Newbie question: How many switches can I install 100m apart?

If you look at groups.google.com, you can read many good answers
to this question.

Basically: it is possible, but far from advisable.

--
Kind regards,
Gerard Bok

Gerard Bok wrote:


802.1d says (or did at one time) that the limit is 7 hops.

You are proposing 3 hops so are well within those limits.

So it's Ok. (and advisable too).

There seems to be the idea that this 7 hop limit may not
be such an issue with modern equipment however it's not
what the standard says.

The 7 hop limit was expected to include thinks like 64k WAN
links and CPUs are now orders of magnitude faster.

If you are running cable put in more than one.

On 21 Jan 2007 15:23:18 -0800 Bod43@hotmail.co.uk wrote:



It is basically not that much a hard limit but a recommendation, so that
the STP topology information cached in the switches does not time out in an
unreasonably short time (or even before the STP BPDU actually would have
arrived). Aside from that, it would actually require to have the 802.1d
Spanning Tree implemented *and* activated to fall under this guideline.
When using RSTP (which is far more common now) or when you know what you
are doing (and are able to set your own max_age), you will not be affected.



It actually does not have any relation to round trip times of your links or
the processing power of your switches. The trouble is, that the "message
age" of a BPDU is increased by one at each bridge. And the sum value of the
message age is subtracted from the max_age (which will be 20 seconds in
most defaults) when building topology information.

--
Denis Jedig
syneticon networks GbR             http://syneticon.net/service /

On 21 Jan 2007 15:23:18 -0800, Bod43@hotmail.co.uk wrote:



I'm far to lasy to read 802.1d, but I recall a 3-4-5 rule :-)



Read back in this newsgroup: from an ethernet point of view,
there is no problem at all.

From a 'cabling', 'safety' and 'regulatory' point of view it is
not advisable --or even prohibited-- to run such a length of
copper between 2 buildings.

--
Kind regards,
Gerard Bok

On Mon, 22 Jan 2007 00:59:06 GMT Gerard Bok wrote:



There is no such rule in any of the standards. It used to be a "safe"
guideline for nowdays rather antiquated equipment to stay within the
boundaries of maximum propagation delays prescribed by the Ethernet
standards. However, as switched networks separate collision domains, it
absolutely does not apply here.



There is nothing in the standards that restrains you from having a 90m
link. Actually, you might have a working unrepeated 150m copper link
operating over standard ethernet equipment, if you get it right. Anything
longer than 90m however would be outside of the structured cabling
standards defined in the ISO/IEC IS 11801 or TIA/EIA 568 respectively.

While it is true that interbuilding links might (and often enough will)
introduce problems with equalizing current between the two linked points,
such links may as well work stable for years, even despite of lightning
strikes.

--
Denis Jedig
syneticon networks GbR             http://syneticon.net/service /

Sure, this will work for a while until you get a nearby
lighting strike.  Then one or more stiches & possibly NICs
will get fried.

The correct way to do interbuilding is with fiber.

-- Robert

Robert Redelmeier wrote:


Quite

You just need switches with fiber uplinks.

But as its "just a couple of pc's" you might find a wireless bridge
cheaper than
having to get fiber pulled - esp if you don't own the land between.

How about two media converters at both ends instead of switches with fiber
uplinks?

as long as there is some glass between the 2 buildings you should be fine.


it may be cheaper to use copper and accept the risk that equipment gets
damaged?



Or use DSL modems if you have phone quality copper between the 2? Again -
may still get damaged.

--
Regards

stephen_hope@xyzworld.com - replace xyz with ntl