70-291: DHCP 20/80 rule

Discussion in 'MCSE' started by ISApose, Nov 19, 2008.

  1. ISApose

    ISApose Guest

    Hello, thanks in advance.

    I have seen 2 different views of the 80/20 rule. One that states to break
    up the scope on 2 different DHCP servers (1 in each subnet), and then what
    the MS Press 70-291 book states.

    It states to use the same scope on both servers, but break up the exclusion
    range, 80% and 20%.

    I'm sure the MS Press book is the one I want to go with, but just wanted to
    confirm.
     
    ISApose, Nov 19, 2008
    #1
    1. Advertisements

  2. ISApose

    Pete Jones Guest

    Slightly confused about the 1 in each subnet, but the rest of it reads like
    two different methods to achieve the same result.

    If you are allocating out 192.168.1.50-254 with subnet mask of
    255.255.255.0, you could either:

    Have the scope on Server A (80%) range from 192.168.1.50 to 192.168.1.165,
    with Server B (20) range as 192.168.1.166 to 192.168.1.254. No exclusions or
    reservations on either server.

    The other option is to have Server A and Server B scopes range from
    192.168.1.50 to 192.168.1.254, and have the exclusion on Server A be
    192.168.1.166 to 192.168.1.254, and the exclusion on Server B be 192.168.1.50
    to 192.168.1.165.

    Both these methods will work, and the goal is the same: to have two DHCP
    servers allocating IPs that a) will not conflict/overlap, (b) will work for
    any machine that makes a request, ie it does not matter which server the
    client gets the address from.

    Personally, I prefer to use exclusions for blocks of IP's that are static,
    and alter the scope ranges when balancing DHCP. However, you should learn
    both, since I can guarantee you will encounter both methods in the real world.
     
    Pete Jones, Nov 20, 2008
    #2
    1. Advertisements

  3. ISApose

    ISApose Guest

    Thanks Pete. I will most definitely keep your advice in mind in the field...
    I am a big fan of real world knowledge and application.

    On this particular question however, I am seeking advice on what strategy
    the 70-291 test is going use. I understand that both will work, just not
    sure which one they will be hitting me on.

    The answer will most likely be multiple choice, so I could probably figure
    out which strategy they are looking for by deductive reasoning, but I just
    wanted to be sure.

    Clarification on the '1 in each subnet' statement: The fault tolerance
    situation that I was studying uses the 80/20 rule on 2 dhcp servers, both in
    different subnets connected by a 1542 compliant router that passes bootp
    packets.
     
    ISApose, Nov 20, 2008
    #3
  4. ISApose

    Pete Jones Guest

    Ah, I think I know a variation of that example.

    Two buildings, Site A, Site B. Server A is in Site A, etc.

    Machines in Site A should get the IP from 192.168.0.0/24
    Machines in Site B should get the IP from 192.168.1.0/24

    The first 49 hosts are static for servers, switches, printers etc. in each
    scope.

    So you assign Server A 192.168.0.50-165, and 192.168.1.166-254 as two
    separate scopes.
    Server B gets 192.168.0.166-254 and 192.168.1.50-165.

    The trick is in making sure that the clients are assigned IP's from the
    right scope. You could use MAC address reservations, but in companies with
    hundreds, thousands or more clients, it can get messy. (My thoughts on
    reservations are that if you are going to use them, why not just assign
    static addresses in the first place?)

    Using the options on the DHCP server AND the router, the server can
    determine the origin network, and choose the right scope. This is done either
    by the relay agent (router) rewriting the source address of the Discover
    packet to its own IP, or the relay agent can add a field within the DHCP
    information called the Gateway Interface Address (giaddr).

    Another option would be to superscope the addresses to 192.168.0.0/23, but
    then you would have to create an exclusion for 192.168.1.1-50, and it would
    mean that broadcasts from one site would reach the other. The whole point of
    having the two separate scopes is to cut down the amount of broadcast traffic
    (and it makes recognising the machines easier when looking through logs and
    such.)
     
    Pete Jones, Nov 20, 2008
    #4
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.