Extending a wireless network

Hi guys

I currently have cable broadband shared around the house with a wireless 802.11g router. It is fine in most of the house, but my daughter's bedroom is at the furthest reach from the router and the connection to it is patchy (keeps losing her MSN connection!)

I was wondering whether it would be possible to get an extra USB

802.11g adapter for my PC in my bedroom and then use that in point to point mode to connect to my daughters, and then bridge the connection to my good quality link to the router. So basically my daughter would access the Internet through two hops, one to my PC and one from there to the router.

Would this work? If so, can anyone give any tips about how to set it up on the two PCs. We are both running XP Home SP2. Would we use a network bridge or Internet Connection Sharing?

Any help much appreciated.


Reply to
Loading thread data ...

If you have a low signal, a reflector will likely fix the problem.

formatting link
EZ-12, printed on photo paper for thick stock, with aluminum foil glued to the sail, provides a substantial boost in signal.
formatting link
You could put one on the router, and anotheron the PC, if it has a stick antenna. You didn't say what was in your daughter's room.

On one PCI card, I added a "Hawking HAI6SDA Directional 6dBi 2.4GHz Antenna"

formatting link

Reply to

Reasons why NOT to use WDS.

  1. It only works with compatible chipsets and preferably identical model access points. Note that the original poster did not specify their equipment list and therefore may also need to replace their existing router with a WDS compatible router.
  2. The extra access point or router costs more than a better antenna. The added reflector types are almost free.
  3. I cuts the maximum throughput in half (which is usually not a problem with DSL intenet access but does cause trouble with client to client file transfers).
  4. Since everything has to be on the same SSID and channel, the extra radio just creates that much extra interference.
  5. It's apparently not very good in the presence of interefence. Lose one packet and each WDS access point has to repeat it twice, which really cuts down on the available air time.
  6. WDS repeaters tend to require omni antennas located at midpoint. Omnis are somewhat more sensitive to multipath and reflections than directional antennas.
  7. Some buggy implimentations of WDS interfere with the use of WPA. This is finally being fixed on current firmware releases.
  8. Setup is a bit complex, but not overwhelming.

formatting link

Reply to
Jeff Liebermann

Why not think about using WDS ( Wireless Distribution Services ) ? Specifically for scenarios like yours. Scenario:

Internal PCs ( x However many ) Internal Router External Router Cable Modem

The Internal Router acts as a bridge to the external. You need to make sure both routers are from the same manufacturer. Your thoughts re Network Bridging/ICS are also viable options.

The reason I prefer using WDS for this is that it's alot easier to move a router from place to place rather than a PC. If you move house for example, you can figure out where to strategically place your routers rather than being forced to make sure PC1 is in "firing range" of PC2. It also scales well, you want/need more distance, just buy another router and add it to the chain. WDS is supported on even the cheapest routers these days.

Rgds, S

Reply to
Steve Berry

Thanks guys for all your help. As a single mum I am struggling a bit with all this, but hopefully I now have some ideas to look into and come up with a solution.


Reply to

To the best of my knowledge, nobody has offered to send me money and make their purchasing decisions for them. I only supply the facts, numbers, technology, and detail. Users make their own decisions.

Nope. I supplied far more information about the relative merits of WDS than you did. You supplies a one line suggestion without the slightest substantiation and without explaining the limitations and potential problems with WDS.

Agreed. Could you cite the exact line in my posting where I hinted, told, demanded, or otherwise suggested what's "right"?

That's true. I didn't list the advantages of WDS because I thought you would supply the necessary detail. After all, it is your suggestion.

I believe I said that but thanks for the details. Generally, the

802.11g wireless network runs faster than the typical US broadband connections. It seems the UK cable broadband is considerably faster, so it might still be a problem.

Client to client transfers are not rate capped like broadband connections. They should go at the full speed of the wireless devices. For example, a 54Mbit/sec connection to an access point will run at about 20-25Mbits/sec thruput when run through a single access point (or wireless router).

However, when you add a WDS store and forward repeater, only one of the two access points can transmit at a time (i.e. simplex operation), thereby cutting the maximum bandwidth in half. The same 54Mbit/sec connection at the end points of a WDS system will only have about

10-12Mbits/sec at best. My measurement with two WRT54G routers and Sveasoft Alchemy can only do about 8Mbits/sec. I'm not sure why and unfortunately don't have the time to investigate.

More commonly, the client radios cannot maintain a 54Mbit/sec connection and connect at lower speed. This would be expected as WDS repeaters are rarely deployed in situations where there is an adequate signal or clear path. So, the thruput can be even less than

10-12Mbits/sec or less. This may be adequate for many users, but I've seen a few complaints where users were expecting much faster client to client performance through WDS.

You might want to spend some time reading:

formatting link
's on the scaleability of mesh networks but the same principles apply to WDS networks. I don't agree with everything on the page, but the analysis and references are good enough to cover the major points.

Huh? There are at least 3 radios on the same channel in the same airspace. Two access points connected with WDS and a client radio. Interference between these is exactly the same as if you had placed a non-WDS access point on the same channel in the vicinity. The FCC does not allow synchronous operation between access points or clients so the interference is always there. I typically see about 10-20% resends at the MAC (radio) level using the "wl" command on the WRT54G radios. That's not too horrible but it's also not that good.

WDS is not clearly defined in IEEE 802.11b/g specs. I can cite chapter and verse if you're interested. WDS in it's current implimentation was thrown together by Broadcom, using their best guess as to what the IEEE was thinking. Atheros did the same with less success. That's why it only works with compatible chipsets. Please note that

formatting link
does not certify Wi-Fi devices for WDS compliance.
formatting link

See above. The interference problem has nothing to do with the WDS implimentation. It's the increased number of active transmitters on the channel in a given airspace that causes problems.

Any benchmarks?

I could write quite a bit about what home and business users really want. Initially, they all want fabulous performance and copious features. That lasts about a month or so. When I arrive, they are more than happy to dump all the goodies and garbage for a major improvement in reliability. Flakey wireless connectivity is the surest way to screw up reliability.

Read what I wrote. I said that they are "somewhat more sensitive..." not that they would totally fail to operate.

I did. That's why I included the URL below so she could see what she would be dealing with.

Any more questions or comments?

Reply to
Jeff Liebermann



May or may not - but not giving them the option to make an informed opinion is wrong. Neither you nor I know what he has.Why do you make the assumption he shouldn't be infomed as neither of us know ?

Here in the UK, you can easily get routers that cost roughly 30 quid which support WDS. If he can't afford that he's free to come back and say so - it's not for you or I to tell him what is/isn't right for him. That's *HIS* choice. Yeah it does cost more but can give you more flexibility - just depends how you want to look at it.

True - but if you're being rate limited by your ISP, that *MAY* be irrelevant. e.g. 2 * 54Mbps routers - cut connection rate in half- if you're still only being fed data by your ISP at 10Mbps ( which is the max NTL here support/will support in the perceivable future) even if the rate is being halved by WDS, that will be irrelevant. What do you mean causes trouble with client to client file transfers ?

True again - but usually not to the extent that it interfers with much else. WDS is an IETF/IEEE derived standard. Why should it cause any more problems than normal radio broadcast signals, assuming the Vendors have implemented it correctly ?

I've used WDS succesfully in the past, and am currently doing so with no problems at all (based on what I need it to do).

I think we're going over the score here - the scenario was for a home user. Most home users won't give a monkey's chuff about whether the TCP/IP stack has to do resends, so long as performance isn't impaired to an unnacceptable level..If this was a business scenario then I'd agree with you.

Yeah and ??? Does that mean WDS doesn't work then ??

Fair point.

Hmm.. Not too bad if you research it enough.

Reply to
Steve Berry

Is the connection in your daughter's PC currently a USB adapter, or a PCI card?

With a USB adapter, and a long cable, you should be able to position the access point for a better signal than you get with the antenna from a PCI card. You could also try a directional antenna if the USB adapter has a connection for one (some do, many do not).

Reply to

Because that's the way they learn. Or rather, that's the way I learn and expect others to do the same. It is impossible to determine the experience level of the person asking a question on usenet or in a mailing list. I therefore tend to supply answers and information well over most peoples experience level. Most beginners will pickup about

10% of my ranting. That's sufficient for them to do some reading, thinking, and research. I try to supply complete, stand alone, detailed answers that if Googled, will supply enough information to solve whatever problem is at hand. Very few people supply enough information to actually solve their problem, so the research aspect is almost a requirement. How "valuable" is my information largely depends on the subsequent research.

What would you have preferred I posted? A one line opinion of she should do consisting of "I think WDS sucks. Get a better antenna"? Is that what you're suggesting?

Single Mom's are capable of learning. I don't insult anyone's intelligence by suggesting otherwise. Incidentally, I will confess that I often write as if I'm lecturing to an audience instead of supplying a personalized answer. If you've been in a position of answering questions from an audience, you'll find that directing the answer to the entire audience, instead of asking them to listen to a personal conversation, is far more effective.

You should read some of my other postings. The "tone" is more insulting, humiliating, brutal, arrogant, and obnoxious than my previous answer. Want tech answers without supplying enough info? Fine, be prepared for my wrath.

In this case, no "tone" is required to understand that I don't like WDS and mesh networks and that methinks a better antenna or location would be better. I supply the information. The person asking the question can supply the conclusions. If she wants to use WDS despite the itemized limitations, then at least she has an understanding of what might be involved, and what to look for when things go wrong.

Note that I did not address the answer to you, but to the original question. If I had addressed it directly to your one line comment, I would have started with something like "How do you know that she can use WDS when you don't have a clue as to what equipment she currently owns".

I'm disappointed. Confuse them with facts? Want me to list a few advantages to WDS? I don't mind.

Would you have preferred a non-confusing, non-detailed, one line pontification instead? I have no clue as to your experience level, therefore I should aim low and presume minimal competence? Don't you think that the detail, despite some of it going over your head, might be a bit more useful than a one line conclusion?

In the colonies, we don't have royalty and don't do that.

I think you'll be surprised. A single WDS repeater actually works fairly well if the connecting laptops can only see the repeater and not the main access point. However, if they are setup in a way where everyone sees everyone else, collisions will become epidemic. Even if you turn on CTS/RTS flow control on the access points, there will still be collisions because the access points and clients transmit asynchronously.

A really simple clue can be obtained with just ping. I suggest using "fping" instead of Microsoft ping. Just ping the DSL gateway IP from the client machine going through a WDS repeater. What you'll see is that the normal ping times vary radically depending on how many retransmissions are required to get the packet through. With WDS, a lost packet requires two retransmissions so the effect will be very obvious.

Most cheapo IP access points hide these resends. There's no internal diagnostic commands that will show multiple transmissions. However, the delays will be evident. If there were no collisions or resends, then the ping times would all be constant and fairly low (about 30msec for 1.5Mbits/sec). Also try the same test without the WDS repeater. It might still vary, but probably not as radically as with a WDS repeater.

You might want to remember that when offering advice. In general, an RF solution or improvement will offer a more stable and reliable connection than adding a repeater. I also commonly suggest adding a

2nd access point via a CAT5 cable to areas where coverage is lacking. Separating the modem/router from the wireless access point is also a good idea because it allows the wireless access point to be elevated from its traditional location under a desk to a much less RF obstructed location on perhaps a bookshelf. Can't do that will a tangle of CAT5 wires going to the switch ports. Lots of other options are available before going to a WDS repeater.

Fine. For homework, chew on this for a while and write a few paragraphs on the advantages of WDS. Quiz in about a week.

Reply to
Jeff Liebermann

Yes they do. For me though that's about informed choice/opinion. The reason I mentioned WDS is that personally I've found it more reliable/efficient/inexpensive than using bridging/ICS- taking the load off the CPU and it scales alot better. Like any technology, it does have its downsides, but you as a user evaluate those downsides against what you need it to do/it's suitability for the task.

Yes - but why is the "Bull in a China Shop" approach - bombarding users with tech information you're not sure they'll even understand beter than taking it slower and trying to ascertain what level of knowledge is actually going to be valuable to them ? If you notice, the poster was a single mum - I doubt she's ever heard of WDS, so what would be the point in hitting that user with info about Omin-directional antennas if they don't understand what the **** that is ?

Ok. True.

The "tone" of your email was "here's reasons not to use WDS" - negative but factual - not here's what you could do with WDS, which is slightly more optimistic.

I didn't specifically for the reason staed above.

Like anything, transmission rates will depend primarily on what ISP you are signed up to, but in Britain it isn't exactly fast ( not if you compare it with the likes of Sweden for example). In the UK, NTL have just recently offered a 10MB service, most users still haven't been upgraded and are currently on either 1/2/3MB connections so the argument for the use of WDS is arguably even stronger-again depending on how you want to look at it.

Ok - that's valuable info thanks - that much I didn't know.

When I get time, I'll try and get round to it.

I bow to your superior knowledge.




Reply to
Steve Berry


See the end.

All noted I did find this educational even if it wasn't how I expected it was going to be. Don't bother with next week - guess you're busy.;)

Reply to
Steve Berry

Lawdamercy! Bottom posting is good, but the whole point is to EDIT!

Reply to
Derek Broughton

Apols - was a last minute before bed thing.

Reply to
Steve Berry

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.