I've written quite a bit about remote computing protocols over the past few months. In a world that's been dominated by RDP and ICA for a decade, it looks like a lot will be shaken up over the next 6-12 months. Certainly a lot of people are buzzing about Qumranet's (now Red Hat's) Spice protocol and Teradici's PC-over-IP, both of which are capable of delivering near-perfect remote desktop experiences, including multimedia and two-way audio. And then there's Microsoft's Calista and VESA's Net2Display, both of which have huge potential but neither of which we know much about. And we can't forget some of the smaller vendors who are making specific add-ons and improvements to RDP. Wyse has something called TCX, where they've added multimedia redirection and propoer multi-monitor support to RDP. (They've also OEMed this to VMware for VDM.) Quest Software's Provision Networks division has added or enhanced so many of RDP's virtual channels that you'd be hard-pressed to tell the different between it and ICA. (Except of course the fact that the Quest solution is about 1/4 the price of a Citrix solution.)
And then there's hp, with their RGS (Remote Graphics Software) protocol. A lof people have traditionally thought (myself included in the early days) that RGS was an extension to RDP. It's not. RGS is a 100% independent protocol for remoting graphics. It works by leveraging the remote host's processing capabilities and essentially turning the entire remote desktop into a streamed video, and it handles remote video and audio with ease. A side-by-side comparison of watching a moive via RGS versus RDP shows RGS is the clear winner. And RGS has some advantages over the protocols that simply "redirect" multimedia (Citrix ICA, Wyse's TCX, or what Quest is planning with RDP) because the RGS solution doesn't require that the specific multimedia codecs match on the client and the server.
So hp finds themselves with a nice little piece of technology without a real home. Sure, hp's managed to find some novel uses for it. They've bundled it in with their VDI and several thin client offerings. (Using RGS is as simple as installing a "remote sender" agent on the host, be it a terminal server for VDI workstation instance.) And hp is using it to connect to the consoles of guest VMs running on the embedded versions of XenServer on their Proliant servers. (Which, in-and-of-itself, is a really cool idea. It's almost like the servers with embedded Xen have an RGS-driven "virtual KVM" built right in. None of this BS about configuring a separate management server and all that.)
But apart from that, what can hp do with RGS? RGS looses value with every passing day as other vendors add features and functions to their protocols and start to meet or exceed what hp can do. But what would the point be for hp to invest more in RGS? It already does fine enough for remote guest console access to XenServers.
hp's RGS challenge is something I've been thinking about for awhile. I was specifically motivated to write this article based on a press release that came out while I was on vacation. It said, in a nutshell, that ClearCube spinoff VDIworks had licensed hp's RGS technology as the display protocol for VDIworks' new product. Maybe this is where RGS is destined to go? Maybe hp can license or OEM it to as many people as possible before some of these other protocols hit the street?
Of course if Red Hat does open source Spice (which is a personal fantasy of mine, not a rumor), then hp might as well just open source RGS, because its commercial value at that point would be essentially zero.