Re: virtual wireless interfaces / WDS
Ken Caruso <[email protected]> Mon, 14 Apr 2003 10:30:58 -0700 (PDT)
| Newsgroups | gmane.network.wireless.bsd.general |
|---|---|
| Message-ID | <[email protected]> |
In discussions I have had with people that know more about this subject than myself, we had concluded that having a virtual interface for each connection (such as sta0, sta1, etc...) might also prove usefull when one is trying to run link state routing protocols (such as ospf) over wireless interfaces. -Ken Ken Caruso [email protected] http://ken.ipl31.net "when in doubt tell the truth" -Mark Twain On Mon, 14 Apr 2003, David Young wrote: > The other night I sketched out some benefits of adding two kinds of > virtual wireless interface to *BSD. > > The first virtual interface type is probably familiar, since Linux has > it. Call it wds(4). It represents a link from AP to AP. A property of > a wds(4) is the RA field of the 4-address 802.11 header. You can bind a > bunch of wds(4) to a bridge(4) and run STP on it. Maybe you're familiar > with this notion in Linux. > > The second virtual interface type I call a sta(4). Each host associated > with a host access point may get its own sta(4). Also, each host in > an Independent BSS may get its own sta(4). Reasons for sta(4): it is > useful to keep per-destination statistics, WEP keys, Tx power, and 802.11 > parameters such as fragmentation threshold, RTS/CTS threshold, etc. I > think that this is all justified (necessitated, really) by the fact that > conditions on a wireless subnet are not ordinarily uniform, as they are > on a wired network, but some STAs may be more distant or they may have > weaker transmitters/receivers, they may or may not be mobile, etc. > > Sta(4) lets one use an existing abstraction in BSD, the interface, to > get link-layer information into a routing daemon. Just for instance, > as soon as a STA disassociates, you may delete its sta(4), which will > provoke an RTM_IFANNOUNCE on a routing socket and also take some ARP > entries out of the routing table (I think). An RTM_IFINFO can indicate > severe packet loss. These sorts of indications are important for routing > in mobile wireless networks. > > A sta(4) lets you track and control a station's data rate and media > (DSSS/OFDM). I think that it will also let you set per-STA queueing > disciplines using, e.g., ALTQ (hmmm---may or may not be useful). > > Like a wds(4), you will ordinarily put all the sta(4) into a bridge(4), > which will take the IP number. > > Is the usefulness of sta(4) clear? > > I imagine that the device attachment will look something like this, > > sta0 at wi0 ... > sta1 at wi0 ... > wds0 at wi0 ... > > Your thoughts? > > Dave > > -- > David Young OJC Technologies > [email protected] Urbana, IL * (217) 278-3933 > -- > *bsd wireless list, a bawug thing <http://www.bawug.org/> > [un]subscribe: http://lists.bawug.org/mailman/listinfo/bsd-wireless/ > -- *bsd wireless list, a bawug thing <http://www.bawug.org/> [un]subscribe: http://lists.bawug.org/mailman/listinfo/bsd-wireless/