Citrix releases "XenClient 1.0," their client hypervisor. Don't waste your time. It's not ready.

Citrix has officially released their client hypervisor, called "XenClient." While this is something that we've been anticipating for well over a year, it appears that this "v1" has some pretty big limitations, and in fact many of the cool features are marked as "experimental.

Citrix has officially released their client hypervisor, called "XenClient." While this is something that we've been anticipating for well over a year, it appears that this "v1" has some pretty big limitations, and in fact many of the cool features are marked as "experimental."

We've been talking about client hypervisors ad nauseam on BrianMadden.com (including when you don't need them and when I think they'll finally catch on), so there's no need to go into the "what" and "why" right now. Suffice to say, whether you believe in them or not, Citrix releasing XenClient 1.0 is a big deal.

XenClient won't be a standalone product per se; instead it will be bundled into and included free with XenDesktop. There's also a "XenClient Express" which is a free edition that anyone can download and play with. (You can use the free edition on as many clients as you want, but central management, image delivery, and backup is performed by a server component called the Citrix Synchronizer, which is only free for use with up to ten clients.)

XenClient is based on the 64-bit version of the open source Xen hypervisor, and Citrix is officially supporting client VMs running the 32-bit versions of Windows XP, Vista, and 7, as well as the x64 version of Windows 7.

XenClient 1.0: Full of limitations?

Downloading and installing XenClient is pretty straightforward -- just make a boot CD from the ISO, configure some BIOS settings, click "next" a few times and you're done! But after reading the fine print you'll start to notice some weird things.

First is that you should keep in mind that Intel and Citrix have a relationship around XenClient. While details of their partnership are officially secret, word on the street is that Intel paid Citrix a lot of money for Citrix to ensure that XenClient only supports Intel vPro-based machines. And if you look at the meager hardware compatibility list that XenClient will actually run on, you'll see that yes, these are all vPro devices.

What's interesting is that about a month ago I blogged about Intel vPro and basically wondered why Intel forced ISVs like Citrix and RingCube to require it and wondered whether there were any real-world advantages to vPro? The defenders of vPro cited examples like TxT (Trusted eXecution Technology) as example of how vPro is awesome with client hypervisors, but then in the XenClient 1.0 User Guide we learn that TxT and TPM are not supported by XenClient!?! WTF???

And then let's not forget that in order for a laptop to have a vPro sticker on it (and, by extension, for it to work with XenClient), it has to have the complete Intel chipset, which means Intel integrated graphics only instead of Nvidia GPUs. (Although I guess this doesn't really matter since 3D support is classified as "experimental" in this release anyway... <sigh>)

At this point I'd like to re-ask my question from last month: Why does XenClient require vPro? What benefits to customers get from this?

Speaking of "experimental..."

So yeah, 3D graphics support is experimental and not officially supported in XenClient 1.0 at this time. Unfortunately it's not the only experimental feature.

"Secure app sharing" is also experimental. This is the feature that is essentially like seamless windows from one VM to another -- meaning that a user can access apps from multiple VMs via a single desktop instead of having to switch to each full desktop to use that VM's apps. Even if this wasn't experimental, configuration looks like a beast, based on local configuration XML files. Yikes! (Check out the User Guide Page 36... ugg!)

XenClient's "Dynamic Mode" is also released as an experimental feature. Dynamic mode is the "shared master" or "one-to-many" disk mode for XenClient, where a single gold master image on the network can be used by many different users at once. In the world of XenClient, users ought to be able to receive and refresh their base images while retaining their personal settings (documents, etc.), although right now that's not supported. Instead XenClient only supports the "one-to-one" or "persistent" image mode (called "Static Mode") where each user "owns" his or her own image. (So how do you update these things? SCCM? Symantec Client Management Suite? Why exactly are we using XenClient again?)

But before you spend any time trying to get the experimental Dynamic Image Mode to work, check out this little gem from the Admin Guide (p24):

Note: The user must manually join an Active Directory domain if required after downloading a Dynamic VM image mode VM.

By the way, the Dynamic Mode is completely different from Citrix Provisioning Server (PVS), which you could use with XenClient as long as your clients were always LAN-connected to your PVS host (which means desktop only, although in that case I think you'd just use PVS with bare metal since XenClient doesn't support enough different desktop hardware SKUs to necessitate a client hypervisor in the first place)!

And then there's the "known issues" list...

And just in case you still have any remaining excitement about XenClient 1.0, there are a whopping 59 (!) "known issues" in the Release Notes. Fifty-nine! And this is a product with a VERY limited HCL which only supports four different guest OS platforms!

Here's a sample of 10 of the 59 known issues. As you read through this, keep in mind that Citrix is actually trying to sell XenClient to real customers. Do you want to train your end users to look out for these ten "gotchas" (and the 49 others)?

Burning a CD using third-party applications may fail on Windows 7
Workaround: Please use the Microsoft Windows 7 CD-burning utility

Internal speaker may continue to emit sound when speakers are plugged in
Workaround: None available.

Issue: The integrated SD card reader available on some platforms presents itself as a PCI device, and is currently not supported for access by guest VMs.
Workaround: Use a USB SD card reader, if possible.

Windows 7 volume control behaves incorrectly at very low values
Workaround: Use higher volume settings or mute.

Cannot set password using localized characters
Issue: Setting a password via the system settings area of XenClient, does not allow input of localized characters.
Workaround: None available.

Hot-unplug from laptop dock causes screen to go blank, requiring a reboot
Issue: Dell E4310 with a 3D Graphics Support-enabled VM can become unresponsive if hotunplugged from a laptop dock
Workaround: Do not hot-unplug a Dell E4310.

Audio suffers when XenClient has heavy load on CPU
Workaround: None available.

Secure Application Sharing application may fail to launch on Application Subscribing VM
Issue: Intermittently a Secure Application Sharing application may fail to launch in the Application Subscribing VM, with no error message.
Workaround: None available.

Unplugging a USB audio device while it is in use on an XP VM can cause the VM to hang
Issue: Unplugging a USB audio device, such as a set of USB headphones while it is actively in use can cause an XP VM to hang.
Workaround: Ensure that the USB audio device is not currently is use (for example, currently playing music) and unplug the device using the standard XP mechanism.

Accessing a re-writeable optical device that is being written to in one VM from a different VM can trigger a BSOD
Issue: Attempting to read from an optical device that is being written to from a different VM can cause an error to be thrown. Subsequently attempting to write to the re-writeable optical device can trigger a BSOD.
Workaround: Ensure that the writing session in one VM has closed before accessing the optical disk in another VM.

Are your users ready for XenClient 1.0?

Ok, so who's ready for this? Who's going to start deploying it to customers and end users? Honestly I can't believe that Citrix is releasing this as a real product.. I guess this way their marketing people can go nuts telling everyone how awesome it is while in real life it's still a beta-quality product.

And actually I should point out that I still love the concept of XenClient. I'm not frustrated with the idea of it. I'm frustrated that Citrix is selling software that's not ready. Now folks are going to try this, see that it sucks, and be set back another five years before we can get them to try client hypervisors again.

Oh, and by the way, how brilliant does VMware's cancel-type-1-replace-with-type-2 strategy seem now?

Join the conversation

17 comments

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

Yet again another case of a companies marketing team dictating their strategy!


I agree this will do more damage in the short term and people will loose faith in the technology. It was clear from the last beta release it's still 12 months from being a GA product!


It also goes to show what a good job others are doing in thing space (Virtual Computer, etc ,etc) right now.


"experimental" features should always be a separate download/install. It makes you look bad to include them in your product! Or at the very least, make them a feature hidden from the end user so the admin can choose when to enable/disable.


Dan.


Cancel

In XenClient 1.0 it's easy enough for admins to protect the users from the experimental features. (In that they can be not installed.) Now protecting the users from the other inconveniences and known issues? That's harder...


Cancel

Thats a start...


If you hide all the experimental features though.... What's left? :)


Cancel

I've been running this 'not ready', 'experimental' stuff since the first Beta on my day to day business laptop and I can objectively say that it has caused me few problems.


I am an IT Pro and can troubleshoot and fix many of the issues that might crop up, but these have been few and far between, the Beta was one of the best Beta products I have worked with.


I agree that v1.0 might not be a product to roll out to the masses, but as with any v1.0 release it will only be the 'bleeding edge' companies that will want to use this immediately, and even these companies will select their pilot users carefully.


I disagree completely that this will negatively impact client hypervisor adoption, in fact I believe the opposite. I applaud Citrix for having the balls to put this product out there, for free so that the industry can dip their toe and start to adjust to the mindset change that will be necessary to adopt this technology.


Cancel

Daniel is probably right: A clear case of MKTG leading the product and not listening to the customer. However, it is also noteworthy that their plan is to include this with XenDesktop (free) and hopefully improve upon the technology regularly.


I suspect Intel saw through the limitations and with their acquisition of Neoclues can move forward in this space more in tune to corporate needs.


My feature list for XenClient would include:


1. Stream the hypervisor into RAM


2. Deployment support for PvS (stream multiple VHDs to the HV)


3. Support for Ghost or RIS


4. True Integration with AD


in the meantime, it is back to Virtual Computer for me.


Cancel

Yeah it's a bummer. I don't think this technology will be mature enough to have any leverage in the enterprise within the next few years.


My next foray into this technology will be when VirtualComputer NxTop Workstation 3.0 is released publicly (It’s a RC as of now). But that will be the free, standalone workstation edition without the central management.


Cancel

I would agree that the latest Virtual Computer hypervisor is certainly beyond this 1.0 product.  I have tested on a non "Pro" laptop using Radeon chipset and found the client to be "pretty reasonable"  VC was not quite to the point of what I want for my personal laptop (but then I'm picky), but very close.  I hope that the vendors don't loose interest because fast adoption doesn't occur.  It needs time to mature.  


We will all be doing this (client side hypervisor) in a few years if and when things mature, unless all we do are smart apps to cloudish back-ends.  While there will be a lot of success in the latter, just like web-apps it will not be able to do everything we need.


Cancel

I was looking at this as a replacement for the use of Parallels or VMWare Fusion on Apple's in my organization. They demoed XenClient (When it was Project Independence) at Synergy with a Mac running the software with no issues.


Then after several months of waiting for information on this being ready for beta, it turns out that it requires vPro to operate, and supports a limited amount of laptops, none of which are Apple.


I do understand that there may be some problems with virtualizing Mac OSX with regards to Apple's license, but why show and tout this feature, and then not have it available?


Cancel

I don't believe vPro is a technical requirement for XenClient, but it's 100% a marketing arrangement. Unless Intel make vPro free and part of all their laptops it's not relevant. A vPro requirement will hold back XenClient. Perhaps @Brian can challenge Citrix on this point in any conversations he has at Synergy.


I agree with those that say client HV will take several years to mature. Also I heard that Neocleus was basically dead because they tried to be a HV company. By the time they woke up it was too late to move into mgmt. Their CTO was an arrogant guy although smart and that pissed a lot of people off. Type 2 is going nowhere fast outside of developer type use cases and some users. MAC is certainly interesting until Apple wake up, but nonetheless it's not a long term strategy. Short term gain only and the real value is the management stack. I like Moka 5 here. Also worth pointing out this is hard and hence VMware has not delivered. Worrying for their credibility and if they say one more time that Type 2 was the plan all along, kick them where it hurts.


Ultimately if the OEMs will need to ship. That is what will determine adoption. This is where the likes of Virtual Computer need to prove traction or they will just be a sort term leader and will not have the ability to compete. I think Moka 5 beats them here as well. I also think Virtual Computer has a problem if they get traction as they are heading for a collision with Systems Management vendors. Hence why I think we are seeing *** management messaging from Citrix. Citrix just wants to get XC out there and evolve it's progress. They can buy a management stack later after the hard work is done that other will find hard to copy. Hence I see Citrix only offering crap management for at least a year.


To the point about why Citrix is selling it. To me it's another marketing sounds bite to encourage XD customers to try it and win mind share. I doubt we will see anybody use it in production. All those resources at Citrix should probably focus on fixing XD first....


Cancel

I cannot agree with you more on the statement that this is product is not ready for final release.  In fact I believe by releasing this a final product this could have a negative impact to Citrix.  How difficult would it be for another company (like VMware)to release something that would exceed xenclients supported systems or features?  Why market it as a solution for laptop users when there are so many other opportunities for a client based hypervisor.  I am afraid that Citirx is going to lose the client hypervisor game even though they were first out of the gate.  I guess we will wait and see.


Cancel

Yep,


Ready for beta customers ....!


If yu are beta company it's fine


Cancel

It least this blog brought attention to this site...


I do work with xenclient for a while and I'm quite pleased about version 1.0. But having said that IT IS A VERSION 1.0 and Cirix never claimed to be ready for a 1000 user deployment with this build....


I agree there are many things on the to-do list, but nothing which realiy would be a hurdle for starting pilots and acquiring experience...


And I also can't see any alternative product, I went through with virtual computing and I'm afraid the ONLY  advantage I've found is the larger HCL!


Cancel

Tsk. Tsk. It's a 1.0 product.....


How many 1.0 products are used in production? How many organizations jump on the next gen of Windows when it's first released? Give it time for adoption.


It's better to release a "working" product now for adoption than to wait... I have been anticipating XC since the Project Independance announcement and I am satisfied with the progress so far.


Another year or two of waiting won't hurt, I am just too busy focusing on getting SBC VDI stamped for approval anyways so it works out just the same.


Cancel

Com'on guys, 1.0 on new tech and you're awaiting something ready for you're view of a full entreprise class deploiement ? R U real experienced IT Pro ? How many comment the fact that VMWare View need 18 months to fully support officially Windows 7 for a VDI solution ?


note : I'm a Citrix fan for the desktop and a VMWare one for the servers (at least for now ;-)


Cancel

Kata,


That's exactly the reason VMware delayed Windows 7 support. Instead of releasing beta software as production, they waited until have a quality release.


Looks like this is not the case o XenClient.


There's a huge different between 1.0 software and  beta quality software !


Cancel

@VMguy


That's a bullish statement.


VMware releases non-production ready sw all the time, "experimental" is just another name for beta.


Don't even insinuate the idea that VMware was waiting until they had a quality release. It just didn't work plain and simple.


XenClient is actually usable, albeit very limiting and not suitable for mass production but usable non the less.


Next you will say VMware "delayed" CVP because they wanted to release something better than XC 1.0... Give me a break.


Cancel

Upfront disclaimer, I work for Citrix as a Sales Engineer in Los Angeles....


For those Citrix customers in the LA area who have attended our Citrix User Groups, you have been following the evolution of XenClient as I have been demoing the product for a bit more than a year now.  To Brian's point, I tell customers that they will not be rolling XC into a production environment of 1000 seats anytime soon.  This is afterall version 1 technology, and it will take some time to work out the kinks.


What Brian does not mention in his article however is that Citrix has opened the invitation to white glove an implementation of 25 users for qualified customers (I don't do the qualifying).  If you are in the LA area and are interested in applying for an implementation such as this contact the local sales team via http://lacitrix.com.  If you are not in the LA area, contact your local sales team to follow up.


A word about the LA Citrix User Groups, during our last sessions we hosted the founder of XenSource Simon Crosby, and the Citrix Product Managers for XenClient Peter Blum and Chaitanya Upadhyay.  During our next round of CUGs November 3rd and 10th we will again be referencing XenClient.  On the 10th we will also be hosting Joe Nord.  Joe is a XenApp architect responsible for all things administration, user profile management, PowerShell, and IP optimization for Branch Repeater.  


Again, if you are interested in applying for a white glove implementation for XenClient please follow up.  We look forward to seeing you at the upcoming CUGs in LA.  Thanks.


RTE


Cancel

-ADS BY GOOGLE

SearchVirtualDesktop

SearchEnterpriseDesktop

SearchServerVirtualization

SearchVMware

Close