Whats the limitation of Repeater ?

Whats the limitation of Repeater ?

Post by dotun ol » Tue, 27 Feb 1996 04:00:00



Can someone pls clarrify this grey area !

What is the limitation on the number of repeaters one  can have on a 10 BAse T
network ?

I am aware that on a 10 Base 5 , with max segement size of 500m , one cannot
have more than 4 repeters so as not to exceed the 2500m limit.

In out case , we have 10 Base T , with maximun lenght less than 100m , The
repeaters are used to concentrate pc's. but we have plenrty of them all being
fed from a master hub.  

I want to be sure , no standard are being broken if any and that the
propagation delays as well as other signal losses will not be significant to
cause  future problems.

Any documentation or references will be appreciated.

Many thanks.

|--------------------[ No Muss , No Fuss ]--------------|

 Criterion Software Limited   -   A  Canon company
 Guildford, UK
 voice : (+44) 01483 406 230
 fax   : (+44) 01483 406 211
 web   : http://www.canon.co.uk/csl/cslhome.html

 
 
 

Whats the limitation of Repeater ?

Post by Joshua Bardwe » Tue, 27 Feb 1996 04:00:00


: What is the limitation on the number of repeaters one  can have on a 10
: BAse T network ?

You can never have, in either a 10BaseT, 10Base2, or 10Base5 network
  - More than 5 total network segments between any two communicating
    stations.
  - More than 4 repeating devices, where a repeating device is a repeater,
    10BaseT hub, etc...
  - More than 3 "link segments," where a link segment is a network segment
    with nodes* off of it.  A run of Thick Ethernet between two
    floors of a building would not be a link segment.

These rules are laid out in the 802.3 spec, and are collectively known
as the "5-4-3 rule".

J
--
Joshua Bardwell                        | "Always remember to never forget:

http://www.veryComputer.com/~gt6234b   |  The people you've met."
                                       |    - TCJ

 
 
 

Whats the limitation of Repeater ?

Post by Kevin Oberma » Tue, 27 Feb 1996 04:00:00



>   - More than 3 "link segments," where a link segment is a network segment
>     with nodes* off of it.  A run of Thick Ethernet between two
>     floors of a building would not be a link segment.

> These rules are laid out in the 802.3 spec, and are collectively known
> as the "5-4-3 rule".

This error gets posted so regularly that I assume some popular test out
there is propagating it! In this case, the statement is unusually
mangled to the point I am not sure what it means, since the definition
of "link segment" is backwards.

To quote 802.3(1993), 13.3,
"When the network path consists of four repeater sets and five segments,
up to three of the segments may be coaxial"

Note that there is no statement on whether there are nodes on a segment
since it doesn't matter. This is a timing issue and access times for
coax are longer than for link segments. 10Base-T and 10Base-F qualify as
link segments.
--
R. Kevin Oberman
Energy Sciences Network (ESnet)
National Energy Research Supercomputer Center (NERSC)

 
 
 

Whats the limitation of Repeater ?

Post by Ken Wa » Wed, 28 Feb 1996 04:00:00



>>   - More than 3 "link segments," where a link segment is a network segment
>>     with nodes* off of it.  A run of Thick Ethernet between two
>>     floors of a building would not be a link segment.

>> These rules are laid out in the 802.3 spec, and are collectively known
>> as the "5-4-3 rule".

>This error gets posted so regularly that I assume some popular test out
>there is propagating it! In this case, the statement is unusually
>mangled to the point I am not sure what it means, since the definition
>of "link segment" is backwards.

>To quote 802.3(1993), 13.3,
>"When the network path consists of four repeater sets and five segments,
>up to three of the segments may be coaxial"

>Note that there is no statement on whether there are nodes on a segment
>since it doesn't matter. This is a timing issue and access times for
>coax are longer than for link segments. 10Base-T and 10Base-F qualify as
>link segments.
>--
>R. Kevin Oberman
>Energy Sciences Network (ESnet)
>National Energy Research Supercomputer Center (NERSC)


   The 5-4-3 rule is a good rule of thumb, but when mixing media types, and
in complicated network topologies, there is a "new" way to calculate valid
network topologies (I believe recently (a year or so ago) in IEEE and
published their earlier by System Architecture Group or SYSTAG, I don't
recall the exact document name off hand, maybe someone else can help out
on that).   Any way...

    The two variables that matter in a network are its Path Delay (PD) and
its Path Delay Variability (PDV).  Every segment type (base5, 2, T, FL, FB),
as a PD associated with it, for each of 3 connection types, "Right", "Middle",
and "Left".  "Right" and "Left" nodes are leafs in a topology, "Right " nodes
are sources of a packet, "Left" are destinations. "Middle" nodes are
intermediate nodes.  There are also Delays for Media distance of each segment.

  The PD's where determined by including a nodes  Carrier Sense, Carrier
propigation, Collision Sense and Collision Propigation Delays,  in both
directions in a network.  So to determine the round Trip Delays of a network,
you only have to calculate the PD's from All "Right" (source) nodes to all
"Left" (destination) nodes. Note that every Leaf node in a network is both
a source and destination.

  So you calculate all the PD's in the network from every source to destination
, and if that is less that 512 bit times, the network diameter is OK.

  Each Node also has a PDV, which relates to how much InterPacket Gap (IPG)
is lost going through a node.  IPG shrinkage can be at most 32 bit times, so
you add up the PDV's of each node in all source to destination pairs, and if
that is less than 32, the PDV of the network is OK.

  If both the PD and PDV are OK, you have a legal Network.

  In the IEEE paper,  there is a table with all PD's (right, middle, left) and
PDV's for all 10base media types.  I'll try and dig up what I have and post.
Also I wrote a C program to do this, but its not very user friendly, but I
guess I could post it.  The interface is very primitive, it prompts for nodes
and their connections (media type and distance) to other nodes, to create the
topology, then calculates the PD's and PDV's.

  What you do find, however, is that for all none 10base-FB links, the 4
repeater rule still holds (the PDV of a non-FB repeater link is 7).  If
there are FB connections, there can be many more repeaters (the PDV of an
FB link is only 2).
---

--------------------------------------------------------------------
Ken Ward                                            ascom-Nexion

508-266-2327                                        Acton, MA  01720
--------------------------------------------------------------------

 
 
 

1. Repeaters, repeaters, repeaters ???

Hi there,

Just a question about repeaters in the copper telephone network.

Where exactly is one located?

Do they exist in the distribution network? (the network from the local
exchange/central office, to cross connection point, to distribution
point, then to customer's premise)

Or do they exist ONLY in the 'primary access network'? (Is this term
correct for the networks which connects within exchanges, also special
connections for customers of T1/E1 services?)

Repeaters are only needed for T1/E1 services, but not the ordinary
telephone (from customer's premise to exchanges), is this correct?

Thanks for your answers,

L S Ng

2. Accessing workstation through Linksys

3. Q:Whats the purpose of Inter Repeater Links?

4. Re : you @ 100mobile.net 839

5. Repeater Limitations

6. DHCP not working when 2nd NIC attached to hub!

7. Physical Limitation Vs Operation Limitation

8. Updating table based upon matching field in second table

9. Whats up on Whats Up Gold ?

10. FS: DEC repeaters, hubs, bridges, fiber, 15x ThinWire 8 port BNC repeaters!

11. Max Repeaters/multiport repeater wiring questions

12. What are the limitation of the Cisco 3845?

13. C3845, Dual Hub Dual DMVPN Hub-To-Spoke, Limitations?