Subnetting

I've seen a lot of good information on this web site since I first started reading the forums and subjects listed under the top menu. It's been a while since I visited this subject. After reading the last few posts I had to go read the sub-netting by Chris. It's good but I suggest you take another look at some of the new CCNA curriculum (V3.x) they do a better job of discussing sub-netting. I had a hard time understanding this myself until I got involved in the CCNP.

Classless involves VLSM (Variable length Subnet Masks) Classful routing protocols (RIP IGRP) do not send subnet mask information in their routing updates and therefore cannot support VLSM. When sub-netting in a classful manner (all subnets have to have the same subnet mask) then the routing protocol assumes a subnet to be the same as the subnet mask on the interface of the router. Classless routing protocols (OSPF, EIGRP, RIP version 2, IS-IS) do send subnet mask information with their routing update information and support VLSM (also known as sub-netting subnets) as well as CIDR (Classless InterDomain Routing), which is the aggregation of network addresses.

I'm sorry but after getting past a few minor mistakes I just could not bring myself to read much past your wrong definition of classless IP addresses. I began to just scan the material and noticed an obvious error in the table where your subnet mask of 255.255.255.254 indicated 1 network. You may argue technically this is correct but you have NO usable IP addresses so the fact that there is a network number and a broadcast address is irrelevant.

You do make an excellent point in saying that you must look at the binary to truly understand what is happening with sub-netting IP addresses. A subnet mask is always all one’s then all zero’s the one’s tell the router which part is the network address. Router’s do not need to know host addresses just networks as they route packets to networks not hosts. Once a packet arrives on a network the host will pick up any packets that belong to them and all other hosts will ignore them. This is true for networks that are using subnetted addresses as well.

When sub-netting all you are doing is adding to the network portion of the address. Of course this takes away (borrows) from the host portion. I find a basic explanation of how a standard IP address works before getting into sub-netting often helps one understand sub-netting. It also helps if you have the opportunity to examine a routing table to emphasize the point that routes do not deal with “usable IP addresses”. They only deal with network addresses.

For example the class C network address of 192.168.1.0 is never configured on an interface or a host. It is said to be the “wire” or “network address”. With this standard class C network address you can configure an interface or host with any IP address from 192.168.1.1 to 192.168.1.254 (giving you 254 usable addresses). Packets with these addresses will be used only by the hosts that have them for individual conversations. IP address 192.168.1.255 is used to talk to all hosts on the network 192.168.1.0 at the same time and is called the broadcast address. All networks have to have a network number and a broadcast address. The range of IP addresses in between the network number and broadcast address are “usable IP addresses” (those configured on interfaces or hosts).

Typically the lowest usable IP address is configured on a router’s Ethernet interface and is the default gateway for all the computers/hosts on that network (LAN). This is getting to be much to long to keep the attention of readers like me so it’s time to quit for now.

Sure would like to see and update of Chris’s sub-netting if he gets a chance to do a little more research I think he could iron out a few misconceptions and develop an excellent topic.

Mew,

Let me thank you for your constructive criticism as we always look forward to our members comments.

After reading your post, I felt that it was necessary for me to reply to it in order to clarify on a few points regarding the Subnetting topic.

Subnetting as you would appreciate is not an easy topic for someone who has just been introduced to it. It requires a special approach and much though was put into it, before attempting to analyse it. The goal undertaken was not to simply “cover” the topic as most sites in the internet do, but to help the reader truly understand what he is reading and answer all possible questions which he might have while he reads the material.

The 255.255.255.254 subnet mask as you so correctly noted is invalid when looking it from a practical point of view as there is no way you can use it. However, when you look at the theoretical side of things, it is valid, regardless if its useable or not. As for the 255.255.255.255 subnetmask – things here are slightly different as its correct in both practical and theoretical terms!

While in a normal network, this mask is only used to identify a broadcast, in remote dialup networks (RAS services) it is the subnetmask usually assigned to a client! If you ever checked your IP address when dialing into an ISP or a remote RAS server, you will notice that your subnetmask is 255.255.255.255!

Because these two “special” cases were correctly added into the table and will most probably go unnoticed by most readers, I marked them with a star (*) but forgot to make a note. Any experienced professional would certainly understand why they are marked and some things are better not explained to new users, to avoid confusion.

You could argue that it might have been better left out, and I’d say that our purpose is to cover everything in as much detail as possible without confusing the reader.

Now, coming to the next point your making – Classless and Classful IP Addresses. I totally agree that these terms are better explained alongside with VLSM, CIDR and routing protocols but our point here is not that. When introducing a new user to IP Addressing and you want him to understand the Classless and Classful “Concept” – you don’t bombard him with VLSM and CIDR cause he is simply going to run away, and I assure you that I’m speaking from my CCNA tutoring experience!

The concept is correctly explained – Classful refers to all subnets having the same default subnet mask (which depends on the IP Range used) while Classless does not. Once you understand it, you can then easily take it and apply it to routing protocols without much confusion, and I’m sure anyone will agree with that.

Closing, if you still consider the approach to be incorrect, or believe that the information is wrong, I’ll be more than happy to further clarify on the topic.

Our carefully designed methods analyzing these topics have earned us much accreditation from major sites, magazines and Cisco themselves – something you won’t find for other network related sites – ! Despite that, we always love to hear comments and criticism from our friends here to help us improve every page we can on this site.

Best Regards,

I concur about the 255.255.255.255 mask.. its what I always get with dial-up here (Chris if you remember I had even asked you about it a long time ago).. it still confuses me.. doesn't it give them an awful lot of routing to do ? And what are the benefits....

Maybe it's because when I first took the academy classes I found the “theoretical” vs. actual approach confusing. Or, maybe it was just my instructor. I avoid that approach myself when teaching subnetting. I find that once the concept that any group of addresses has to have a network number a broadcast address and that the usable IP addresses are the range in between is understood subnetting is easier to teach. This is why I do not expound upon the special cases until later either.

I had to re-read my post because I did not recall discussing the subnet mask 255.255.255.255. It’s been a long time since I used dial up for may service. Cable modems are great. I’m sure you are right. This is the recommended mask for loopback interfaces on routers as well. In terms of subnet mask ranging from 255.0.0.0 to 255.255.255.255 I have never seen a use for 255.255.255.254. If you know of an example of where this would be used please share it.

As to the real difference of opinion that we have and that is your definition of Classless and Classful IP address I want to do some research and I will get back to you. I really like the examples you give especially with the use of binary ones. The way you demonstrate the use of the mask is excellent. I don’t want my strict obsession with vocabulary definitions to distract from that.

After many discussions with other students and other instructors (than my first) my understanding of subnetting greatly improved. My initial frustrations with subnetting actually became a motivation for me to help others avoid such frustrations if possible.

The terms classful and classless refer to the way routing protocols deal with IP addresses not the IP addresses themselves. You can subnet a classful IP address in a classful manner and have classful routing protocols work with them efficiently. On the other hand if you subnet in a classless manner classful routing protocols will not be able to function properly.

So for Chris to say in reference to borrowing three bits to create a subnet “…On the other hand, the use of an IP Address with a subnet mask other than the default results in the standard Host bits (the Bits used to indentify the HOST ID) being divided in to two parts: a Subnet ID and Host ID. These type of IP Addresses are called Classless IP Addresses.” is not correct.

To be classless you have to extend this to VLSM (Variable Length Subnet Masks). Only when you use more than one subnet mask in a network does it become classless. If you use only one subnet mask then this is still a classful address and the classful routing protocols will work well with these subnets.

When you subnet in a classful manner you will still have three parts to the IP address. The network portion, the subnet portion, and the host portion and it will be classful and work with classful routing protocols such as RIP and IGRP. However if you subnet a subnet (a definition of VLSM) then you have a classless subnet and only classless routing protocols can function properly with them.

The reason for this it that the older classful routing protocols do not send subnetting information with their routing updates and classless routing protocols do. A classful routing protocol can recognize the subnet mask used on the interfaces of the router upon which it is configured. As long as that subnet mask does not vary the protocol will assume that all connected networks have the same subnet mask. At the same time they will assume that all other networks that are advertised to them have the default classful subnet mask. Otherwise it assumes that something is wrong and will not function properly. However, classless routing protocols send subnet mask information with their updates and can function properly with various subnet masks.

The Cisco academy is in a transition period in which it has allowed instructors to choose to continue with CCNA curriculum that does not address VLSM (Variable Length Subnet Masks) or CIDR (Classless InterDomain Routing) in other words version 2.X (hopefully at least 2.1.4) or they can upgrade to the version 3.X curriculum. This is new this year and it is a major revision.

Cisco has determined that it is necessary that among other things the CCNA include VLSM and CIDR. Instructors and students that have not studied the CCNP in the past or have not carefully reviewed the new CCNA curriculum are only projecting their theories or guessing in terms of the distinction between classless and classful.

As of July 31, 2004 Instructors that have not completed new training will no longer be allowed to teach Cisco classes. I suggest that students that have studied under the older curriculum update their knowledge.

Just a quick message to everyone who replied to my post on subnetting!This has certanily cleared it up for me.

Many thanks