to thank you.

Here is the nomination form. You can nominate yourself or someone else. We're also taking nominations from people and programs within VMware, so you actually don't even have to nominate yourself and you still could wake up one day with an award. You might want to fill out the form anyway, because we may not be aware of everything you're doing.

Awardees will get an award certificate and a little gift, but the real award is being able to call yourself a VMware vExpert for the year, getting listed in a public directory, and getting access to our private community. The vExpert award is not any sort of technical certification like VCP, and in fact vExpert isn't really about your raw technical expertise at all. However, the people who get this kind of award tend to both (1) display a lot of technical expertise as they help others; and (2) work their tails off in the community on top of their day jobs.

The VMware vExpert program will likely get talked about as "VMware's MVP," and while we'd like to acknowledge inspiration from Microsoft's very successful program, please don't compare them directly. Microsoft MVP has been going on for over a decade and includes a 4-day conference -- and the vExpert won't even get you a free ticket to VMworld! :-) Our first goal simply is to publicly thank you and acknowledge all the hard work you've been doing on behalf of virtualization and VMware. That being said, when you get this award you'll be visible to various teams inside VMware and may be tapped to participate in various activities.

If you aren't active in online places like the VMware Communities or the blogs, or if you aren't active in your local virtualization community, then this award isn't for you. However, if you've been thinking about getting started, I hope these awards are an inspiration. Being active in the community is a great way to get exposure to technologies that you don't normally touch in your day job and to keep learning every day. (As the Roman smart dude Seneca put it "We learn by teaching.") In general having a higher professional profile is a career accelerator. I hope at this point most bosses agree that the VMware Communities and sites like it are valuable for all IT professionals in 2009.

Finally, I also want to thank everybody in the community that I've had the privilege to work with. You're super-professional, super-knowledgeable, and you've taught me to tell my LUNs from a hole in the ground. Being able to help shepherd these awards out the door is both personally and professionally rewarding. I'm looking forward to hanging out with the VMware vExpert Award winners of 2009!

John Troyer