Options

Why do we subnet?

DerekAustin26DerekAustin26 Member Posts: 275
Whats the point?

You can setup a network without subnetting and it'll work just fine. VLANs segment & segregate your broadcast domains. So does Subnetting do the same? If not, why subnet?

Comments

  • Options
    KaminskyKaminsky Member Posts: 1,235
    Primary historic reason. IP4 wasted a lot of IP address space. Subnetting/VLANs also help reduce traffic on the LAN.

    People think that when Ipv6 is fully out, address space won't be a problem but you can expect technology to have your washing machine messaging the support company when it wants to be serviced, so it won't be long before IPv6 needs to conserve address space.

    Sure there are better answers but that's it in a nutshell.
    Kam.
  • Options
    SlowhandSlowhand Mod Posts: 5,161 Mod
    How would you reach those different segments on different VLANs or on opposite sides of a router, without subnets? If we all simply used a single address-space, every computer on your network would have to have a public IP address, leased from InterNIC more than likely. Even then, it would be challenging for an ISP to assign a block of addresses, such as a /29, for example, to a single customer.

    Back to the private address spaces, (192.168.n.n, 172.16.n.n - 172.31.n.n, and 10.n.n.n,) you need to be able to be able to give addresses to those separate networks. If every port on a router had the same address space, let's use 10.0.0.0 for this example, then it would receive packets and have no idea which port to send them out of. If all VLANs had the same network address, a layer 3 switch or router wouldn't know where to send packets destined for 10.0.0.1, and they wouldn't know the difference between that address and 10.1.1.1.

    Here are some breakdowns of how subnetting works, which will be key to understanding why we use it:

    LearnToSubnet.com
    Subnet Addressing Tutorial from Network World
    Cisco's IP Addressing and Subnetting Tutorial

    Free Microsoft Training: Microsoft Learn
    Free PowerShell Resources: Top PowerShell Blogs
    Free DevOps/Azure Resources: Visual Studio Dev Essentials

    Let it never be said that I didn't do the very least I could do.
  • Options
    rwwest7rwwest7 Member Posts: 300
    Couldn't get enough with the SMTP/POP thread?

    Well, we subnet because we have to
  • Options
    /usr/usr Member Posts: 1,768
    Couldn't get enough with the SMTP/POP thread?

    I didn't even realize it was the same OP until you said that...

    I'm not even getting started in this thread.
  • Options
    KaminskyKaminsky Member Posts: 1,235
    Let me give you an example.

    My first job in IT, large site, we all shared (as in hubs) a single class C address. There was no internet so that didn't come into it.

    Main problem was slowness of the network. Initially due to the hubs bouncing out everything to everyone and swamping the lan. Mainly a collision domain issue (collision, backoff, wait on a site wide basis - was hellish. You pc would just work in little spurts of activity). When traffic started approaching only 30%-40% then lan was practically unusable and everyone complained. (luckily not to me!)

    With switches instead of hubs, this same thing would happen with broadcasts although switch design would help a lot with the collisions, broadcasts would still present a big problem especially early in the morning when people are starting work and IP addresses are having to be relearnt.

    With subnetting and VLANs, you can break up your address space into much smaller chunks (which is really good for security) so traffic passing through your NIC card is only for a relatively small number of users. Anyone suddenly decides to download a whopping great file, they are not going to swamp the whole site but just those in the local area.

    Now, if you imagine all the companies in the world and especially with the internet, there is no way it could work without subnetting, even with the massive address space of IPv6.

    Have a read through the WIKI on it for more info. Subnetwork - Wikipedia, the free encyclopedia
    Kam.
  • Options
    amp2030amp2030 Member Posts: 253
    Whats the point?

    You can setup a network without subnetting and it'll work just fine. VLANs segment & segregate your broadcast domains. So does Subnetting do the same? If not, why subnet?

    What is your definition of 'just fine'?
  • Options
    rwwest7rwwest7 Member Posts: 300
    Stop biting this guys bait.
  • Options
    HeroPsychoHeroPsycho Inactive Imported Users Posts: 1,940
    I'd respond with something helpful, but I saw the last thread about POP3/SMTP, and I decided to live my life instead.
    Good luck to all!
  • Options
    KaminskyKaminsky Member Posts: 1,235
    Now .. Now .. There's nothing wrong with passion. Coupled with manners would be nice, I agree .. but it is quite a valid theoretical question.

    At the end of the day, the answer boils down to, if you don't know how to, you will lose out painfully to someone that does.
    Kam.
  • Options
    NeekoNeeko Member Posts: 170
    We don't need to subnet. In fact, the whole interweb should be on one subnet IMHO.
  • Options
    daniel2009daniel2009 Member Posts: 20 ■□□□□□□□□□
    Before subnets existed, networks were divided in classes.
    A company could get a class A, class B and/or class C network.
    Class A was 16777215-2 = 16777213 private IPs.
    Class B was 65536-2 = 65534 private IPs
    Class C was 256-2 = 254 private IPs

    An IP address is only 32 bit long.
    11111111.11111111.11111111.11111111

    This means there can be only 4294967295 (2^32) private IPs.

    IETF figured in 1994 or somewhere around there that there would be soon no more IPs available if each company would register 16 million private ips.


    There would be no more IPs if
    Total - (Class A * amount) - (Class B * amount) - (Class C * amount) is near or is 0.


    4294967295 - ( 16777213 * a ) - ( 65534 * b ) - (254 * c ) =~ 0


    On the internet there are lists of these classes still maintained.
    For example there are 66 class A adresses.
    IP network numbers - Class A

    There are around 12000 class B addressess
    IP network numbers - Class B

    There are around 800000 class C addresses
    IP network numbers - Class C


    So
    4294967295 - 1107296058 - 786408000 - 203200000 = 2198063237 free.

    This is about 51,2% which is used.



    There would be ( 2198063237 / 16777213 ) = 131 class A available.
    So 131 companies would register a Class A network, and the internet would be full.

    Subnets are a solution to make this happen less fast. With them one can create networks of different sizes.
    I always wondered why they just didn't make an IP 128 bits long instead of 32
  • Options
    DerekAustin26DerekAustin26 Member Posts: 275
    I think i got it.. Correct me if I'm wrong

    If you have too many devices connected to one port on a router, you can create a seperate segment on a seperate port on that router(to improve performance) and then configure the Subnet so that the Router can Identify with which port is designated to which network.

    Cuz if not, the router would just broadcast incoming traffic out both ports if they were both the exact same subnet?


    Sound correct?
  • Options
    KaminskyKaminsky Member Posts: 1,235
    You're getting the idea but you really need to go slowly through the CCENT book.

    The more you rush and want the answer now, the slower it will take you.
    The more you jump to conclusions, the more confused and frustrated you will become.

    Being the first step into the networking certs, common sense says CCNA should be a doddle if someone just applies themselves.... Because it covers a broad ground base of networking basic theory, the CCNA is a very difficult exam as it requires a LOT of reading and practice. It is renown for slaying those that don’t do their homework properly.(especially cocky new MCSEs)
    I think i got it.. Correct me if I'm wrong

    If you have too many devices connected to one port on a router, you can create a seperate segment on a seperate port on that router(to improve performance) and then configure the Subnet so that the Router can Identify with which port is designated to which network.

    Cuz if not, the router would just broadcast incoming traffic out both ports if they were both the exact same subnet?

    Sound correct?

    Firslty, routers have very few connections. Switches have tons which is where your average PC/printer will connect to. The switch in turn will either connect to a single port on a much more powerfull switch (along with many other access switches) or in small networks, to a single FA port on a router.

    The wiring on each port on a switch stops collision domains but it doesn't stop broadcast domains. The only device that does is a router.

    Vlans are used to seperate hosts/subnets on a switch but for one vlan to talk to another usually requires a connection from the switch to a router which does the routing between the seperate Vlans.


    I know your eager to learn and have all the answers and wiz through this CCNA cert but the best place for you right now is to stick your head into the ccent book and don't come out until it says Index. You'll have a million and one questions after each chapter but don't jump the gun. That will be your first reading. Most people that obtain the cert read it through many times.

    You'll be a much better network engineer for it.
    Kam.
Sign In or Register to comment.