Wireless PCIe To Enable Remote Graphics Cards

J. Dzhugashvili writes "If you read Slashdot, odds are you already know about WiGig and the 7Gbps wireless networking it promises. The people at Atheros and Wilocity are now working on an interesting application for the spec: wireless PCI Express. In a nutshell, wPCIe enables a PCI Express switch with local and remote components linked by a 60GHz connection. The first applications, which will start sampling next year, will let you connect your laptop to a base station with all kinds of storage controllers, networking controllers, and yes, an external graphics processor. wPCIe works transparently to the operating system, which only sees additional devices connected over PCI Express. And as icing on the cake, wPCie controllers will let you connect to standard Wi-Fi networks, too."

Comments

Question

[The Clockwork Troll]

To those in the know, why will this succeed where UWB/wireless USB failed in the market?

Remote graphics seems like an even more esoteric need than the remote mass storage, printing, cameras that UWB would have offered?

Re:Question

[fuzzyfuzzyfungus]

I have no idea whether it will go anywhere; but I'd assume that the one major strike in its favor is that, unlike wireless USB, wireless PCIe addresses use cases that basic boring ethernet/wifi do not.

The performance of early wireless USB hardware was pretty shit, and it was uncommon and ill standardized, so you usually still had to plug a dongle in, just to get performance worse than plugging in a cable. Plus, basic NAS/print server boxes had become really cheap and fairly easy to use. Anybody who wasn't a technophobe or living in a box(and thus not the target market for pricey and sometimes flakey wireless USB) already had his mass storage and printers shared over a network, wired or wireless, and his human interface devices wireless via bluetooth or proprietary RF, if he cared about that. Wireless USB didn't really enable any novel use cases that anybody cared about.

On the other hand, there is basically no way of plugging in "internal" expansion cards over a network(in the home context, I'm sure that some quite clever things have been done with I/O virtualization over infiniband, or whatever). Particularly with the rise of various "switchable graphics" technologies, I assume that the use case is basically this: User has nice thin, light, long-running laptop. They come home, sit within a dozen meters of a little box(containing a graphics card or two, with one head connected to their TV), and suddenly their laptop has a screaming gamer's graphics card supplementing the onboard card, either usable on the built-in screen, or via the second head connected to the TV, or both.(Analogs could be imagined for professional mobile workstation applications, where simply sitting near your desk connects you to a quartet of CUDA cards and an SAS controller with 4Us worth of drives hanging off it.

Will the market care, enough to bring the volume up and the price down? I have no idea. However, it at least has the advantage of allowing things not otherwise possible, unlike wireless USB, which pretty much covered the same ground as a mixture of bluetooth peripherals and resource sharing protocols over TCP/IP; but years later and without the standardization..

Good news, everyone!

[Drakkenmensch]

We'll soon have ONE MORE wireless signal to keep track of, when all those we already have work so well together!

Re:"Band"-aid

[Anonymous Coward]

Wait, are you saying that not requiring a license for 2.4GHz was a bad thing? That's the only reason wifi took off.
What sucks about wifi is the minuscule width of that unlicensed band.

Neat idea but it'll suck where it needs to shine

[jtownatpunk.net]

Let's say I've got even a little building with 50 people who want to use this. Will I be able to pack 50 of these point-to-point units into a building and have all of these systems perform at peak capacity without stepping all over each other? That would be amazing.

And, aside from the technical issues of getting it to work well in a dense environment, there's still one cord that needs to be connected to the laptop. Power. If I have to plug that in, I may as well snap the laptop into a docking station and skip the wireless connection entirely. One connection is one connection and I won't have to worry about interference, security, bandwidth, etc.

Re:I must admit...

[blueg3]

You're unlikely to be able to *alter* PCI traffic, though you could perhaps *insert* PCI traffic.

Still, people figured out properly encrypting wireless links some time ago. Tempest is primarily interesting because the signals you're looking at are unintentional (and often unknown) side effects and they often deal with links that are impossible or unreasonable to encrypt.