VMware accidentally leaks the feature list of View 5. Drum roll… underwhelmed

Yesterday VMware accidentally posted the Release Notes for View 5.0, the latest version of their VDI product which they'll be announcing at VMworld 2011 in Las Vegas later this month.

Yesterday VMware accidentally posted the Release Notes for View 5.0, the latest version of their VDI product which they'll be announcing at VMworld 2011 in Las Vegas later this month. The leaked document was quickly deleted, but not before plenty of people downloaded and re-shared it via twitter.

The features included are pretty much exactly what Bridget predicted a few weeks ago on SearchVirtualDesktop.com, although she also mentioned a type of disk image streaming that's so far not included. (Could this be the much ballyhooed VMware "OnDemand" technology they talked up in 2007 that has since vanished? :)

So what's new in View 5.0? According to the document, the big things are:

Persona Management (a.k.a. the Virtual Profiles product from RTO Software)

Well it's about goddamned time! Originally purchased by VMware in February 2010 the RTO Software Virtual Profiles product has been conspicuously missing from the past few versions of View. VMware originally included "Persona" (as they call it) in the betas of View 4.5, but then in vanished before the final release. The official reason given alternated between "the code wasn't up to VMware's standards" and "RTO didn't support Windows 7," although a third (and more plausible) reason some have suggested is that VMware was just waiting out the clock so they didn't have to provide a new version of Virtual Profiles to Symantec. Either way, it looks like we'll finally get it in View 5.

What will be interesting is whether VMware supports using this Persona Management for non-View Windows desktops in addition to View desktops. While a more simple VDI-only profile management might have made sense when VMware first bought RTO, the world has sort of evolved since then. Now the user profile management vendors have moved on to support single user configurations that span both physical and virtual environments, and in fact a lot of AppSense and RES Software's biggest sales are coming from purely physical desktop environments.

So if VMware comes out with a VDI-only (or View-only) profile manager, that will definitely be lame and not usable. (I say "not usable" since it's unlikely that anyone will be able to move 100% to View, which means they'll also have users logging into non-View Windows workstations. And if that's the case, they need a product to manage the user environment of those, so why not just use that other product to manage both the physical and virtual desktops?)

I'm crossing my fingers that VMware View Persona will also work in non-View environments, but I'm not holding my breath. Same goes for the Browns winning the Super Bowl this year.

PC-over-IP WAN performance optimization

The second major feature of View 5.0 will be performance enhancements to the PC-over-IP protocol, especially for WAN environments. Attendees of BriForum 2011 in Chicago last month actually know quite a bit about this, as Teradici CTO Randy Groves spoke at length about it in Teradici's sponsored session. (Watch a video of his presentation.) Specific new PC-over-IP capabilities include an updated and more efficient lossless codec, client-side bitmap caching (for clients that are able to), and the ability to disable the "build to lossless" capability.

The other recent big news for PC-over-IP from Teradici, while not specifically View 5-related, is that they finally released their APEX 2800 server offload card. (Here's a video I recorded of the card in action earlier this year.) Each card can support up to 64 1900x1200 displays, and you can put as many cards as you can fit in a single server. All the hooks to make it work are already built into View 4.5 and of course it will also work with View 5.

By the way, in his blog post about the new capabilities of PC-over-IP, VMware's Warren Ponder wrote, "we believe that the protocol debate has distracted from what really matters.  We think that it’s time for the industry to move on." Amen brother. After using both HDX and PCoIP exclusively for one month each earlier this year, I can emphatically say that they're both… fine. I wouldn't pick a product based solely on the remote protocol.

3D graphics support (with vSphere 5)

vSphere 5 introduces a new version of the virtual machine: Version 8. (The VM format gets updated every so often, and in general you can't boot an newer VM on an older platform without first converting it.) One of the new features of Version 8 VMs is that they'll support 3D graphics, including Aero and basic 3D apps. (And if I understand correctly, it can do this without a GPU, although I'm not 100% clear how that works. I'll find out more at VMworld.)

So if the VM can handle 3D and Aero, then VMware can remote it with PC-over-IP (and possibly RDP, again, I don't know for sure). But 3D graphics support when VMs are running on vSphere 5 is an official new feature, (which is nice, because VMware's sort of been implying that for over two years now).

That's it?

So it seems that those are the only major new features. Honestly I don't know what I was expecting, but I'm not really blown away by this. In fact what has the core View team been up to? Persona was something they bought. The PC-over-IP stuff came from Vancouver. And the 3D graphics is more of a core platform feature. So what are the new features of View itself?

Still no client hypervisor. Still no Mac PC-over-IP client. Still no real Remote Desktop Session Host support. Still no real integration with other desktop products.

Maybe they'll surprise me and blow me away at VMworld.

Fingers crossed! (Go Browns!)


Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Probably the most major ommission which hasnt been commented on is the dissapearance of CBRC support (see:- intellicache for View). Caching of golden images to cheap local SSD storage to remedy SAN crushing boot storms was the key feature many were waiting for in View 5 - and its "done an RTO".

Maybe we'll see it in View 6.5 ?


This just shows that VMware still has little focus on their "End User Computing" division.  Citrix continues to pull away in features and if VMware doesn't start innovating like they do with vSphere then they will not have a chance.


It;s my understanding from talking to some of the product managers for EUC (End User Computing) at VMware that the development cycle of the client has been separated from that of the platform, to speed up releases. This leaves me hopeful that a Mac PCoIP client might be in the works sooner rather than later!


I'm personally rooting for the Bills.. but same premise :)

I'm actually hoping VMware come to their senses and open up both the vGPU and "View Accelerator" API's so that people running XenDesktop on vSphere 5 can utilize them.

I don't see the benefit of VMware's vSphere tdev eam wasting any of their time building View-specific VDI enhancement features.


I believe the View Accelerator (a function of CBRC) is missing from View 5.0, despite being a listed feature for vSphere 5.

VMware really need to get thier head in the game with VDI before the RemoteFX/Hyper-V ecosystem spins up and destroys the benefits of PCOIP as the only viable host based VDI rendering protocol.

As soon as Citrix put together a viable RemoteFX (when on LAN) + HDX (when on WAN) capable version of Xendesktop, running on a Hyper-V platform (for the virtualised GPU capability) with Dynamic memory they'll clean up.

The biggest pain point to that roadmap right now is that vSpheres memory management is class leading - and really benefits VDI. Xenserver is okay, but its memory overcommit system is a real nuciance. If Microsoft can improve memory management, dedupe and overcommit in Hyper-V, and provide a more optimum management console, they're set!


I won't be surprised if this mediocrity feature set additions continues forever for View. Based on their history, they just want people on vSphere so they will just try to add the bare minimum "me too's" and stride to be just a "good enough" VDI platform added onto vSphere.

Smells like Microsoft 2.0.


VMware View 5's new feature set: www.youtube.com/watch


I personally want to see how good the WAN/low bandwidth enhancements are before passing any judgement. If they have done a really good job here it could shift the equation a bit.


There is a beta running at the moment - a vmware rep may be able to get you access if you ask nicely.


I'd like to think that product success isn't measured in terms of number of features, but rather the impact of the features in the release - as Steve comments above.

It's hard to understate the amount of engineering that has gone into the View 5 release, especially around key features like PCoIP optimization - this isn't a single feature, but rather an integrated set of capabilities that all work together to dramatically improve the user experience and reduce the bandwidth.  We are all extremely proud of the View 5 release and the beta is validating our faith.

The notes that leaked out were an early set, and yes, we have a few surprises up our sleeve for VMWorld. Keep in mind we're concerned with really pushing the envelope in key areas and not just baffling customers with a huge number of half baked features.  VMWare is an engineering focused company and we deliver solid, stable, and reliable products.

We've also separated out our client releases from View releases, so new clients are made available as soon as ready.  We're working on several other projects that are not part of the view release, but are pushing the envelope and some will be announced & shown at VMWorld.

Lastly, VMware has been shipping the market leading client hypervisor solutions for many years - Fusion and Workstation are doing extremely well  and are the right model for BYOPC - the market is continuing to validate that this is the right model.  

I'd be happy to meet with anyone at VMWorld to discuss any of these points - come to the View 5 Super session and grab me afterwards, or we can meet for a beer one night at the bar. We're serious about this space, and I'd love to hear your thoughts.


Uh oh, the VMware Marketecture man is back!



Ok, some valid points around PCoIP optimization but...

Why, and I must repeat WHY does VMware continue to relentlessly feed us garbage that CHVD is the best option for BYOPC. Clearly Citrix retracted this because their type-1 hypervisor is intrusive (yet superior in every other way), but they also realized that server hosted is the best case for BYO models. Now, in case they need to be offline with the data (which is a very niche use case) then they would need it client hosted which is where type-1 and type-2 client hypervisors come into play.

And for security and performance reasons, type-2 is my least favorite choice.



The BYOPD model relies on the ability to run enterprise apps in a secure container and we agree that it makes sense in many cases to have them server hosted.  But VMWare also offers the option to run locally with a secure container, and the success of Fusion and Workstation has demonstrated that there are many use cases where this makes sense.

Likewise our Mobile Virtualization Platform (MVP) is bringing the same option to mobile devices.  You can see this in action at VMWorld next week.

It's all about choice.




I don't think anyone is critisizing the improvements VMware are making with View at every itteration - the complaint seems to be that the wrong (or at least, less important) features are getting the focus. Now obviously its your job to draw attention to the good stuff, but its also got to be useful to you guys commercially if we feed back the things that we would appreciate seeing soonest ;)

The Delay of the RTO integration was painful. Its good to see its finally in there.

Likewise, any improvement in PCOIP performance is welcome.

Also - the variety of options for BYOPD deployments are also great.

However - are these factors preventing new deployments going ahead? To myself and others I have spoken to who are very much interested in deploying VDI at a business wide level, we need VMware (and your competetors) to address some key issues first:-

1) Virtualised hGPU capabilities (your all getting there - with Microsoft in the lead at present). Users want to watch Youtube videos and play Flash games. Annoying? Yes - if everyone interested in VDI worked in a locked down callcentre we'd be using Terminal Services anyway.

2) CPU cost of streaming protocol - I don't like having half my host resources hogged delivering those damn youtube videos (even if they are nicely audio synced). The APEX2800 should fix this for you guys - lets hope it isnt too expensive for scale out vs scale up.

3) Storage and commodisation of platform - This is the biggy The reason so many niche hardware vendors (V3, Nutanix etc) keep popping up is that they see a problem you arent fixing fast enough in software. The seeming lack of View Accelerator/CBRC support in View 5 is the single biggest flaw in my opinion with this upcoming release, and dwarves all of the benefits. This one feature totally alters the way VDI hosting platforms can be designed - saving us a tonne of money and design complexity headaches. Why you haven't moved heaven and earth to get this into v5.0 on day 1 is beyond me.

While it would be crazy of me to claim that my usage case is the only/main one, it certainly seems that it is the area where VMware could both hit Citrix and Microsoft on thier own turf, and expand the VDI segment as a whole.

BYOPD, client hypervisor etc are all nice to haves for certain cases (within the VDI field which is a niche as it is!) - but surely the big money will come from being in the driving seat of amarket boom that will only occur when you all realise that what the bulk of users want is for you to offer a VDI solution that gives as good (and flexible) a user experience as a Thick client, while remaining performant, flexiblem economical enough to convince ever corporate IT department that any minor pain or cost involved in migrating from those thick clients will be soothed by excellent maintainability and managability.

Basically - focus on features that will allow you to give us something that is at least as good as a line of PC's, for roughly the same price all up, and you'll start appealing to the massive market who don't care about user owned hardware.


Interesting link - seems that it is possible to enable CBRC through the command line!




Thanks for the great comments - this is the best feedback that I've had from this site.

I am not a marketing guy - my job is to define the features and capabilities that go into every release.  I run Product Management for all the Desktop Products, so you are talking to the right guy.

If we can get you into an NDA roadmap briefing I think you'll be pleasantly surprised that every item on your list is mentioned and our visions and priorities align.

We're also doing what we can to accelerate feature delivery and moving to a more agile development model, plus decoupling many components, such as the client releases.

If you're at VMWorld please let's meet up and discuss - would love to get more feedback.



Thanks Phil - its nice to see senior guys in Product who actually want to know what those down at the coalface actually think and will take the time to communicate on sites like this - the best we often see out here in the wilds of the web are the more adventurous (or lost) marketing drones!

Unfortunately I won't be at VMworld this year as I'm on business in sunny Singapore next week (much as I'd prefer the dry heat of Vegas!) but finding out whats on the radar post view5 issomething I'd love to hear about.

Its interesting to hear your trying Agile - its great when done right - oddly enough I've got a View vs Xen POC on the go right now to allow some of our own Dev teams (Who are in the process of moving away from waterfall to agile) to have thier custom configured environments that follow them from desk-to-hotdesk-to-home - thats the kind of situation where a well deployed VDI solution can have real and immediate benefits; even within the present limitations in the ecosystem.

I'd like nothing better than to push out a VDI deployment across the board but as with so many decision makers I want to be confident that if I do the product I settle on will do everything I need in terms of software featureset and hardware platform design fundamentals; Imposing new limitations or issues doesn't make for happy users - and my experience has shown me that getting user acceptance is always the biggest challenge when adopting new and different ways of working...

Hopefully going agile will help you speed up the release cycle of View and we'll all get to play with new features that much quicker - I guess it will be nice for you to be able redirect resources to focus on different features as need arise - with the pace and scope of the VDI market being what it is at present I expect choosing the next "key" feature to focus for a sprint can be a moving target based on what your competetors announce or what (impatient and opinionated) users (not that I know any of those...) demand.


#Phil (Wonderdog)


Thanks @Wonderdog

Safe travels to Singapore and please reach out when you get back and we'll organize an NDA briefing to get your thoughts.



You guys are better then Microsoft at selling "futures"

At least Microsoft learned how to make money from it