Cable equivalents

The literature that I can find on-line (Cisco) says that in planning a
100BaseT wired network, the maximum cable distance between devices should
not exceed 400m. If there is any repeater (switch or hub) in the path, then
this should be considered as contributing 95m to the length. But what
happens if there is a wireless link or access point in the route? How much
cable should this be considered equivalent to?

TIA
Chris

In article <dko4as$qds$(E-Mail Removed)>,
(E-Mail Removed) says...
> The literature that I can find on-line (Cisco) says that in planning a
> 100BaseT wired network, the maximum cable distance between devices should
> not exceed 400m. If there is any repeater (switch or hub) in the path, then
> this should be considered as contributing 95m to the length. But what
> happens if there is a wireless link or access point in the route? How much
> cable should this be considered equivalent to?
>
There isn't a cable equivalent - the Ethernet limitations are down to
timing issues with signal propagation on the cable, and the cable stops
where it meets the wireless device.

"Chris Watts" <(E-Mail Removed)> wrote in message
news:dko4as$qds$(E-Mail Removed)...
> The literature that I can find on-line (Cisco) says that in planning a
> 100BaseT wired network, the maximum cable distance between devices should
> not exceed 400m. If there is any repeater (switch or hub) in the path,
> then this should be considered as contributing 95m to the length.
> But what happens if there is a wireless link or access point in the
> route? How much cable should this be considered equivalent to?

The reason for the limit you mention is to do with collision detection, and
applies only to devices within a collision domain. A collision domain is a
single, half-duplex Ethernet network. Collision domains end at bridges.

A switch is not a repeater, it is a bridge. If, for example, you use two APs
to join two wired networks, each of those APs is a bridge. Thus the limit
does not apply in either of these cases.

Alex

"Alex Fraser" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed). uk...
> "Chris Watts" <(E-Mail Removed)> wrote in message
> news:dko4as$qds$(E-Mail Removed)...
> > The literature that I can find on-line (Cisco) says that in planning a
> > 100BaseT wired network, the maximum cable distance between devices
should
> > not exceed 400m. If there is any repeater (switch or hub) in the path,
> > then this should be considered as contributing 95m to the length.
> > But what happens if there is a wireless link or access point in the
> > route? How much cable should this be considered equivalent to?
>
> The reason for the limit you mention is to do with collision detection,
and
> applies only to devices within a collision domain. A collision domain is a
> single, half-duplex Ethernet network. Collision domains end at bridges.
>
> A switch is not a repeater, it is a bridge. If, for example, you use two
APs
> to join two wired networks, each of those APs is a bridge. Thus the limit
> does not apply in either of these cases.
>
> Alex
Thanks Alex. I can understand and appreciate that. Indeed I do now see
that two wireless APs will act as a bridge hence what you say must be true.
But CISCO definitely disagree with you about a switch being a bridge - they
say it is a repeater (albeit a selective one) and hence contributes a
latency equivalent to 95m of
cable - their figures.

Chris

"Chris Watts" <(E-Mail Removed)> wrote in message
news:43706179$0$63076$(E-Mail Removed)...
[snip]
> Thanks Alex. I can understand and appreciate that. Indeed I do now see
> that two wireless APs will act as a bridge hence what you say must be
> true. But CISCO definitely disagree with you about a switch being a
> bridge - they say it is a repeater (albeit a selective one) and hence
> contributes a latency equivalent to 95m of cable - their figures.

You must be misinterpreting what you've read, or what you've read is wrong.
Switches do contribute latency, but that latency is irrelevant from the
point of view of collision detection because it doesn't appear within a
collision domain.

What exactly have you read?

Alex

"Alex Fraser" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> "Chris Watts" <(E-Mail Removed)> wrote in message
> news:43706179$0$63076$(E-Mail Removed)...
> [snip]
> > Thanks Alex. I can understand and appreciate that. Indeed I do now see
> > that two wireless APs will act as a bridge hence what you say must be
> > true. But CISCO definitely disagree with you about a switch being a
> > bridge - they say it is a repeater (albeit a selective one) and hence
> > contributes a latency equivalent to 95m of cable - their figures.
>
> You must be misinterpreting what you've read, or what you've read is
wrong.
> Switches do contribute latency, but that latency is irrelevant from the
> point of view of collision detection because it doesn't appear within a
> collision domain.
>
> What exactly have you read?
>
> Alex
See 100BaseT cabling guidelines
in
http://www.cisco.com/en/US/products/..._installation_
guide_chapter09186a008007eb00

Chris

"Chris Watts" <(E-Mail Removed)> wrote in message
news:4370b619$0$82637$(E-Mail Removed)...
> "Alex Fraser" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
> > "Chris Watts" <(E-Mail Removed)> wrote in message
> > news:43706179$0$63076$(E-Mail Removed)...
[snip]
> > > But CISCO definitely disagree with you about a switch being a bridge
> > > - they say it is a repeater (albeit a selective one) and hence
> > > contributes a latency equivalent to 95m of cable - their figures.
> >
> > You must be misinterpreting what you've read, or what you've read is
> > wrong.
[snip]
> > What exactly have you read?
>
> See 100BaseT cabling guidelines
> in
>
http://www.cisco.com/en/US/products/..._installation_
> guide_chapter09186a008007eb00

For some reason (cookies probably) that link didn't work, but searching for
"100BaseT cabling guidelines" on Cisco's site resulted in the same URL but
working...

The only source of confusion I can see is that the document is talking about
modules that can be added to a Catalyst 2820. In its basic form the 2820
appears to be a 25 port 10BASE-T switch, but some of the modules are
repeaters - these are distinguished by the term "shared".

Alex