http://www.engadget.com/entry/1234000497063795/

Just to remind all the haters that Google Labs isn’t the only research game in town, Microsoft Research has just busted out VirtualWiFi: a bit of software that makes your computer believe a single WiFi card is multiple configurable cards, allowing connections with multiple networks. There are all sorts of applications for this — beyond tricking people into thinking that you have more WiFi cards than them — such as connecting to multiple ad-hoc networks, or getting your internet from a for-pay WiFi subscription and then sharing it over ad-hoc (Engadget does not endorse any such usage of VirtualWiFi and predicts that your laptop will turn into a heap of ash if such usage is attempted). Sounds like fun, no?

http://research.microsoft.com/netres/projects/virtualwifi/software.htm


HOWEVER

The current version of VirtualWiFi does not implement some features. Please keep checking this page for updates. The features not implemented in this release of VirtualWiFi are:
# Buffering at remote nodes: We have prototyped this feature and presented the results. However, it is not implemented in this release.
# WEP and 802.1X: The current version of VirtualWiFi does not support networks using WEP or 802.1X. (booo, no wep?)
# Multiple cards: The kernel implementation of VirtualWiFi supports multiple cards. However, we have not incorporated this support in the user level code of this release.

[Edited on October 18, 2005 at 3:30 PM. Reason : .]

[Edited on October 18, 2005 at 3:33 PM. Reason : ,.]

10/18/2005 3:29:31 PM

This has spectacular ramifications/applications.



Thanks for the links!

10/18/2005 3:31:34 PM

which reminds me

are any of you familiar with the client/server application where:

the client buffers outgoing traffic and hops on to any open wireless access point it can find, when it can find one, for burst transmission/reception to the server

the server buffers incoming traffic for the client, does NAT/tunneling for the client, maintains the clients connections, and waits for the client to connect for burst transmission/reception of buffered data

the point being to maintain a somewhat unstable connection in a wireless device moving through an area of high wireless availability

I found this on google a long time ago and haven't been able to find it again

10/18/2005 6:05:58 PM

I see what you're talking about...that is a good application but I doubt you want something like that on a laptop.

10/18/2005 6:08:05 PM

yes I do

I'm on the road a lot with my laptop around campus here and it'd be real nice to have even retardedly high latency access to AIM

10/18/2005 6:11:21 PM

NCSU does not have a cellular network...

10/18/2005 6:14:16 PM

yes, but the 20 or so insecure access points per square mile in student apartments and shit are a reality

10/18/2005 6:16:29 PM

not exactly sure about how application layer networking works, but assuming AIM has its own software timeout mechanism in place (which i'd bet it would), wouldn't that defeat the point? you'd end up being automatically signed off in the 10 or so seconds without a connection, and your send buffer would be useless at that point.

10/19/2005 8:02:47 AM