+- How2pass.com Forums (https://www.how2pass.com/forum)
+-- Forum: CCNP (https://www.how2pass.com/forum/forum-6.html)
+--- Forum: CCNP ENARSI 300-410 Forum (https://www.how2pass.com/forum/forum-15.html)
+--- Thread: Possible Correction Needed - QID RTE029 (/thread-800.html)
Possible Correction Needed - QID RTE029 - zdormanjones - 07-01-2013
The answer to this question claims that traffic from the Internet to the data center is load balanced across 3 paths, but I think the correct answer is 4 paths.
It's true that Router A load balances across its 3 links, but the question states that all routers have the variance configured. Therefore, Router B will load balance across two of its links, making a fourth path.
I set this up in GNS3 with the ratios of the metrics for the various links exactly the same as presented in the question. I checked the routing table on A, and sure enough, it had 3 entries for the "data center" (loopback) connected to router H: through B, D, and F. Router B had two routes to the data center: through C and E.
Then I set up Wireshark on Router E on the links to Routers B, D, and F, as well as on Router C on the link facing router B. Finally, I sent 90 icmp packets from Router A to the "data center," and the results were as I predicted:
A->D->E->H: 40 packets
A->F->E->H: 30 packets
A->B->E->H: 12 packets
A->B->C->H: 8 packets
I will grant that the question states that the K-values are set to their default, while I set them to only compute delay on all routers. However, I can't conceive of any way that would contaminate the results.