Prev: High capacity access point
Next: WPA on Quickertek(Ralink) card on Mac Powerbook not working
From: seaweedsl on 8 Apr 2008 17:14
On Apr 7, 3:15 pm, msg <msg@_cybertheque.org_> wrote:

> 802.11b has distinct coverage advantages, and often greater channel
> capacity in consumer-grade access points.

Greater channer capacity? In that B can handle more clients per
channel? That seems to be the opposite of what I hear. I admit
that I'm just parrotting what I've learned from the discussions on
this forum, so if there's more to it, then please elaborate.

As far as coverage advantages - are you referring to the greater
transmitting power listed for B? It seems like Jeff disabused me of
that interpretation once already but I forgot the details.


No one should seriously
> rely on embedded security in those products, WEP or WPA; a VPN
> (IPsec for example) solution makes more sense and scales better
> in a public (campus/dorm) setting.
>

What are the security problems in WPA beyond cracking weak
passwords ? I have not heard...

WEP, on the other hand is now well known to be crackable. You are
lumping them both together?

Perhaps what you are saying is that if one is connected to a "hostile"
LAN like a coffee shop, then wireless security does not protect one
from threats within that LAN? And in such cases, VPN to a non-
hostile (if you have one) LAN is best?

That makes sense. Sort of like saying that a deadbolt will not
protect you from strangers already in your house. But that's not to
say that there's no reason to lock your doors ! Or that we shouldn't
use a good deadbolt instead of an easily opened door latch. Or that
administrators don't prefer to control who is allowed in the LAN.

Appreciate any clarification. Since I'm responsible for the security
of a small group sharing a connection, I want to know if there's
something further I should be doing besides using strong pw WPA.
Everyone within is trustable - a friendly LAN, if you will.

Steve
From: msg on 9 Apr 2008 00:18
seaweedsl wrote:
> On Apr 7, 3:15 pm, msg <msg@_cybertheque.org_> wrote:
>
>
>>802.11b has distinct coverage advantages, and often greater channel
>>capacity in consumer-grade access points.
>
>
> Greater channer capacity? In that B can handle more clients per
> channel? That seems to be the opposite of what I hear. I admit
> that I'm just parrotting what I've learned from the discussions on
> this forum, so if there's more to it, then please elaborate.

I just meant that for a given LAN PHY bandwidth, more 802.11b
connections could be aggregated at 11 Mbps than at 54 MBps
(unrelated to 'client' connections), assuming max rates in
both formats. Capacity is obviously dependent on available
resources in the access point/router such as available ram
for buffers and tables and processing speed of the cpu.

> As far as coverage advantages - are you referring to the greater
> transmitting power listed for B? It seems like Jeff disabused me of
> that interpretation once already but I forgot the details.

Jeff says that OFDM is more robust, and perhaps the shorter
packets at higher speeds suffer less from destructive multipath
interference, but I have memory of analyses that show 802.11b to be
generally more robust in open-air settings. He also says that
slower OFDM speeds compared against equivalent DSSS speeds provide
superior channel capacity; I have no experience to evaluate this.

> No one should seriously
>
>>rely on embedded security in those products, WEP or WPA; a VPN
>>(IPsec for example) solution makes more sense and scales better
>>in a public (campus/dorm) setting.
>
> What are the security problems in WPA beyond cracking weak
> passwords ? I have not heard...
>
> WEP, on the other hand is now well known to be crackable. You are
> lumping them both together?

Just a philosophy; in a setting larger than an informal group, especially
where participants are not entirely trusted, use the best security
available at the least cost -- e.g. a stand-alone router based on
a secure o/s that can be configured and sized to the requirements of
the application. I use OpenBSD, IPsec VPNs and run the access points wide
open.

>
> Perhaps what you are saying is that if one is connected to a "hostile"
> LAN like a coffee shop, then wireless security does not protect one
> from threats within that LAN? And in such cases, VPN to a non-
> hostile (if you have one) LAN is best?
>
> That makes sense. Sort of like saying that a deadbolt will not
> protect you from strangers already in your house. But that's not to
> say that there's no reason to lock your doors ! Or that we shouldn't
> use a good deadbolt instead of an easily opened door latch. Or that
> administrators don't prefer to control who is allowed in the LAN.

Indeed, all of the above.

Regards,

Michael
From: seaweedsl on 9 Apr 2008 13:47
On Apr 8, 11:18 pm, msg <msg@_cybertheque.org_> wrote:

> I use OpenBSD, IPsec VPNs and run the access points wide
> open.
>
>
> Michael

Thanks for the clarifications, especially on your own approach.

I conclude that for our system as well as most homes and small
offices, G with WPA is still most efficient (vs B) and appropriate
path to wireless security.

Most of us just don't have time to go into alternative OSs etc, but
it's good to know what the serious techs are doing !

Cheers,
Steve
From: Jeff Liebermann on 9 Apr 2008 14:28
On Tue, 08 Apr 2008 23:18:31 -0500, msg <msg@_cybertheque.org_> wrote:

>Jeff says that OFDM is more robust, and perhaps the shorter
>packets at higher speeds suffer less from destructive multipath
>interference, but I have memory of analyses that show 802.11b to be
>generally more robust in open-air settings.

MatLab?

Shorter packet do have a higher probability of delivery in the
presense of fixed rate interference. That's also the reason for
packet fragmentation, which splits large packets into smaller pieces
so that the chances of getting clobbered by interfence is less. Of
course, with smaller packets, the packet overhead is increased,
resulting in a loss in thruput.
<http://www.wi-fiplanet.com/tutorials/article.php/1468331>

The problem with 802.11b modulation methods is that an interfering RF
blast at any inband frequency, during transmission, is fatal to the
entire packet. The data is trashed and needs to be resent. However,
802.11g OFDM consists of 52(?) carriers, each of which carry part of
the data. If one carrier gets trashed by interference or frequency
selective fading, the other carriers will still make it through and
get decoded.

>He also says that
>slower OFDM speeds compared against equivalent DSSS speeds provide
>superior channel capacity; I have no experience to evaluate this.

It's not a huge difference. With 802.11b, all management packets are
sent at 1Mbit/sec. That takes more airtime than the same 802.11g
management packets sent at 6Mbits/sec.

A very rough indication is the difference in thruput between
11Mbits/sec 802.11b versus 12Mbits/sec 802.11g. You'll be lucky to
get more than 4.5Mbits/sec thruput with 11Mbits/sec, but can easily
obtain 6Mbits/sec at 12Mbits/sec. OFDM (with 802.11b compatibility
off) has very close to 50% of raw data rate thruput. CCK is perhaps
about 40%. Not a huge difference, but noticeable.

Also see:
<http://pdos.csail.mit.edu/roofnet/doku.php?id=interesting>
These are some observations on the performance of the MIT Roofnet mesh
network in real conditions. Note that it's all 802.11b. Also note
the rather lousy "probability of delivery". I don't consider such
chronic and typical packet losses to be "robust". Similar networks
implimented using 802.11g work much better, with far less packet loss,
but over a smaller range/area.

--
Jeff Liebermann jeffl(a)cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
First  |  Prev  | 
Pages: 1 2 3
Prev: High capacity access point
Next: WPA on Quickertek(Ralink) card on Mac Powerbook not working