Archived from groups: comp.dcom.lans.ethernet (
More info?)
Walter Roberson wrote:
> In article <cr1spk$4k7$1@news.onet.pl>,
> Lukasz Ponski <lponski@wpk.p.lodz.pl> wrote:
> :I am looking for a norm specifying the space between computers in a
> :single room. A room (let's say 16 square meters) is supposed to have the
> :maximum number of workstations.
> :Is there any RFC about this issue?
>
> Are you talking about in ethernet terms, or in cooling terms, or in
> fire regulation terms, in terms of EMF limits, or in psychological
> terms of how closely you can pack people together without them feeling
> "crowded" ?
>
> There is no meaningful minimum distance between systems in modern
> ethernet (unlike the old 10Base2 days.) I seem to recall that there may
> still be a theoretical limit of 1023 stations per segment, but I don't
> know if anyone bothers to enforce that, especially considering that if
> you are using a modern fully-switched full-duplex topology, there are
> only ever at most 2 stations per segment.
>
> The limits to your cooling capacity depend on what you can afford. If
> you go to the bother of putting in full raised flooring and a Leibart
> unit capable of raising a small tornado, or if you go to the trouble of
> water-cooling everything, you can get a lot more units into a small
> area than if you have the relatively stagnant air conditions of a
> typical bedroom. Manufacturers often give guideance about how much
> space to allow around a system, based upon the assumption that the
> device will be used in its typical environment (e.g., a home PC in a
> convection-cooled home, a business PC in an air-conditioned office
> designed to keep -people- comfortable.) The space recommendations for
> each model differ. If you have a high-end graphics board or a 3 GHz
> CPU, then you are going to need more cooling space than if you have an
> older generation of equipment. A CRT usually needs more cooling than a
> flat panel display.
>
> Fire regulations vary greatly from area to area, and the most directly
> applicable fire regulations might prove to be those governing
> electrical circuit densities; that is going to depend upon the local
> by-laws and upon the kind of wiring used. Household-quality aluminium
> wiring isn't going to stand up to as much heat as industry-quality
> copper wiring.
>
> EMF regulations... you'd have to look at the EU directives for that.
> Canada [my location] has no regulation corresponding to the EU limits
> on low-level EMF. The standard that we use at work is that we only
> allow authorized/trained people into past the 5 Gauss line. That limit
> isn't because of the effect of magnetic fields on the human body: the
> limit is for safety, providing a generous margin to be sure that
> nothing the person happens to be carrying is dragged in to the magnetic
> field. It is my understanding that the EU does have limits that
> are grounded in concerns about the possible effects of low-level EMF
> on the human body.
And then there is the question of what constitutes a "workstation". If
you're talking CPUs then you can get a rather large number of blade servers
in a single rack. If you're talking terminal or thin client then it's
really more a matter of how many chairs and tables you can squeeze
together. If you're talking standalone PC then it may depend on what
specific hardware you're using--you should be able to get more workers with
laptops in a room than you can with dual-processor engineering workstations
with multiple displays for example.
--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)