Skip navigation
Cisco Learning Home > Certifications > Routing & Switching (CCNP) > Discussions

_Communities

This Question is Answered 1 Helpful Answer available (2 pts)
2242 Views 4 Replies Latest reply: Aug 2, 2011 7:54 AM by latticesam RSS

Currently Being Moderated

Load balancing questions

Jul 13, 2011 7:40 AM

latticesam 33 posts since
Jul 15, 2010

1). I have read somewhere that " Eigrp update packets route reliable change information only to the affected routers"

I thought that the update is forwarded to every neighbor. how does eigrp gets to know which router is affected and which is not?

 

2).Eigrp has been configured to operate over frame relay mutipoint connections.what should the bandwidth command be set to?

 

  The answer said  :the cir rate of lowest speed connection multipiled by no. of pvc but according to Official Guide it should be the sum of CIRs of all PVCs.

 

3).  I  have two questions attached in images,

In the implimentation plan question, what is baseline?

In the Load balancing question , im not able to figure out the succesor and feasible succesor routes left alone load-balancing paths, please help.

 

4). How do we verify that load-balancing is in operation( without looking at the variance setting) ?

 

5). What is hold-down state in eigrp?

 

I highly appreciate your help.

Thanks

  • Ethan 21 posts since
    Aug 27, 2008
    Currently Being Moderated
    1. Jul 26, 2011 12:23 PM (in response to latticesam)
    Re: Load balancing questions

    Howdy,

     

     

    I’ll take a shot at answering a few of your questions. Let’s start on the first one:

     

    1). I have read somewhere that " Eigrp update packets route reliable change information only to the affected routers"

     

    I thought that the update is forwarded to every neighbor. how does eigrp gets to know which router is affected and which is not?”

     

     

     

    Let’s take a look at a quote from “CCNP: Building Scalable Internetworks v5.0.3.0” (this is the only curriculum I have on me, at the moment). Under section 2.1.1, it states “Propagation of partial updates is automatically bounded so that only those routers that need the information are updated.

     

    What you can derive from this is that EIGRP will not propagate queries and/or updates to neighbors that it does not need to. Think of the word “bounded” as boundaries. Let’s use an example:

     

    Suppose I have three routers running EIGRP, we will call them R1, R2, and R3. Assume they are daisy chained together, as such: R1- - - - -R2- - - - -R3. The routers are advertising the following networks, with R2 summarizing between R1 and R3:

     

    R1 – 10.1.1.0/24

     

    R2 – 10.0.0.0/8 and 172.16.0.0/16

     

    R3 – 172.16.1.0 /24

     

    Let’s say that I bring up a new interface on R1 within the 10.0.0.0/8 network. Say, for example, 10.1.2.0/24. R1, assuming this address is included in the EIGRP routing process via the network command, will inform R2 of this new network. R2 knows that it does not need to propagate the new update pertaining to 10.1.2.0/24 to R3, as it is summarizing the entire 10.0.0.0/8 network to R3 (effectively making it the boundary for the 10.0.0.0/8 network). This is the concept of bounded updates. Updates are propagated only to affected routers. R3 will not be affected by routes within the 10.0.0.0/8 network going up or down, as it has a route from R2 for 10.0.0.0/8, therefore it does not need to receive updates.

     

    I would assume (as I do not have experience in the other scenarios), that various other scenarios influence these bounded updates, as well.

     

    Cisco documentation: http://www.cisco.com/en/US/tech/tk365/technologies_white_paper09186a0080094cb7.shtml#summaffect.

     

    Good Blog I found: http://ccietobe.blogspot.com/2009/01/eigrp-bounded-updates.html

     

    I can try and elaborate on this further if this isn’t clearing the concept up for you. Just let me know.

     

     

    On to the next question (2)!

     

     

    2).Eigrp has been configured to operate over frame relay mutipoint connections.what should the bandwidth command be set to?

     

    First, let’s remember that there are two different commands in this scenario; bandwidth, and bandwidth-percent. Both of these commands are for informational purposes only and cannot be used to influence the actual bandwidth of the interface.

     

    EIGRP, by default, can use up to 50 percent of an interface’s reported bandwidth. The aforementioned commands allow us to modify this aspect of EIGRP for special scenarios, helping us to either ensure that EIGRP traffic is not lost when critical, or to ensure that bandwidth remains for passing data during times when EIGRP is busy.

     

    The two answers that you are confused between are two different scenarios. When you have multiple PVCs with differing CIRs, you should multiply the lowest CIR by the number of PVCs. When you have multiple PVCs with the same CIRs, you should multiply the CIR by the number of PVCs, and use that value to configure the bandwidth, as EIGRP will divide evenly amongst the PVCs. I would highly recommend you read Cisco’s documentation on this topic: http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080094063.shtml

     

     

     

     

     

    Question 3!

     

    In terms of your baseline question, Cisco defines a "baseline" as follows (from http://www.cisco.com/en/US/tech/tk869/tk769/technologies_white_paper09186a008014fb3b.shtml#what):

     

    A baseline is a process for studying the network at regular intervals to ensure that the network is working as designed. It is more than a single report detailing the health of the network at a certain point in time. By following the baseline process, you can obtain the following information:

     

    ·         Gain valuable information on the health of the hardware and software

     

    ·         Determine the current utilization of network resources

     

    ·         Make accurate decisions about network alarm thresholds

     

    ·         Identify current network problems

     

    ·         Predict future problems

     

    So, in this specific scenario, you would want to baseline the network performance between the two hosts prior to and after the move to load balanced routing, in order to verify and document the improvements made by those changes.

     

    In terms to the second part of this questions, referring to load balancing, remember that EIGRP will choose the route with the lowest Feasible Distance, which is the Advertised Distance (The distance from the advertising router to the destination network) + the cost between the local router and the next hop (advertising) router.

     

    So, let’s first figure out which route will be chosen as the best route, based on the lowest FD.

     

    ·         Path offered by B: AD = 20 + distance between A and B = 40 = FD of 60.*

     

    ·         Path offered by D: AD = 20 + distance between A and D = 10 = FD of 30.

     

    ·         Path offered by F: AD = 20 + distance between A and F = 20 = FD of 40.*

     

    So, as you can see when comparing the FDs of the path’s offered from Router A’s three neighbors, the path offered by Router D has the lowest FD, 30, which will be chosen as the best route.

     

    By default, EIGRP will load balance across equal cost routes, using up to 4 different paths by default. In order to achieve unequal-cost­ load balancing, you need to employ the use of the variance command. The variance command specifics how many multiples of the best route’s FD will be considered as the maximum metric for installing unequal-cost routes for load balancing. So, in other words, if I have configured a variance of 2, the formula would work as follows: 2 * FD of best route (30) = 60 = Routes with a FD of up to 60 can be installed as routes in the routing table and used for load balancing.

     

    Hopefully this makes sense so far. Now, going back to the question:

     

    The path offered by Router D has been installed as the best path, as it has the lowest FD (30). The “variance 2” command has been configured, allowing for routes with a FD of up to 60 to be installed, as well. The path offered by Router B has an FD of 60, which meets the allowable FD generated by the variance formula, so this path will be installed. The path offered by Router F falls within the allowable FD (60), as well, so it will be installed as a route in the routing table, ultimately allowing unequal-cost load balancing across all three paths (ADEH, ABEH, and AFEH).*

     

    You fourth question ties into this as well. Want to verify load balancing? A quick an easy way is to just do a “show ip route” command. You will see more than one route installed for a single destination within the routing table. Let’s say I’m load balancing to the destination network 192.168.151.0/24. You will see something like this in your routing table under the “show ip route” command.

     

    D    192.168.151.0/24 [90/3328] via 172.30.253.81, 7w0d, GigabitEthernet3/2

     

              [90/3328] via 172.30.252.81, 7w0d, GigabitEthernet3/1

     

     

     

    Last question!

     

    “5). What is hold-down state in eigrp?

     

    I believe you are referring to the hold down timer. In this case, I’ll just pull a quote from Cisco’s documentation, again (http://www.cisco.com/en/US/tech/tk365/technologies_white_paper09186a0080094cb7.shtml):

     

    “The rate at which EIGRP sends hello packets is called the hello interval, and you can adjust it per interface with the ip hello-interval eigrp command. The hold time is the amount of time that a router will consider a neighbor alive without receiving a hello packet. The hold time is typically three times the hello interval, by default, 15 seconds and 180 seconds. You can adjust the hold time with the ip hold-time eigrp command.

     

     

     

    I know that was a lot of information, but hopefully that helps. It’s always good for me to go back and review this stuff, as well J.

     

     

    *EDIT: I apologize, I had glanced at your load balancing question and then written my response to said question based on memory, and pieces of my response were incorrect. I've corrected the erroneous information within my response. Hopefully this makes more sense now.

     

    Ethan

  • icewulfie 17 posts since
    Aug 18, 2008
    Currently Being Moderated
    3. Jul 26, 2011 8:58 AM (in response to latticesam)
    Re: Load balancing questions
    • A-B-C-H has an FD of 70, AD of 30. Not in routing table.
    • A-B-E-H has an FD of 60, AD of 20. Feasible successor.
    • A-D-E-H has an FD of 30, AD of 20. Successor.
    • A-F-E-H has an FD of 40, AD of 20. Feasible successor.
    • A-F-G-H has an FD of 60, AD of 40. Not in routing table.

     

        A will check the FD of all the routes and see that ADEH has the best FD, therefore becoming the sucessor. All the routes that have an AD less than the FD of the successor route will become feasible successors; AFEH and ABEH qualify to the rule, so they become a backup route. Since we have a variance of 2, router A will load balance between feasible successors as long as their FD is 2x or less than the FD of the successor; AFEH and ABEH both qualify for load balancing as both their FD is at most twice as bad as A-D-E-H.

     

    Router A will then load balance on all of its 3 paths to Router B, D and F.

Actions

More Like This

  • Retrieving data ...

Bookmarked By (0)