Gabe Knuth

Early-bird registration for BriForum ends today!

Gabe Knuth — Fri, 19 Jun 2009 17:54:00 GMT

Just a quick note to let you all know that today is the last day that members of BrianMadden.com can receive the special $300 discount off the registration fee for BriForum. If you're not already a registered user, you can sign up today and still qualify! When you have a user account, contact Jackie Anderson at (781) 657-1380 to receive the discount code.

We also are offering a discount to past attendees, so if you'd like to come but haven't yet made arrangements, give Jackie a call today!

VDIworks releases new VDI remote desktop connection protocol. Yawn or Yay?

Gabe Knuth — Thu, 18 Jun 2009 04:00:00 GMT

This week, VDIworks launched a new connection protocol called VideoOverIP (VOIP? Taken. How about V/IP?) that purports to "drastically improve the user experience delivered in a hosted virtual desktop model." And while this isn't the first time that sort of thing has been advertised, rarely does an announcement from a relatively unknown company come with such promise. VDIworks was formerly the software arm of ClearCube until it was spun off earlier this year to bring ClearCube functionality to the VDI world.

ClearCube provides an excellent turnkey solution that uses proprietary blade PC's or servers, connection protocols and client devices to centralize desktops and deliver the user experience remotely. VDIworks has the goal of bringing ClearCube-like desktop delivery technology to mass-market hardware, rather than to proprietary equipment. This was timed so that VDIworks entered the VDI space in its infancy, while still bringing many years of desktop delivery experience. At its inception, VDIworks (who still has close ties to ClearCube) made available its VDI management software.

Since then, VDIworks has expanded its offering to include integration with System Center and an offline VDI solution called VDIworks2Go. VDIworks2Go is not the Type-1 hypervisor that's been all the rage lately. Instead it is checking out the server side VM to a local Type-2 hypervisor, which in this case is VMware Player. Of course, this means that the backend must be VMware ESX (Hyper-V and Xen are also supported on the backend), but there are plans to create an offline solution for Hyper-V and Xen as well.

That bring us to this week and the announcement of VideoOverIP. When I first heard about it, I thought "Great, ANOTHER protocol that only does a few things well..." Brian beat me to it and emailed VDIworks to ask why they would build yet another protocol when there are already so many out there. Here's the response from VDIworks CEO Amir Husain:

Unfortunately, if you survey the remoting protocol landscape you find that there are no options that are completely free of ties to particular brands of thinclients, hypervisors or server backends. There are some relatively more open protocols available, but they inevitably turn out to be low performance. So if, as a customer, you want to preserve your choice of hardware and hypervisor, while still expecting a decent user experience, you're flat out of luck. That's the issue we saw repeatedly and it became a customer refrain. That's why we developed VideoOverIP.

We don't believe one protocol will dominate all others anytime soon. We believe in heterogenous environments, flexibility, co-existence, and integration. If you've recently peered inside a typical corporate datacenter, you'll probably agree that openness and vendor agnosticism is the name of the game. In fact, that's preciely why our end-to-end physical and virtual managment solution, VDP, supports over a half dozen different protocols, including HP RGS, PC-over-IP, Clearcube TDA, Microsoft RDP, Wyse TCX and others.

So the short answer reads more like "All the other protocols are either not available to us or not good enough for us."

I L-O-V-E that answer, and here's why:

First, it's confident. I mean come on - there's so many protocols out there, even ClearCube's own TDA, and they didn't think any were good enough.
Second, it shows dedication. Plenty of other companies have made enhancements to existing protocols to provide a better or more realistic VDI experience, and they've done a decent job with that, but creating a new protocol in such a crowded house shows real tenacity.
Last, it signifies a real, forward-looking movement towards change. Rather than trying to extend something that already exists, VDIworks decided to do some real innovation while supporting all the existing protocols.

All that said, I still haven't played with it, and the industry has yet to put it through the ringer to see if it really holds up in a real-world environment with remote and LAN users. The bandwidth and preformance figures are unknown at this point, but enough buzz has been generated that it shouldn't be long before real data appears. VideoOverIP can be downloaded from VDIworks' website today, and my goal is to set up a comparison similar to the one we did at Qumranet's office last year. Stay tuned for the results of that and the final verdict on whether or not this is the real deal.

As for VDIworks in general, it's interesting to see a company who's sole purpose in life is providing a VDI solution, as opposed to an existing company evolving into the space using existing technology. The pieces look to be coming together, and they might have a pretty robust suite of products on the table in a very short period of time.

RedHat SPICE vs RDP vs ICA performance video

Gabe Knuth — Thu, 18 Jun 2009 04:00:00 GMT

I realized as I was writing today's VDIworks article that we never published the performance videos that we recorded last year when we visited Qumranet's (now Red Hat's) offices in the Bay Area. The information we got is very interesting, but publication was hung up when Red Hat acquired Qumranet, and the videos got lost in the shuffle.

The test methodology was pretty simple - we wanted to see how each protocol (SPICE, RDP, and ICA) performed in a controlled environment. We did this before we stress-tested the SolidICE and XenServer-based solutions, so this has nothing to do with server sizing or anything like that. This is a simple setup with one client talking to server running a single XP virtual machine. In each scenario, we ran through the same tests and recorded what the user actually sees as well as what the CPU utilization and network bandwidth metrics within PerfMon showed. Each of these scenarios can be viewed from within the player after it is launched:

ICA Protocol - Business apps on one screen
ICA Protocol - Business apps on four screens
ICA Protocol - Multimedia apps on four screens
RDP Protocol - Business apps on one screen
RDP Protocol - Business apps on four screens
RDP Protocol - Multimedia apps on four screens
SPICE Protocol - Business apps on one screen
SPICE Protocol - Business apps on four screens
SPICE Protocol - Multimedia apps on four screens

To watch the videos, click here or on the screenshot below. During each scenario, you can click the Swap button under the small video to swap videos from small to large. It's especially useful in this case since you'll sometimes want to see the user experience in the big screen instead of staring at PerfMon for the entire segment.

Why do we need "Software" client hypervisors?

Gabe Knuth — Wed, 03 Jun 2009 04:39:00 GMT

For at the past few years, I've talked about how I think that platform hypervisors are going to move towards the hardware in the future, leaving the current market players sitting not on their own hypervisors, but on a suite of tools to manage/extend the capabilities of the built-in hypervisor on whatever piece of hardware (and therefore embedded hypervisor) is purchased.

I think the same can be said for client hypervisors (called CHV from now on...unless I start getting paid by the letter). In fact, I think that it might be even more important for CHV's to be embedded into the hardware than platform hypervisors. The reason? Adoption. I believe there are the three key points to get CHV's adopted into every day IT, from the home PC to the corporate workstation:

CHV's need to be everywhere - every laptop, every desktop, every netbook. Nobody should have to go "get" a CHV--it should just be there already. Users should be using it (or have it available to them) without knowing it.
CHV's need to work with everything - offline VDI (all vendors), virtual security appliances (Neocleus style), operating systems, directories, networks, devices, you name it.
Everything needs to work with a CHV - the reverse is true, too. All OS's should have full access to all the devices (manageable, of course). None of this "one machine has access to the GPU" stuff.

One way I can see to achieve this is to embed the hypervisor directly on to the hardware. Forget Xen, ESX, and KVM. Forget key fobs and memory cards. The key to widespread CHV adoption is embedding it on every single new machine out there, lacing the hypervisor in directly with the hardware at the lowest possible level, while still leaving hooks so that management utilities can dig in and control access to devices, OS's, and system configurations as needed.

Now, I realize that the four companies (Citrix, VMware, Virtual Computer, and Neocleus) currently developing or selling CHV's are going to have a hard time with this, but let's face it, it's better for them, too. Instead of spending all their time adapting to each piece of hardware old and new, why not let the hardware do the work while they focus on what really makes CHV's cool - the management capabilities.

CHV vendors aren't facing the same market that the platform virtualization vendors faced when the "V-word" was new. Hell, platform virtualization almost sells itself. Run 10 servers on one piece of hardware? I'm in. Consolidate the space/resource consumption of your data center by 1/10th and recoup your new hardware expenses in a year or two? Sign me up.

CHV's, however, aren't in demand. There's no ridiculously obvious cost savings or increased efficiencies associated with them. You can't put a CHV on a laptop and have two people use the same device (I mean, you could, I guess, but they're not designed for that). In fact, the opposite is true - you now have a single device for a single user that potentially can require two, three, or more OS licenses! Add to that the fact that there are four CHV vendors currently in the space (With more coming, I'd bet - no KVM-based solutions are out there yet), and you've got a cluster...and not the redundant kind.

And so, the solution to widespread adoption is to embed it. Change the decision that people have to make from "Should we look at all these different client hypervisors?" to "Should we use this technology built in to the computer?" Make it like HyperThreading.

So how does a CHV get embedded into the hardware? Of course, Intel and AMD could sit down and come up with a way to build a CHV into their processors or chipsets, but BIOS manufacturers could also get in the mix. In fact, Phoenix Technologies (yes, THAT Phoenix Technologies) has a lightweight hypervisor called HyperCore that is actually based on Xen (just like three of the four CHV vendors). The first use of HyperCore is a solution called HyperSpace, released earlier this week, which is a sort of pre-boot environment that allows you to run certain Linux applications without booting into Windows. Admittedly, that's not a CHV like we've been talking about, but it is a hypervisor at a lower level in the system. The fact that it's based on Xen means that there may be room to expand on it and make it more robust. So what if Citrix, who's already going to give XenClient away for free, were to partner with Phoenix and other BIOS manufacturers to find a way to include XenClient in the BIOS?

The last thing I want to mention is EFI, or Extensible Firmware Interface. EFI is less like a bootstrap-on-a-chip BIOS and more like a shim OS between the hardware and the operating system (sound familiar?). All current generation operating systems support booting from EFI-based machines, so maybe it's worth looking for ways to build the hypervisor into the EFI on new computers. It's still not in the hardware, but it is one level lower than where we are today.

That's it for this drawn out thought! I'm anxious to hear what you all have to say, so post in the comments or shoot me an email.

VMware announces vSphere - the foundation for the next generation data center

Gabe Knuth — Wed, 22 Apr 2009 04:00:00 GMT

Yesterday, VMware announced vSphere 4 - the successor to VI 3.5, and its association with vCloud - a relatively new initiative to provide local, private cloud computing environments in combination with third-party external cloud providers in an effort to provide high availability and remote access to cloud computing environments.

I Twittered (tweeted?) about the launch event while watching it via the web (watch the archive here), but you can't do much in 140 character bursts. The overall message that VMware is trying to convey is that with vSphere 4 there no reason that you cannot virtualize any of your servers/applications. Between vendor representatives and live demos, VMware showed that vSphere 4 is considerably more capable and more efficient with power, storage, and resources than its predecessors. That's mostly words for now, but I'm sure when it's released (Q2 2009, so very soon) there will be plenty of benchmarks and comparisons to look at.

Add to that vCloud--VMware's initiative to create a standardized cloud computing platform that enables enterprises to build their own, private cloud that can also be replicated to a live off-site cloud computing (SaaS) provider--and you've got yourself what looks to be a next-gen datacenter. Actually, this software, combined with the hardware that Cisco is now selling as part of its Unified Computing System, touches nearly all the bases. The only gaping hole I see is storage, but I'm sure VMware can put you in touch with EMC :) With that in place, you can run a number of OS's with whatever application delivery mechanism you like (web, SBC, VDI, traditional), all the while having your entire data center existing virtually in an off-site cloud.

Cool technology aside, for me the biggest takeaway as it relates to our niche and the things we talk about on this site is that there was very little talk about VDI or DaaS. That's not an accident - I really feel like this doesn't have a lot to do with VDI. In fact, this falls in a level below VDI or SBC or any other application delivery method. This is about data center virtualization, not application or desktop virtualization. VMware CEO Paul Maritz did mention that VMware View is the plumbing to connect users to the applications in the cloud, but that only serves to confirm that vSphere is not a VDI platform - it's a data center platform. In some ways, it's almost better to just think of vSphere as hardware--hardware that you can use to support whatever application delivery/VDI solution you want to use.

Some other thoughts:

$3495 per processor! Holy frijoles! Admittedly, I'm not sure what a high end VI 3.5 license costs, but I do know that price means that a commodity, dual proc quad core box that runs what...$4k...will cost upwards of $11k by the time you add vSphere to it. Still, the same economies of scale apply - you can run, say, 10 (or whatever) servers on that one box, saving you $29k in hardware and a considerable amount on operational costs. The numbers work just fine in the end, but still...sticker shock!
There was no word on the pricing for the vCloud SaaS stuff, but that's probably quite a bit more complex. If anyone has had any dealings with this kind of situation and can share any thoughts about the process, feel free to do so in the comments.
Those "tweener" companies (as in be"tween" SMB and Enterprise) will stand to benefit from the SMB offerings. For some, the free ESXi and/or VMware Server work just fine, but there is a gray area that the vSphere SMB solutions should fill. If that can also be combined with vCloud for offsite federation of virtualized servers, I think that's a winner. SMB's typically aren't the best at keeping and maintaining backups (as well as extra hardware), so being able to have a hot-site-in-a-box would be extremely helpful.

Random notes:

Even though someone from VMware was probably not following my Twitter feed during the event, I'm going to choose to believe that someone read my tweet that said "I swear I see a rack full of Sun servers there. Someone should put an Oracle sticker on it," because about an hour later that's exactly what happened - a piece of paper with the word "Oracle" on it was stuck over top of the Sun logo.

The best line from this event came from John Chambers, CEO of Cisco. When talking about the relationship between VMware and Cisco, he said "really good execution will always beat good innovation," which is very close one of my all time favorite quotes by Steve Case, former AOL CEO & Chairmain: "In the end, a vision without the ability to execute is probably an hallucination."

I really enjoyed this from a presentation standpoint. Having been involved with the production side of events like this, I thought it was well done. Even the live demos went well (assuming they weren't pre-recorded :). There was still a fair share of marketing/gladhanding, but hey, it would've felt weird without it. So, to whoever put it together - nice job.

Gartner predicts $65.7 billion in VDI revenue and 49 million users in 2013

Gabe Knuth — Wed, 15 Apr 2009 04:01:00 GMT

"Wowsers!"

That's what Brian said to me when he forwarded me a link to an article about a recent Gartner report. The article reports on what Gartner expects the "Hosted Virtual Desktop" (that's Gartner's term for VDI, but more specifically, server-side hosted desktops - not including client hypervisors) market to do over the next five years.

For starters, in the next year Gartner expects the worldwide HVD market to grow from $1.3 billion to $1.5 billion--about a 15% gain, but that's not very "wowsers"-worthy.

What sticks out is that Gartner predicts the HVD market will grow to be a massive $65.7 billion market by 2013. There's the "Wowsers!" moment, especially when you consider that the number of units used to calculate this number range from 500,000 units in 2009 to 49 million units.

That means that the cost of a typical VDI implementation at this point is $2600 per unit this year, but by 2013 the cost per unit will drop almost 50% to $1340.

The article goes on to say that right now, VDI makes up "less than 1 percent of the worldwide professional PC market," and that in 2013, that number will grow to "more than 40 percent of the worldwide professional PC market."

So why is this so amazing? Well, for starters, the VDI market going from $1.3 billion to $65.7 billion is pretty interesting. Not only does that serve to validate this technology, but it shows that in a down economy VDI is going to grow at a phenomenal rate. And since Hosted Virtual Desktops doesn't include client hypervisors, that means that 49 million units number leaves out what could be another pretty sizable group of users.

Those figures take into account a drop in hardware prices, but an increase in the need for servers, bandwidth, and supporting software. Not an increase in price for those, necessarily, but an increase in the actual numbers needed to support the increase in users/servers.

Enough numbers. What's it going to take to for a vendor to take advantage of that kind of growth?

Remote Display Protocol - If a solution doesn't have a good remote display protocol, they're out. Those companies can only hope to piggyback on RDP7 (or 8 by 2013?) or hope that VESA brings something to the table with Net2Display that everyone can take advantage of.
Client Hypervisor - Having a client hypervisor might not make or break a company's chances, but it could provide a leg up. My guess is that there will be some interoperability, or at least some third party products that work with the big names. I imagine it as more of a value-add than a make-or-break feature, but who knows--five years is an eternity in this space.
An All-Encompassing Solution - Having an end to end solution, however, is important. This could mean an end to end software solution, but it could also mean an end to end hardware/software solution. For instance, Cisco continues to invest in software (although not in the VDI space), and has recently made a splash with the announcement that they will be selling blade servers as part of an overall data center solution. Imagine the overall solution if Cisco were to buy Citrix for their VDI capabilities and, say, DataCore for their storage virtualization technology. Wowsers.

So now I'll ask you - do you think these numbers are too optimistic or unrealistic? What do you think the VDI vendors need to do to capitalize on that kind of growth?

Is Net2Display dead or just dying?

Gabe Knuth — Mon, 13 Apr 2009 17:10:00 GMT

I was thinking about remote protocols today after reading about the funding that Teradici received, and it occurred to me that we haven't heard anything about Net2Display lately.

Net2Display (link goes to article from 2007), a projected open standard display protocol from VESA, first showed up on the radar in late 2006 after a video engineering conference. There was ample speculation in the following year, but since late 2007, we haven't heard anything new. That begs the question, "Is Net2Display dead?" We don't even know where VESA left off with it, so it's hard to speculate whether or not it's tied up in politics, technological problems, or dead and gone. Still, I'll try :)

The political tie-up reasoning:

While some will argue that an open standard is good for everyone, there's going to be a group of folks that don't agree. In this case, that group consists of private companies that are either licensing or are really close to licensing their own product that would compete with that open standard.

Net2Display has (had?) several companies on it's committee, namely IBM, Teradici, DeskTone, Avocent, and VMware. Those names are the most recent list I can find, having come from an article by Brian from late 2007. In fact, Brian pulled those names from a PDF from 2006!

Times have changed for a lot of those companies. In addition to everyone feeling a bit more competitive in a slumping economy, new relationships have been formed - namely Teradici and VMware. With two of the main proponents of Net2Display working on their own partnership that would isolate VMware's product as the best remote display solution (remember, Citrix says their working on something, but nobody's seen anything), wouldn't it be in their best interest to keep that solution close to their vest?

Would a VMware/Teradici co-op be better for those companies than an open standard...probably. I could make an argument that everyone could benefit from it, actually. With an adequate solution out there, Microsoft and Citrix will have to amp up their development efforts, and we could find ourselves surrounded by some really amazing protocols.

The technological problems reasoning:

VESA could have put the wraps on this protocol a while ago if they just decided they wanted to have a hand in the game, but the fact of the matter is that this is a tough problem to solve. There are many limitations in the current technologies that are constants - no matter what you do to the protocol, they will always exist. That means for a new protocol spec to be created, the wheel needs to be re-invented.

Other companies are working on solutions, but a broad local and remote, fast connection and slow solution does not yet exist. If VESA is waiting for this, then that would explain the delay. The complexities of a solution that robust are astounding.

The dead and gone reasoning:

Of course, it could just be that VESA decided to bag the whole thing and leave it alone. Maybe they decided it was best to let the dust settle, then come back and choose a direction. As I said, no solution currently exists that addresses all the needs of a next-gen remote display protocol, so they could be waiting to see if a new technology or approach emerges that they can use to build a foundation for a new protocol. It's a great idea. Hell, in 2006, it was revolutionary, but maybe it's a victim of bad timing from conception, through competition, and into a bad economy?

It's also possible that nobody "decided" anything. The Net2Display initiative is being driven by a committee, since it's an open standard being developed, in part, by for-profit companies, there could be a lack of incentive to put noses to the grindstone in an effort to get something developed. It could just be lingering in limbo waiting for someone to take the bull by the horns.

If not Net2Display, what will the next standard be?

At this point, it's hard to tell. The relationship VMware has with Teradici could result in something, but something more pervasive might be necessary to become a "standard." That's where Microsoft comes in. With the inclusion of the Calista feature-set into RDP 7, it very well could become the new de facto standard. Details are sparse as to what that feature set will be, but it stands to reason something will be included in the final release of Windows Server 2008 R2. If Net2Display is going to be the standard, it better get something out to the community pretty quickly.

I could be 100% wrong about reasons behind the Net2Display silence. I'd love to see it still around and moving along at a nice pace, but it's just been so long since any new information has surfaced. If anyone has any information, or if anyone involved with Net2Display cares to share the status of the project, please feel free to email me or comment.

Who's using the Citrix Branch Repeater?

Gabe Knuth — Mon, 13 Apr 2009 04:11:00 GMT

It's not meant to be a condescending question, "Who's using the Citrix Branch Repeater?" I really just want to know who's got it and how they use it. Since I first heard about the device sometime last year (or was it the year before...these product announcements are starting to run together in my head), I've been intrigued by it, but I've yet to get my hands on one (Maybe I'm not asking the right person. If you have one in a lab and can let me peck at it for a bit, send me an email).

In case you don't know about the device, it's used in tandem with a Citrix Repeater (formerly known as WANScaler) to deploy applications to a branch office. The Branch Repeater functions as the remote-side WANScaler and as a local site cache for virtual (ICA-provided) applications.

In addition to the "regular" Branch Repeater, Citrix also offers a Citrix Branch Repeater with Windows Server, which is basically a Windows Server implementation with a WANScaler client on it, tweaked to work as an appliance. This device does the same thing as the regular Branch Repeater, but adds in the functionality of serving as an on-site domain controller, DNS/DHCP server, and file/print server. Also, in addition to the thin-protocol application caching, it also servers as a remote cache for streamed applications.

Obviously, the Windows Server variant is the more intriguing of the two, since it's essentially a "remote site in a box." In past lives, I would've jumped at the chance to send out to a remote site a single box that could do what this one does. Properly managed, you could bring up a site in the fraction of the time it would take to create individual servers to handle each role, then ship, travel, and install them. Operations costs would be reduced, too, because there would be less power and space consumed at the remote sites, which in my experience are short on “real” data center space anyway. I’ve seen a lot of branch offices where the server(s) are placed in a corner of a telco closet, where they’re most likely used as a multipurpose table for anyone working in that room.

So, having no hands-on experience with it myself, can anyone share their experiences? Is the plain Branch Repeater worthwhile when there's a more powerful/feature-rich option available? For the Windows Server version, Citrix lists all those Windows specific features (DNS/DHCP, etc...), and tacks on an "and more," so is the Branch Repeater with Windows Server even more extensible? Can you, for instance, add Exchange to it and keep a local message store for the site where it resides? Or maybe put SQL on it with a local replica of the CRM data store? Maybe I’m asking too much :)

Please, post your thoughts in the comments below or send me an email at gabe@brianmadden.com.

Cisco now sells blade servers...wait, what?

Gabe Knuth — Tue, 17 Mar 2009 06:33:00 GMT

Today, Cisco announced their new Unified Computing System (UCS). UCS is a complete data center platform that incorporates network and blade servers that integrate into your existing Cisco core, as well as storage connectivity. The system can support up to 40 blade server chassis that each hold eight blades for a total of 320 blade servers.

We've been keeping our eyes on this for a while now, trying to see if it would actually happen or not. There's been quite a bit of chatter both for and against the idea of a Cisco blade server, but the detractors have primarily been talking about Cisco's Nexus line of chassis/virtualization technologies.

The question we've spent the most time asking ourselves has been, "If this happens, what does this have to do with us?" Cisco's virtualization vision has been evolving since at least 2007, when they invested heavily in VMware shortly before their IPO. Until recently, the most notable product result of that relationship is the Nexus 1000v switch, which is a software implementation of a Cisco switch that runs within the VMware vNetwork switch framework. That's worth mentioning, but hardly a fundamental application delivery concept.

So now that Cisco has entered the x86 server space and now competes with the very companies they once made products for (like the IBM BladeCenter line), what does this mean? Will you consider using Cisco products for your servers, or is your organization structured in a way that will make that hard? Companies I've worked for in the past have lacked the cooperation or integration between the network and server groups to simply drop in a solution that bridged the boundaries. Is the reason this is such a big deal simply because a networking company now sells servers, or is there a bigger revolution on the horizon? Cisco obviously has the latter in mind, but will it work out?

VMware releases open source VDI client

Gabe Knuth — Tue, 03 Feb 2009 19:14:00 GMT

VMware announced this morning that they've created an open source version of the VMware View client software called the VMware View Open Client. This new client will be compatible not only with VMware View, but also VMware VDM 2.0 and 2.1.

This client will be available to anyone, though it's intended for partners who previously had to use VMware's XML APIs to interface with VMware View. The new client will allow partners to create a native VMware View Client for non-x86 platforms as well as machines that are not running a flavor of Windows or Linux.

By "Partners," VMware (for the most part) means "terminal manufacturers" that want to bundle in a View client, but have products that don't run the traditional versions of Windows or Linux and/or on traditional hardware. VMware was cautious to note, however, that terminal makers should still certify their products through the VMware Ready Certification program since, in theory, anyone can now make a View client software package. If the terminal or software package is not certified, there will be no support from VMware.

Technical Details

Some features of the View Open Client include:

Ability to create a secure tunnel using SSL
Support for two-factor authentication with RSA SecurID
Novell SLETC Add-on RPM package
Full command line interface

The following features are NOT included in this release:

USB redirection
Multiple desktop sessions
Multimedia redirection

The system requirements for the VMware View Open Client pre-compiled binaries are:

i586 compatible processor
2MB of secondary storage
128MB RAM

VMware View Open Client is available under the GNU Lesser General Public License version 2.1 (LGPL v 2.1) and is accessible from http://code.google.com/p/vmware-view-open-client/.

BriForum DVD's for sale (finally!)

Gabe Knuth — Tue, 27 Jan 2009 19:44:00 GMT

At long last, we're able to make the DVD's from BriForum 2008 in Chicago available for sale!

This past BriForum was our biggest yet, with over 60 sessions presented by more than 20 speakers to over 400 attendees. Now you can see every single session through our special conference video player. For a preview of what the videos on the DVD will look like, click here to watch the session titled "The Excruciating Detail of the XenApp Logon Process," presented by Brian Madden and Kevin Goodman. Of course, the DVD does not contain any ads :)

The DVD sets are available for $495 from our online store.

Ericom's True Media Experience, or Reverse Seamless Windows

Gabe Knuth — Wed, 17 Dec 2008 07:39:00 GMT

Earlier this week, I had a briefing with Ericom and came away with an interesting bit of information about their upcoming PowerTerm Webconnect release. "True Media Experience," which uses a technology developed by Ericom called Reverse Seamless Application (RS) technology is a feature that takes application windows from the client device and laces them in seamlessly with their virtual desktop session, even preserving the z-order of each window (even alt-tab works). By running certain applications locally and seamlessly integrating them with applications running remotely, Ericom believes that True Media Experience (I'm going to call it TME now) is a viable solution to poor performance by graphics-intensive or multimedia applications in VDI scenarios.

With TME, administrators can specify where applications should run--local or remote. The desktop and Start menu shortcuts for local applications are available in the virtual desktop, so when a user launches an application the experience remains familiar. Using TME to run, say, Windows Media Player locally (but still appearing to within the VDI session) will provide the user with a rich experience while avoiding performance, bandwidth, and resource concerns on the back end.

Administrators can also configure TME to behave differently for local or remote documents/URL's. For example, the users' Internet browsers could be configured to run off the local device, but exclusions can be made so that corporate intranet sites are presented through the remote session. This allows for users to work from home or in other locations without WAN/LAN/VPN access, yet still have access to all the features of TME and all the features of being at their desk.

True Media Experience and Reverse Seamless application technology are slated for a late Q1, early Q2 release as part of Ericom's PowerTerm Webconnect product. During the demo I saw, the administration components were not shown, and I was told that an admin interface is in the works. By the second release of the feature, they expect to have a full featured admin interface in addition to other new features such as the ability to tunnel document/URL requests for remote items through the RDP protocol, so that all documents or URL's can be accessed locally and securely without the need to maintain lists of local and remote addresses.

Some of you might be thinking "Haven't I heard this before?" If you've been paying attention, you have! Back in April, mixed in with all the LufLogix madness, Brian wrote about Citrix Project Alice--one of the research projects going on at Citrix's Advanced Products Group in Sydney, Australia. One of the features of Project Alice is called "Looking Glass" (catching the Alice in Wonderland theme?), and is an attempt to do the exact same thing that Ericom's TME is doing (well, maybe not exact. Ericom's reverse seamless method is patent pending). As yet, Citrix hasn't rolled the feature into any of the products, and to my knowledge nobody else has released anything like it, either.

I'm left wondering if other companies have given up on (or are looking past) the technology, and here's why: It seems to me that offloading processing of certain types of content or applications to the local device is more of a stop-gap than a be-all, end-all solution. It doesn't really solve any of the protocol performance issues that are preventing VDI from providing a local experience, it just relocates some of the processing to the local device. So what if you're using a thin client? You can't offload much of anything to the local client, unless you're running embedded XP. TME is sharp enough to pick up on the fact that the client doesn't have the capabilities to run things locally, so the user will still be able to access the applications on the remote side (assuming they're installed, of course), but we're right back to where we started with the user experience.

So what about those other companies? I'm hoping that they're working on is the next-gen remote desktop protocol. Rather than focus on stop-gap measures, maybe Citrix is focusing on a solution. Please don't get me wrong - Ericom's solution is really cool, and it's something that can help get you over the hump until the next-gen protocol comes through, but in a few years I hope we'll have a complete solution that can provide a local experience with little or no client side involvement. That's the only way to get to Any (device) Any (connection) Any (application).

XenApp 5 Feature Profile - Citrix Universal XPS Printer Driver

Gabe Knuth — Tue, 16 Dec 2008 05:58:00 GMT

I hate printing. Sadly, it's one of those necessary evils like backups, patch Tuesday, and renewing your driver's license (where I've actually helped fix a printer...sigh).

That said, among the many changes in XenApp 5, Citrix has added some printing enhancements. In addition to the normal, EMF-based Universal Printer Driver (UPD) and the ubiquitous PCL drivers that we've come to know over the different names and versions, Citrix has included a new feature named the Citrix Universal XPS Printer Driver.

XPS, or XML Paper Specification, printing is a new feature of Windows Server 2008 that attempts to create a platform-agnostic electronic document format similar to PDF files. XPS printing capabilities are already built in to Windows Server 2008, Windows Vista, and Office 2007. XPS printing is also available on Windows Server 2003 and Windows XP with the installation of .Net Framework 3.0 or 3.5. Installing either of those versions of the .Net Framework will install the Microsoft component of XPS printing called "Microsoft XPS Document Writer."

NOTE: This next step is only if you're running Windows Server 2003. If you're running Windows Server 2008, you can skip to the next section - Configuring XPS Printing.

After that, you'll need to install the Citrix XPS Universal Printer component on to your application servers. To do this, run the Add Printer Drier wizard and install the drivers located on the XenApp installation CD. For Windows 2003 x86 servers, the location is \XenApp Server\w2k8\Program Files\Citrix\Drivers. For Windows 2003 x64 servers, the location is \XenApp Server\w2k8x64\Program Files\Citrix\Drivers. You'll be asked what type of printer you would like to install, and in this case it should be the Citrix XPS Universal Printer.

If successful, you should see the Citrix XPS Universal Printer driver in the print server properties (In the Printers folder, go to File | Server Properties).

Configuring XPS Printing

To configure your farm to use XPS printing, you'll need to create a printing policy that will dictate what kind of drivers the users' sessions will use. This step hasn't changed in many years - you simply select whether or not the session should use only the universal driver, the universal driver only if a native driver is unavailable, or only the native driver.

One thing to keep in mind, however, is that the default setting for XenApp servers is to use the EMF-based universal printer introduced with Presentation Server 4.0. To configure your servers to attempt to use the XPS driver first, you'll need to edit the registry.

Open RegEdit and browse to the HKLM\Software\Citrix\UniversalPrintDrivers key.
On the right side, you'll see the Driver List registry value. Right-click that value and select Modify.
Note that the default string is a list of abbreviations that looks like "EMF;XPS;PCL5c;PCL4;PS", and that XPS is the second in the list. To configure your server to try XPS first, simply move XPS; to the front of the list, so that it looks like this: XPS;EMF;PCL5c;PCL4;PS.
Once you've made the change, restart the Citrix Print Manager service.

What you've done is modified the priority of the universal printer drivers on the server. Each time a new session starts, the server will check to see if the client has A) a policy enforcing a universal print driver, and B) capabilities to support each of the methods in the registry key, in order. If a client does not have the ability to support XPS printing, then negotiations will move on to EMF printing, and down through the line until something compatible is found.

How does it work?

An XPS print job follows this process:

A user tells an application to print, and the application sends the print data to the printer object within the user's session.
The printer object on the server then sends the print data over to the XPS printer driver on the server, where it's converted into an XPS file (it's actually named *.xps).
From there, the file is sent to the client device via the printing virtual channel to something called the XPS Print Helper, which is installed with the XenApp Plugin for Hosted Applications.
At that point, if the XPS printer is not attached to a certain printer, the user is prompted to select a printer. If the XPS printer is attached to a certain printer (something that happens at the beginning of a session), the printer selection screen does not appear and the job is sent directly to the printer.

Wrap-up

Another day, another new feature. Thing is - this is more of an extension of a feature that Microsoft has added, but I'm glad that Citrix is taking advantage of it so soon. If you're using it and care to share your comments, please leave them below.

CherryPal - A cloud-based computer for $249

Gabe Knuth — Fri, 12 Dec 2008 16:12:00 GMT

This morning I was checking out Digg and saw a post about something called a CherryPal. Thinking maybe it had something to do with CherryOS, I decided to check it out. What I found had nothing to do with a PowerPC emulator for the PC. In fact, the CherryPal is a device used to connect users to a cloud (not "The" Cloud, since that doesn't exist).

In the fifteen minutes since I first heard about CherryPal, here's what I've found out: The initial review that I read claims that the device is rather slow, but it appears there is room for improvement. The device runs a triple-core, 400MHz Freescale processor with 256MB of memory and an 8GB solid state drive. Also included in the 10oz, 1.3" high x 5.8" long x 4.2" wide box is 802.11b/g WiFi, 10/100 ethernet, audio out, VGA out, and (only) two USB ports.

Among the applications accessible from the device (whether or not they're local or in the cloud I'm not sure) are Firefox, OpenOffice, Music Player, MPlayer for CD's and DVD's, instant messenging, and Skype.

A large part of the marketing effort for CherryPal is focused on the device being very "green". The device has no moving parts and only consumes 2 watts of power, which is pretty impressive considering the 500 watt power supplies in the PC's these days, but nothing new to the thin client world. CherryPal also says the devices use 80% fewer parts and will last for "10 years or more" - another benefit that we've heard before.

At first the device seems pretty cool, even though its a new spin on the same old thin client, but there are some glaring shortcomings:

It seems to me like they should include more USB ports, because there's no way to connect anything to the device after a keyboard and mouse--like a DVD player or an external hard drive.
A Skype conversation wouldn't be that useful because there's no microphone jack on the device, so you'd need to have some sort of supported USB audio device.

That's all I know about the device for now. I suppose I'll keep an eye on it and see what happens. If any of you have seen one in action and can share the experience with us, please leave a comment.

XenApp 5 Feature Profile: Preferential Load Balancing

Gabe Knuth — Wed, 03 Dec 2008 06:00:00 GMT

What is Preferential Load Balancing?

Preferential Load Balancing (PLB) is a new feature included in XenApp 5 (Platinum only) that uses administrator-configured levels of importance to assign more or less CPU cycles to users and/or applications. For instance, dispatchers and their applications in a trucking company might be assigned a higher level of importance because their job is more critical to the company's day to day operations than, say, data entry people using Excel. This increased level of importance results in more CPU cycles being dedicated to them and fewer to the less important users/applications.

Note: Preferential Load Balancing actually has very little to do with conventional Load Balancing, since they accomplish their objectives in very different ways. PLB is, in fact, an extension of the CPU Management feature that has been built in to XenApp/Presentation Server for some time.

"So what," you might ask? "I've been putting executives (and myself) on their own servers for years now!" This is quite a bit different than just managing the number of sessions on a server (resource-based or otherwise). With PLB, important users will actually receive more or less CPU cycles for their session based on something called a Resource Allotment.

Resource Allotment

This is really the crux of PLB. Rather than simply say "this user is more important" and leave it there, Citrix has devised a more complicated, and more fair, way of assigning CPU cycles to users. To do this, two importance levels can be specified by the administrators - the published application importance level and the session importance level. The combination of these two importance levels is the Resource Allotment. The higher the Resource Allotment for a session, the more CPU cyeles are dedicated to that session.

Note: Published application importance levels are configured in the properties of the published app, while session importance levels are configured via XenApp policy.

Resource Allotment values are determined by multiplying the published application importance level by the session importance level. Each level can be a number between 1 and 3, with 1 being the lowest, 2 being normal, and 3 being the highest level of importance. So, a session that has an application importance level of 3 and a session importance level of 2 would have a Resource Allotment value of 6 (3 x 2). By default, all applications and sessions have a value of 2 - Normal.

Preferential Load Balancing uses Resource Allotment not only to define the availability of resources to a particular session, but also direct the session to the proper server. In an environment using PLB, the server with the smallest aggregate Resource Allotment value of all the sessions will be the next to receive a session. This means that if Server 1 has a 15 sessions with a sum Resource Allotment value of 30 and Server 2 has 20 sessions with a sum Resource Allotment value of 26, Server 2 will receive the next session.

Enabling Preferential Load Balancing

To turn on Preferential Load Balancing, you'll first need to make some decisions about what should receive the priority - applications, users, or both. In some cases you might want to assign a high level of importance to the application itself, such as a the CRM at a marketing firm. In other cases, you might want to assign a high level of importance to a specific user or group, such as a vice president or CxO. Used together, this means that an important user who is also using an important application can have a very high number. The same is true in reverse - a very unimportant user using a very unimportant application could be relegated to a small slice of the CPU cycles on a given box.

Before configuring PLB for applications and sessions, you must enable it on each server (or farm-wide). To do this, open the Access Management Console and bring up the properties of the farm or server. Under Server Default | Memory/CPU | CPU Utilization Management, select CPU sharing based on Resource Allotments.

Next, to turn on PLB at the application level, open the Access Management Console and pull up the properties of the published application that you wish to modify. Then, under Advanced | Limits, you'll see a dropdown box labelled "Application Importance." Here you can select High, Normal, or Low, which correspond to 3, 2, or 1 when calculating the importance value of the application.

To turn on PLB for users or groups, open the XenApp Advanced Configuration console (the old CMC) and create a new policy (or edit an existing one). In the policy's properties, under Service Level, you'll see the Session Importance policy. When this policy is enabled, you'll be able to change the importance level in the same way as you do with an application. You'll still need to apply the policy to users, groups, or in whatever other way you'd like to enforce the new rule.

Important Considerations

Multiple applications per session

If a user is running multiple applications via the same session, there is the possibility that the individual applications will have different importance levels associated with them. In this case, the Resource Allotment value for that session is the highest of the individual application/session pairs. This means that if the user's importance level (assigned by policy) is a 3, and they are running Notepad with a value of 1 and SAP with a value of 3, the Resource Allotment value would be 9 because it is the higher of the two session X application values.

Unused Cycles

In situations where a session with elevated importance is not using all their CPU cycles, the unused cycles will be shared by the other sessions on the server until the more important session needs them. Think about it like sitting in the courtside seats at a basketball game...just until the ticket holder arrives.

CPU Management Tweaks

If you're granting additional resources to users using the CPU Utilization and Reservations registry keys as part of the CPU Management feature, Citrix recommends using PLB instead. In fact, if PLB is used, the entire registry key is ignored.

Wrap up

Compared to the last feature I looked at--Special Folder Redirection--Preferential Load Balancing is looking pretty useful. If you've already ponied up the money for Platinum, it's at least worth a shot. It can eliminate some of the performance problems that cost the company money at the same time as it placates the directors and other high-importance people. I'm curious to see what the people that have deployed PLB on a large scale feel about it. It's hard to gauge the performance and overall experience in a small lab, so if anyone has any experiences they'd care to share, please do so in the comments.