Linksys Range Extender Question

Linksys provides firmware upgrades. I don't know of any 3rd party firmware.

The channel must be the same as the AP you wish it to associate with.

The AP association is specified by the channel, the MAC of the AP, and the WEP key.

The WRE54G ignores the SSID it is configured with, and will automatically change to use the SSID of the AP.

They don't make it exactly obvious though! Of course there are so few configuration options available that when I did it I just clicked on the one and only thing I hadn't tried, and there it was... :-)

I'm not sure what the previous firmware load did, but it certainly makes no difference, with the latest firmware, what you enter for the SSID.

Well, the other side of that is the situation where you have a permanently located client which just happens to be in a fringe area for both the AP and the repeater. Using the same SSID you cannot select which one it connects to, and it might well flip back and forth between them with odd results. If the two had different SSID's the client would be able to chose one or the other. (That can be done when a WRT54G is used as a repeater with 3rd party firmware.)

Thank you . I figures out there was a Linksys firmware upgrade and how to do it. Duh. Dumb me.

Linksys told me to use a different SSID I think because they felt you might pick up the weaker signal by accident. But you are right if you start roaming around your house, you would have to reconnect. Duh....Not my day for thinking.

Thanks Patty

For Firmware: Go to the Linksys website and click on support. Select Technical Support, Click on Product Pages, Select Driver.Firmware in the dropdown box, then select your product model.

A range extender is a repeater and must be on the same channel as the Router. I see no reason in having 2 separate SSIDs, if you want to roam around the house and stay connected then stick with one SSID, otherwise you will disconnect and reconnect everytime you get in or out of range of one or the other.

My experience has been that my client hardware can be given the MAC address of the AP... and it will be totally ignored if there is another AP with the same SSID and a better signal!

That is with Broadcom based hardware, YMMV of course.

One example I've seen was where an AP and a repeater were both within range. The client was directly across the road from the AP, while the repeater was down the road several yards. Usually the AP would have a 20 dB better signal than the repeater; however, frequently a vehicle driving on the road between the AP and the client would result in a path fade, and the client would switch to the repeater, with a barely but usable signal.

Indeed the client was reluctant to switch... and would remain associated with the repeater until its signal would also take a severe enough hit. In the mean time, performance would be seriously degraded compared to what would have been available from the AP (the repeater of course would at best have half the data rate).

In essence, it avoided the momentary hit by switching, which indeed resulted in avoiding a total loss of signal perhaps, but also provided noticeably lower performance for a significant length of time too.

I'm pretty sure I've tried it with both HyperWRT and Sveasoft firmware, but can only positively recall checking that with Sveasoft.

snipped-for-privacy@barrow.com (Floyd L. Davidson) wrote in news: snipped-for-privacy@barrow.com:

With the facilities of WinXP, that is correct. But most (?) wireless card utilities allow explicit selection between access points with the same SSID by using the BSSID.

I've not seen this in practice. Most (?) wireless card drivers seem reluctant to change...

This is what I do in mixed 802.11g and 802.11b networks - use different APs with SSID specific to each. It improves performance, even though it would be possible to have a single large Extended Service Set.

Ah. I didn't know that - many thanks. Is this the Sveasoft firmware, or a different variety?

Kind regards

Has Linksys allowed for WPA encryption with the range extender yet?

Dave G

"Dave G" wrote in news:2SBVd.98505 $ snipped-for-privacy@tornado.tampabay.rr.com:

That simple question has a relatively complex answer...

The fundamental issue is the the strict definition and standardisation of WDS, Wireless Distribution System. While WDS is described in IEEE 802.1 it is not explicit, and implementations have differed between manufacturers. There is currently an IEEE task group looking at this as part of its work on mesh networks (TGs or 802.11s)

Since WPA encryption with TKIP uses the MAC address to derive (part of) the dymanic key, and in a WDS there are multiple MAC addresses, this leads to what might politely be called 'confusion' as the key is dynamically changed. In general, WPA will fail when used across a WDS link.

Currently it is only likely to work with compatible firmware and a compatible wireless chipset. I understand that the Sveasoft third party firmware for the WRT54G allows it, and also the Apple Airport Extreme Base Station. Both these are based on a Broadcom chipset.

All encryption-related issues are closely linked to the hardware, since the encryption algorithms are numerically intensive and rely on processing at the hardware level rather than in the general purpose cpu within a router's (or other device's) communications processor. And Broadcom has implemented WDS functions in its hardware and reference firmware, which makes it easier for vendors with Broadcom- based products...

So even if it were confirmed that Linksys had implemented a 'fix', it would likely be highly product specific.

Hope this helps - anyone else?

(...)

Good explanation. See:

18th line down proclaims "WPA over WDS" support. No clue which of the 4ea MAC addresses in WPA they use as part of the WPA-PSK (pre-shared key) or whether its does both WPA with RADIUS or just WPA-PSK.

Nope. I've never actually implimented a WDS bridge as a repeater. I tend to favour two back to back wireless bridges on different channels running full duplex. Therefore, no store and forward bandwidth reduction. More expensive, more complicated, more antennas, but twice as fast.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

I use and recommend this configuration. While there is a measureable throughput drop and an increase in latency, the drop in throughput is significantly less than the single radio repeater device.

However, my experience is that this configuration is also likely to fail with WPA. The reason is that such back-to-back solutions are generally built from a pair of multi-mode access points, one in Access Point mode, the other in Wireless Client (or AP Client) mode.

And all (?) recent multi-mode access points that I'm aware of implement Repeater mode. In other words, because they use WDS for the repeater mode, they are also likely to use WDS for Wireless Client mode. Which means that the frames will again have 4 MAC addresses - and as you point out, which of these is used for the key is unspecified.

I have tried and failed to get WPA-PSK working across a wireless client device, so I suspect it won't work when part of a 'back-to- back repeater' configuration. However, perhaps I didn't try hard enough - to be honest, I was expecting it to fail :(

Kind regards

Not mine. I use back to back DWL-900AP+, WAP54G, or other "transparent bridge" radios. They can shovel multiple MAC addresses and do NOT attempt to use the client mode. One radio is in bridge mode, the other in access point mode. The internet end of the puzzle is also a DWL-900AP+ or WAP54G in bridge mode. It then connects to some random router which then connects to a DSL or cable modem. Lots of boxes, but it always works. Sometimes, there's also a bandwidth manager in the chain (in various places depending upon what I'm trying to do). In any case, the only device that's doing client mode is the users wireless client.

Agreed. WDS and client mode seem to go together. I'm going to resist the temptation to look at the Sveasoft Alchemy source code to see how they implimented WDS. I got work to do tonight.

I don't see why it shouldn't work. The only open question I can think of is which MAC address to use for part of the encryption key. It's impossible to tell which key is used by sniffing, but looking at the Alchemy source should offer a clue (or answer). However, if someone became creative with the WPA-PSK implimentation, even if both ends used the proper MAC address, there will still be some form of failure.