Lab 11.3.4: Configuring TCP Load Distribution with Frame Relay

Objectives:

  • To configure a Frame Relay connection between our WebHeads router in Orlando and the ISP's router in Houston, and configure TCP Load Distribution on the WebHeads router.

Scenario:

We work for a company that hosts web sites. The company, "WebHeads" currently has only one main web server, hosting a large number of web sites. The web sites hosted at WebHeads are growing increasingly more popular with Internet users. We have done testing and found that the frame relay link and our router can easily handle more information, but our server is having a hard time keeping up with the requests for web pages. One option is to buy a new more powerful web server, but our superiors are afraid of the loss of business if the server ever failed, and the down time would cost the WebHeads company a lot of money. Our solution instead was to buy multiple servers which have duplicate information loaded on them, that way if one should ever fail, the others would still be up and operating, and the company would not have any down time or loss of revenue. Since the original server was at IP address 4.4.4.4, we want to keep that same address for all of the servers in use. The way that we can accomplish this is through the use of Load Distribution on our router. It is our job to set up this connection and to correctly configure Load Distribution on the WebHeads router.

Notes: 

Lab Tasks:

  1. Cable the lab as shown in the diagram.
  2. The next task is to set up the routers and Frame Relay connection as in previous labs:
    • Configure IGRP routing with the AS# 707.
    • Configure the correct IP addresses on each of the routers.
    • Configure the encapsulation type, DLCI, LMI type, and any frame relay maps.
    • Note: Houston, use DLCI 17 to get to WebHeads. WebHeads use DLCI 18 to get to Houston.

    Don't forget the no shutdown command on the interfaces.

  3. Before configuring Load distribution on the WebHeads router, test your configurations by pinging the other router. This way you are insuring that your basic configuration is correct before changing it.
  4. The first step in configuring load distribution is to define a pool of addresses containing the IP addresses of the real web servers. Since we are going to take down the original server when load distribution is complete, its original address of 4.4.4.4 is going to be a virtual address that is shared by all of the web servers. To the outside world it will only look like one server even though each of the servers are sharing part of the load. The servers still need real IP addresses, and the router needs to know what those addresses are so that it can send the first request to the first server, the next address to the next server, etc. Use the following command:

    WebHeads(config)#ip nat pool rot-hosts 4.4.4.10 4.4.4.12 prefix-length 24 type rotary

    What does the prefix-length in the previous command refer to?

    What does the term rotary mean?

  5. Next we need to define a standard access list permitting the address of the virtual server(the address that all of the servers will be sharing).

    WebHeads(config)#access-list 5 permit 4.4.4.4
  6. Now is the time to enter the command to establish dynamic inside destination translation, identifying the access list defined in the prior step.

    WebHeads(config)#ip nat inside destination list 5 pool rot-hosts

    Explain in your own words what the previous command is telling the router.

  7. We next need to tell the router which interface is connected to inside network and which interface is connected to the outside.

    WebHeads(config)#interface fast ethernet 0/0
    WebHeads(config-if)#ip nat inside

    WebHeads(config)#interface serial 0/0
    WebHeads(config-if)#ip nat outside

  8. Now we get to verify the routers address translation.

    WebHeads#show ip nat translations

    What information did the router respond with?



    WebHeads#show ip nat statistics

    What information did the router respond with?

  9. For lab testing purposes set the default pages of the 3 different servers to say three different things so you will know which web server you have hit. Now it is time to test our new configuration. From a workstation on the ISP routers network open Internet Explorer or Netscape Navigator. For an IP address, use 4.4.4.4, the IP address of our virtual server.

    WebHeads#show ip nat translations

    What information did the router respond with?

    WebHeads#show ip nat translations verbose

    What additional information did the router respond with?

    WebHeads#show ip nat statistics

    What information did the router respond with?

  10. From the workstation refresh the Internet browser window a few times and observe the resulting web pages. Check ip nat translations again on the router.

What were the inside IP addresses that the NAT table rotated through?