Every BriForum finds a few recurring themes, and this year I had a handful of conversations with people about a rumor floating around that VMware was thinking of parting ways with Teradici (or vice versa) and the PC-over-IP protocol. Nobody could speak as to why either company would make such a move, but hearing it from more than one person is enough to trigger an investigation or, at the very least, some thought cycles.
Bear in mind that I'm not writing here that this is happening right now and all those zero clients you've bought over the years are worthless. This is just some healthy speculation into the relationship between VMware, it's protocol of choice, and the company that creates it.
One one hand, VMware and Teradici are a match made in heaven. Teradici makes the hardware and the protocol, and VMware integrates it into one of the top desktop virtualization platforms on the market. Both companies are doing what they do best, and nobody is extending into areas where they shouldn't be. The PCoIP protocol itself is fine, although at this point they pretty much all are, broadly speaking.
If you being looking for reasons to back up the rumor, though, you don't have to look far. When the relationship was first forged, VMware was 100% focused on VDI desktops. Today, they also have solutions for managing physical desktops with Horizon Mirage. Not a deal-breaker by any means, but indicative of a shift in desktop philosophy (a shift that I agree with) that doesn't make any money for Teradici.
Still, when it comes to virtual desktops, VMware is focused only on VDI, disregarding RDSH-based solutions. Sure, they've added the ability to connect to XenApp and RDSH environments into their solutions, but at the end of that day that's frustrating to organizations because they have two protocols to deliver to users. If you purchased a PCoIP thin client, you'd have to nest an RDP session inside a View session to deliver it to the device.
It's well within the capabilities of VMware to deliver an RDSH solution, but as yet they haven't done it. I believe this has frustrated Teradici, who recently released their own RDSH solution called Arch that uses the PCoIP protocol. Deployed in conjunction with View (it has to be–it depends on the View broker), you're able to publish applications from terminal servers to end users as well as desktops.
It could be that VMware left it up to Teradici on purpose as part of their agreement, but I don't get that impression. Why would a hardware company start making a desktop virtualization software platform? The answer, I'm sure, is that they thought VMware was leaving a big hole and not addressing a concern that customers have. The people at Teradici talk to customers all the time about buying thin clients, and I'd bet that the majority of those conversations say "You know, these zero clients look great, but since there's no XenApp functionality in View I have to be able to support my XenApp applications, too, so I need more than one protocol." Teradici did it because there was a hole in the market that VMware wasn't addressing and that directly impacted their sales.
So, if there is something brewing between Teradici and VMware, what does that mean for us? There's an interesting company called Framehawk that, should VMware acquire or partner with it, could make the transition relatively seamless. Depending on the agreement with Teradici, they may never actually have to turn off PCoIP, allowing companies to continue using it at their discretion, while also allowing users to use a new solution.
Framehawk came onto our radar from the consumerization vector because they've built a protocol from the ground up to deliver Windows applications via mobile connections. The Lightweight Framebuffer Protocol (LFP) was designed by people with a NASA background, where long delays and unreliable connections to spacecraft are the norm. LFP performs marvelously over 4G and 3G connections, and could be just the ticket for a company with a VDI solution looking to adopt a new protocol.
The other important thing about Framehawk is that what they do operates out of band from the rest of your desktop virtualization environment. Framehawk's system establishes a connection to your desktop environment via native protocols, then re-encodes the information into LFP before delivering it to the end user. It's very similar to Oracle's Secure Global Desktop or the late Sun Ray Server in that it can take many different application inputs and output them all via one protocol to end users. Framehawk can even do this from their own cloud service by connecting to your environment over a VPN. Since all you're sending to them is remote desktop protocol data, and all they're sending to users is re-encoded LFP data, it's a snap with little-to-no added security risk.
If VMware were to acquire Framehawk, they wouldn't necessarily even have to part ways with Teradici. Since Framehawk exists outside of the primary desktop virtualization solution, VMware could give you the choice of which protocol to use.
What else could VMware do if they were to break up with Teradici? RemoteFX support is an option that would make the most sense from a technology standpoint. RemoteFX has blended together the best elements of RDP, HDX, and PCoIP, using redirection, remoting, and UDP in the most appropriate situations, and is certainly the topic of much conversation around the industry. The thing is, VMware and Microsoft hate each other, so don't expect and grand gestures from either side.
Right now, everyone is mucking around with HTML5. I'd venture that Ericom is putting the most development effort behind it, but everyone has a solution these days. HTML5 isn't without its drawbacks, though. Native protocols are still going to provide the best performance, so if VMware and Teradici part ways I wouldn't expect to see them put all their weight behind HTML5.
Realistically, I'd say that Framehawk and LFP give them an ultra-flexible solution that can deliver any application from any platform. It currently only works on mobile devices, but as thin clients become more like screen-less phones (Android on a stick?), that just bolsters the case for a protocol like this that doesn't carry with it a lot of overhead. I'm sure that Windows-based clients can be quickly made as well, which would expand LFP from a WAN solution to a LAN one as well. Of course, there's a lot we don't know about in terms of video performance, bidirectional audio, and so on, but things like that can be added down the line, too (or, if you need them in specific situations, just deploy those desktops via PCoIP).
So that's what I know/think. Don't expect to see any big changes happen at VMworld this year. All indications are that there will be a minor release of View, saving the major platform release for next year. Who knows, though…perhaps we'll see a new View with View 6. If you want to see more of Framehawk, you can check out the video Justin shot of their booth demo at BriForum.