Yesterday Teradici released an update to their firmware that adds capabilities which enable their PC-over-IP remote display protocol to work much better over WAN-type connections with lower bandwidth peaks and higher latency.
In case you’re not familiar with Teradici, they’re a chip manufacturer whose custom chips allow remote desktop hosts and thin clients (with their chips) to communicate over their PC-over-IP protocol. PC-over-IP provides a user experience which is more-or-less indistinguishable from a local experience (even with multimedia, peripherals, etc.) when the client and remote host are connected via a LAN. But until now, they haven’t had a good story for WAN scenarios, as PC-over-IP didn’t really work too well in those cases. (If you’ve never seen PC-over-IP in action, I highly recommend you watch the 16-minute video I recorded last month showing the unboxing, installation, and demo of the Teradici hardware communicating via the PC-over-IP protocol.)
Yesterday’s announcement was about the availability of version 2.1 of the Teradici PC-over-IP firmware that supports a variety of WAN-like network scenarios. What’s cool about it is this does NOT require new hardware, and it’s NOT a new “WAN mode” or anything. Instead this is just some smarter algorithms running on the current generation of hardware that enable it to work over WAN connections.
- A new “dynamic image quality adjustments” feature will figure out what quality image to show based on the current network conditions.
- A “progressive image refinement” feature that will lower the quality of an image that’s in motion, and then raise the quality when the image stops. (Citrix ICA has something similar with their “progressive image display.”) Teradici’s been calling this “build to lossless," which is a cool way to phrase it.
- A new “local pointer” means that the mouse pointer (and its associated movements) are rendered locally on the client, rather than only being sent as commands to the remote host and waiting for the refreshes to be sent to the client. (This is how most of the software-based remote display solutions work, like RDP and ICA.)
- Finally, a “dynamic bandwidth sharing” capability ensures that multiple users sharing a single WAN connection can each get the best experience possible. (This is a combination of bandwidth shaping and automatic capabilities to detect the characteristics of the network which then adjust the bandwidth consumed on-the-fly. And like all good solutions, there are controls that let you "unfair share" it, should you need to protect certain users in certain situations. :) All of this is dynamic, allowing the system (and the admins) to change these things in real-time.
Teradici released this YouTube video showing version 2.1 in action with various applications running over 90ms and 150ms links, with 5mbps, 2mbps, and 1mbps bandwidth spike limits. As you watch this video, keep in mind how remote display protocols work with respect to bandwidth. The video shows PC-over-IP running with various bandwidth maximums defined, but on average, most Teradici users end up consuming more like 250kbps.
From a loistical perspective, a client running the new 2.1 firmware will be able to communicate with a remote host running 2.0 firmware (and vice versa), although you can’t mix-and-match 1.x and 2.x version of PC-over-IP. (But again, it’s all the same hardware, so it should be no problem to reflash your 1.x stuff with PC-over-IP 2.1.) Since Teradici only makes the chips, you need to go to you hardware vendor to get the latest firmware. Yesterday represents the day it was released to the OEM partners, so it could take some time to trickle down into general availability, depending on your hardware.
By the way, this 2.1 announcement is not related to anything that Teradici is doing with VMware for an upcoming “software only” implementation of PC-over-IP that may be integrated into a future version of VMware’s View product, although it does do more to prove the concept that yes, it is architecturally possible for PC-over-IP to work on a WAN.