Im getting confused between subnets and segments

x_Danny_x

can someone tell me the differences, please. I having problems understanding them

RussS

Subnet - each independant network on a TCP/IP network is known as a subnet.

Segment - the bus cable that the computers are attached to. This can be the length in a network, or part or whole of a line connecting 2 or more networks.

Therefore we can say that subnet is a description used when referring to a network (or part of one) and segment is used when referring to the cabling.

Webmaster

A network divided in to smaller parts based on Layer 3 information is a subnet (i.e. a subnet has a separate IP range, is separated from other subnets by a layer 3 networking device, i.e. a router). Parts of a network based on Layer 2 information is typically called a segment. But the term segment can be used to denote a part of a network or WAN or LAN, etc, etc, as well... in other words, the word segment is often used as the English word segment (not as if it is IT terminology/jargon).

To confuse you even more a subnet can consist of multiple segments and a segment can consist of multiple subnets.

Each subnet is typically one broadcast domain and each (Layer 2-)segment is typically one colission domain.

RussS

Dang - and I thought I was making this a little simpler for Danny

Webmaster

RussS wrote:

Dang - and I thought I was making this a little simpler for Danny

Yeah, I did go Enforcer-style on him there, didn't I..

Well, I think your explanation is solid for the Network+ exam. As I tried to mention the word segment is used often not as a 'stand-alone' term, but rather in combination, such as your example a cable segment, or a WAN segment, or a LAN segment. Therefor you won't see an exam question asking what a segment is (while you might be asked what a subnet is). But, and I'm not trying to confuse anyone here, something like the following is possible:

You are the admin of a large LAN. You have noticed the LAN is often saturated with broadcast traffic. You want to segment your LAN to create multiple broadcast domains. Which of the following devices should you use?

a. bridge
b. hub
c. router
d. WAP

(answer is router. A switch with VLANs could also be used instead.)

Let's give this another try: as segment is a physical part of a network, a subnet is a logical part based on addressing information

One long road (physical segment) can have multiple street names (subnets). A single street name can cover multiple roads (I know, the latter is not common, I've seen it though )

henkk

Hi,

Nice to see this topic on subnetting. Something that seems to get everyone confused. Now, I've known for a while now HOW to do subnetting, but I have to humbly admit that I don't really know why.

So far, I think I've done it to break up broadcast domains, but regarding what's been said above, I realise I've always thought of a broadcast domain as being pretty much the same as a collision domain.

Please enlighten me! What's the difference, and WHY would I want to subnet, and if the answer is "to break up a broadcast domain" then WHY would we want to do that?

Henk

p.s. I hope I'm not the only subnetting ignoramus out there!

x_Danny_x

Man my head hurts for trying to understand all of these posts

I got the subnet down and a little bit of the segment because of what was mention dealing with the physical cable. Though a little confused about the WAN segment and LAN segment,

What about the 5-4-3 rule of the repeater? Does this describe the cable as being the segment, that is what I thought of after reading the notes here




@Webmaster, you said that a subnet is a Broadcast Domain?

Yes a little info then on what is a Broadcast Domain and Collision Domain!? I would like to know too as the poster named Henkk

Webmaster

Subnetting is one of the more difficult topics indeed (at least until you master it), but this topic is not specifically about subnetting. 'Subnetting' refers to calculating the optimal subnet mask for a given situation, playing with bits, dividing a (sub-)network in to small subnetworks, making optimal use of the IP address space, and changing classful into classles. Subnetting, in this context, is not part of the Network+ exams. However, for the CCNA exam and other Cisco and some Microsoft exams you must master the art of subnetting.

Please enlighten me! What's the difference, and WHY would I want to subnet, and if the answer is "to break up a broadcast domain" then WHY would we want to do that?

Understanding the difference between a broadcast domain and colission domain is required to understand what a certain network device does. For that reason, I included it as the first two paragraphs of my Network Components TechNotes:
www.techexams.net/technotes/networkplus/networkcomponents.shtml

It also shows exactly how a device 'works with' each of the two domain types, in full color

The reason you want to separate a large network into multiple broadcast domains is mentioned in the example multiple choice question I posted above. After reading those TechNotes you will probably understand why: to reduce colissions and broadcast traffic. Remember that a network like Ethernet is a broadcast network (has nothing to do with Layer 3 broadcasts to an IP broadcast address such as 192.168.1.255), this means that every node in the network hears all traffic, also traffic between two other nodes. Just like a lot of people in a single room, all shouting to communicate with someone else. Only the nodes that the traffic is addressed to will pick it up from the cable. If there are too many nodes on the network there will be an excessive amount of 'broadcasted traffic' (not just broadcasts destined for the Layer2 broadcast MAC address FF:FF:FF:FF:FF:FF). So if you put those people in two different rooms, two people shouting to eachother are more likely to hear eachother. If someone in one room wants to talk with someone in the other room, there will have to be some sort of intermediate between the room.

If the intermediate standing in the doorway between the two rooms is a repeater it will just pick up all the shouting from one room and shout it out in the other room (and vice versa).

If it were a bridge, it would forward the shouting unless it knows that the destination is someone in the same room as the shouting came from (based on the MAC address of the people/nodes).

If it were a router it would not forward the shouting at all, it would demand the people to write down their messages and put it in an envelope (ie. Layer 3/IP packet) and address it (with the destination IP address). That destination address is made up of the room number and the persons personal number.

I hope this will actually be of help

Webmaster

x_Danny_x wrote:

]@Webmaster, you said that a subnet is a Broadcast Domain?

I said 'typically' the terms are not interchangable or anything. As you probably know, routers route traffic between IP networks or IP subnets. Router also separate broadcast domains. I.e. if a router interconnects two ethernet networks, each network is its own broadcast domain, cause it is separated by the router which does not forward layer 2 broadcast traffic (ignore to avoid confusion: layer 3 neither if securely configured else its vulnerable to Smurf :smurf: attacks), each network connected to the router is also a different subnet.

x_Danny_x wrote:

What about the 5-4-3 rule of the repeater? Does this describe the cable as being the segment, that is what I thought of after reading the notes here

Yes, that it correct. Each port on the repeater is connected to a single segment. That doesn't mean that the amount of ports equals the amount of segments as you can see in the following example:
[repeater]port

---

segmentA

---

port[repeater]

henkk

Johan, you mention in your (brilliant) technotes that: "Broadcast frames are frames explicitly directed to all nodes on the LAN."

Would an ARP cache refresh be an example of such broadcast frame?

(a bit on ARP, as far as I understand it. A source host needs to find out the physical address of a destination host. It then "broadcasts" a request to every other host on the network with the destination IP address included in the request. Every other computer looks at this request, and if its IP address is the same as the one in the request, it sends a reply giving its MAC address. The source host then updates its ARP cache with the new MAC address next to that IP address)

What would be other examples of a broadcast frame?

Henk

Webmaster

Yes, ARP requests are a very good example of Layer 2 broadcasts.

Thanks for the word about the TechNotes. Though after talking about this so much, I keep seeing room for improvement...

What would be other examples of a broadcast frame?

Other frames that have FF:FF:FF:FF:FF:FF as the destination MAC address, examples are:
- DHCP (I.e. discover)
- RIP broadcasts
- NETBIOS broadcasts
- Some novell protocols.

As you probably noticed it's mostly used for sending out probes, ie. to find a MAC address, DHCP server, Netbios name/IPaddres combi, etc. etc. or, such as RIP, it is used to send information to anyone on who can hear it and wants to pick it up.

I hope this helps!

x_Danny_x

Webmaster, So basically this is what the difference between a segment and a subnet

The difference between a segment and a subnet is as follows: a segment is a physical network construction, whereas a subnet is simply a higher-level software configuration.

In a segment the computers are physically grouped together by cable while subnets are defined with software configuration

Please say Im right

Webmaster

You are right

Layer 3 addresses, which define the boundaries of a subnet, are indeed assigned in software.