5-4-3 rule - worth remembering?

Talic

My Mike Meyers N+ is confusing the crap out of me with the 5-4-3 rule. The more I try to memorize it the more annoying it seems to get. Does it show up on the test? What about any of the Cisco testing? I remember reading in the book that layer 2 switches pretty much throws the rule out but I'm worry about CompTIA's annoying habit with testing old concepts.

I think its stupid how they put a computer in between repeaters and using them to connect hubs. Talk about a huge breach of fault tolerance. The rest of the book is ok, just him trying to explain that particular subject annoys me.

Now back to learning subnetmasking... bit sharing... bleh.

freetech

It didn't show up on the CCNA test, per se, and I don't think that's in any of the subject domains, but...
I have seen this problem bring a network to its knees in real life. Unfortunately it does happen and I would be prepared for it on the test.
The reason that it is important is that any host has to be able to send out a packet and sense a collision BEFORE another packet gets sent out. If you have too may devices-repeaters-hubs, repeating out packets, in line, in sequence, it produces enough lag time to make that impossible. It really kills network reliability.
Hope that somewhat simplistic explanation helps.

bighornsheep

It's a good rule to remember, for those of you who are wondering what the rule is: 5 network segments are connected by 4 repeaters whereby no more than 3 segments can be populated with node devices.

I can't say I've ever encountered a situation related to this rule or have heard of stories that involve this rule because most if not all ethernet networks now adays are fully switched.

However, the concept is worth noting, which is attenuation or the concept of "signal loss", you can't expect 100% of your packets to travel end-to-end with no loss.

Plantwiz

Talic wrote:

My Mike Meyers N+ is confusing the crap out of me with the 5-4-3 rule. The more I try to memorize it the more annoying it seems to get. Does it show up on the test? .....

It can, but more importantly you may run into it in a live environement and you'll wish you invested the time to understand it BEFORE you put some exam letters behind your name.

Good comments above.
Just to add some additional links for your review:

http://www.webopedia.com/TERM/5/5_4_3_rule.html

http://compnetworking.about.com/od/networkdesign/l/blfaq018.htm

http://compnetworking.about.com/b/2004/02/26/the-5-4-3-2-1-rule.htm

http://www.windowsnetworking.com/articles_tutorials/netrule.html

http://fcit.usf.edu/network/chap5/chap5.htm

Talic

Oh, I think I got it now. They are saying that they don't want you to daisy chain segments together and going crazy with the repeaters. Having all the hubs running to the backbone is ok because they only go through their segment, the backbone, and the segment leading to the node they are talking to. The rule forbids having a populated segment hooked into another populated segment and expecting it to be networked into the backbone. I can see why they were worried about it back then.

Only thing is, when switches start to come into play does it completely throw out the rule? What if they use hubs plugging into the switch's port that separates collision domain? Say, if some company gets cheap and they run out of ports on a switch and they decide to take a regular hub and connect that to the switch which leads to the backbone. I would assume this would still violate the 5-4-3 rule because thats 4 populated segments the signal would have to go through. The hub, the switch, the backbone, and receiving switch.

freetech

Switches cancel out the 5-4-3 rule.
Because each port on a switch defines its own collision domain and also because a duplexed port cannot, by definition, have a collision.

bighornsheep

freetech wrote:

because a duplexed port cannot, by definition, have a collision.

KEYWORD is "by definition"........

paulbarron

You might like to consider a different version of the rule: 5 - 4 -3 -2 -1
5 Segments
4 Repeaters
3 Occupied segments
2 Unoccupied segments
1 Collision domain

Slowhand

paulbarron wrote:

You might like to consider a different version of the rule: 5 - 4 -3 -2 -1
5 Segments
4 Repeaters
3 Occupied segments
2 Unoccupied segments
1 Collision domain

Not a bad way to look at it, at all. Welcome to the forum, paulbarron, and thanks for the helpful mnemonic twist to an old device.

Goldmember

Who uses repeaters nowadays???

Hubs were great 8 years ago when I first learned about them.

In most modern networks switches are just as inexpensive as hubs and get rid of collision domains.

Here is the new 5 -4 -3 rule

If you have and extra 5 bucks, get rid of 4 hubs, and get 3 switches...

Slowhand

Talic wrote:

Now back to learning subnetmasking... bit sharing... bleh.

Heh, subnetting is going to be the backbone of your continuing network-studies. Just wait until you get to CCNA, you'll be learning to subnet and do binary calculations in your head.

Talic

Slowhand wrote:

Talic wrote:

Now back to learning subnetmasking... bit sharing... bleh.

Heh, subnetting is going to be the backbone of your continuing network-studies. Just wait until you get to CCNA, you'll be learning to subnet and do binary calculations in your head.

Only thing is I'm thinking about focusing on *Nix rather then Cisco; so I might not be able to put it to much use. Although, I might take some Cisco classes at school just for the sake of credits. Not much to choose from when you're trying to get a degree in IT.

Slowhand

Ah, okay. Still, the networking skills you're learning now, as well as the ones you'll learn in the Cisco classes, will carry over into your Unix/Linux studies. Subnetting, routing protocols, packet-switching, the OSI model, ACLs, and a whole lot more are universal. You'll see a lot of it as you delve deeper into *NIX networking concepts, especially when working at the kernel-level.

Talic

Thats good to hear. I was just worried about taking the classes and having to remember more commands to configure switches which wouldn't help me with anything else.

salem21

Is there any word definition for the one large collision domain or meaning?

MrXpert

Talic wrote: »

My Mike Meyers N+ is confusing the crap out of me with the 5-4-3 rule. The more I try to memorize it the more annoying it seems to get. Does it show up on the test? What about any of the Cisco testing? I remember reading in the book that layer 2 switches pretty much throws the rule out but I'm worry about CompTIA's annoying habit with testing old concepts.

I think its stupid how they put a computer in between repeaters and using them to connect hubs. Talk about a huge breach of fault tolerance. The rest of the book is ok, just him trying to explain that particular subject annoys me.

Now back to learning subnetmasking... bit sharing... bleh.

hey I do like your avatar. That's Brent Spiner from TNG
love the blue screen effect.