Hub / Switch daisy chaining limitation

What is the number of daisy chaining limitation for hubs and for switches?  Please include a supporting link as well.

Thank you
Who is Participating?
timyangccConnect With a Mentor Commented:
The 5-4-3 rules applies to the original 10Mbps CSMA/CD networks, other  interfaces have very different characteristics.  I will try to list them here.  The best source is of cause the IEEE 802.3 standard document, if you have enough time and energy to read and comprehend that pile of papers.

10Mbps CSMA/CD :  The kind of network is very rare these days.  The segments often relates to thick coaxial cable. And user stations attach to such cable by using a separate media attachment unit, which attach to the cable by drilling a hole into the cable and inserting a pin.  A later edition, often called cheapernet, consists of thinner coaxial cables daisy chained with the use of BNC connectors. If you are using this kind of network, good luck. The 5-4-3 rule may apply.

The Hub.   The hub starts with the twisted pair technology. The twisted pair cable is actually the link segment of the original Ethernet, not the user segment.  The 5-4-3 rule still apply if you map the hub to repeater, and the cable as link segments.  But you do'nt have the user segments as in the 5-4-3 rule. So you stick with 4 repeaters or hubs in a chain.

Half-duplex   The whole reason for this 5-4-3 rule is to keep the timing of the collision response. The CSMA/CD technology is often called half-duplex, meaning the traffic can not go  in different directions at the same time. The reason is because everybody is sharing the same cable, and if you receive while transmitting, it is because someone else is transmitting too, and you detect a collision and require a back-off and retransmission. The number of repeaters and cables is specified because the data from the far end of the network needs some time to travel to you, and if you detect the collision too late, the rule of the protocol breaks, and the data will collide without retransmisison will be lost.

100Mbps    The 100Mbps CSMA/CD technology is a different beast.  It works only on twisted pair cables, and no longer  works with coaxial cables. The rules are more stringent, because the timing is much tighter than 10Mbps.  So there are two kinds of repeaters, class I and II, and essentially you can hardly daisy chain these hubs.  If you have a stackable hub, then it can only connects to stations and bridges, and never to another hub.

The timing is actually based on the signal travelling time, so if you use shorter cables, these rules can relax a little. If the farthest nodes can communicate without  "Late Collision" errors, then the network topology is fine.

Full duplex. With the introduction of bridges and switches, the Ethernet protoocol went through a revolutional change.  In a point-to-point link you actually do'nt need the collision mechanism because you have only one partner and you can talk to each other at the same time if you forget about this collision mess.  So a modern full-duplex network works  in much simpler manner than its predecessors.  To conserve backward compatibility, the half-duplex network can still be used and connected, and such links follow the old rules.  In a full duplex world and fast switches, there are almost no restriction in daisy chaining.  There are other  factors like the spanning tree in a switched network, but it is another story.
Standard ethernet "hop" limit for repeating links is three (3).

(sorry, I'm in training right now else I'd have more time to provide detailed links!).

THREE is the limit.   Daisy chain any more than that, and you end up with packet loss.
I always thought the limit was 4...I'll see if I can find some links that verify one way or the other...
Protect Your Employees from Wi-Fi Threats

As Wi-Fi growth and popularity continues to climb, not everyone understands the risks that come with connecting to public Wi-Fi or even offering Wi-Fi to employees, visitors and guests. Download the resource kit to make sure your safe wherever business takes you!

Check out the "5-4-3" Rule...

I remember learning about this in a networking class way back...for some reason 4 repeaters was stuck in my head.

According to the IEEE 802.3 specifications, a network using repeaters between communicating stations (PCs) is subject to the "5-4-3" rule of repeater placement on the network:

5 segments - connected on the network
4 repeaters
3 segments - of the 5 segments can have stations connected. The other two segments must be inter-repeater link segments with no stations connected.
Carefully read the packaging also.  I had purchsed several switches that had "Desktop" on them (3com desktop switches 10/100) that only allowed a total of 4 mac address per port, was a real pain troubleshooting why the distant switch only 3 people could log on at one time, heheh

Also, I believe the 543 rule only applies to 10base-T and not 100BASE-TX ??

"A switched Ethernet network should be exempt from the 5-4-3 rule because each switch has a buffer to temporarily store data and all nodes can access a switched Ethernet LAN simultaneously"

With the price so low, I would use nothing except switches unless their was a very specific reason to use a hub.  For example to monitor incoming traffic I have a hub setup just inside my router, before my firewall, and my monitoring device is also connected to the hub so it can watch traffic flow by.
Daisy chain one connected to another is no recommended if you can avoid it.  If you must have several switches/hubs to have a 1-2 central switches and have all the connections come back to these.

-= Felix =-
Mazaraat, it's funny...that's a contradiction...the quote you pulled refers to a "switched network", yet the rule applies to a switched network, hence the "4 - switch" part of the rule.

I don't know for sure, but I think it's just "good practice" to follow the 5-4-3 rule. Depending on your specific hardware use, this rule may be bent in favor of more or less switches depending on the quality of the hardware. You are likely not going to be able to uplink 5 $30 switches and get the same kind of single at the other end than with 5 $2000 switches. I think the rule is in place because for *most* networks, this will enable a reliable data transfer.
The 5-4-3 rule applies to a single collision domain. Switches start a new collision domain on each port, so the 5-4-3 rule does not apply to ethernet. There are no bonus points for following the 5-4-3 rule when switching.

On this CCNA website I found that this might be some useful information:

"The IEEE 802.3u specification for 100BaseTX networks allows a maximum of two repeater (hub) networks and a total network diameter of approximately 200 meters.  A link segment, which is defined as a point-to-point connection between two Medium Independent Interface (MII) devices, can be up to 100 meters."

Yes. This is correct.  I was a little lazy in putting out the exact spec.
This essentially says the same thing.  If you have a stackable hub system,
a packet goes into one unit, passes the inter-repeater link, and goes out
from another unit, it essentially passes through two hubs. The spec actually tries
to keep the user link segment to consistent  100 meters. The little more
than 200 meters includes the short inter-repeater link cable.
From user's point of  view, you see a stackable hub and think it is a
single hub. You can only have every user segment 100 meters, but you
can not link to any other hub with just a link cable. Every other
network device linked to this hub has to be bridges.
The post on 10/19 was the accepted answer. Am I missing something?
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.