Exam 400-007 topic 1 question 101 discussion

A) Correct answer. Traffic aggregate to the link on core switch. Imagine 40 servers with 10Gbps for each, then core switch may need 400Gbps bandwidth but not exist now. So spine and leaf topology is introduced. B) Incorrect. It is advantage on security for separating network. C) Incorrect. It is advantage as you can add as much as switch you like. If you think heavy loading on core switch, then add another core switch as another tree. D) Incorrect. Latency depend on distance, device and link performance. 3 layers design should be local and no distance problem. Device and link performance may be occurred due to traffic aggregration, but mentioned in A more appropriate.

D. - With all those hops you are adding latency to the end to end traffic, we have no information over the bandwidth in play

The correct answer is D. high latency. In a traditional three-tier architecture (Core, Distribution, and Access layers), east-west traffic (traffic between devices within the same data center or between different pods) has to traverse through the distribution and core layers. This introduces additional hops and processing, which can increase latency.

Refer the section “Disadvantages” in this link. https://networklessons.com/cisco/ccna-200-301/spine-and-leaf-architecture It does show both BW and Latency as disadvantages. However, Latency is more optimum here. Latency: The server-to-server latency could be high, depending on the traffic path. East-west traffic between pods has to go through the distribution and core layers.

Latency, because of the hops.

A is the most correct one. Oversubcription ratio

Compared to connecting access or distro layers (non 3 tier) what 3-tiers adds is latency/ Compared to collapsed core-distro, =2 tier, what 3-tier adds is latency, but only if core is routing. There is no mention of Layer 2 or layer 3 solutions. ECMP can balance BW to better potential. Only UCMP on a non 3-tier (not what they are asking, would go for solving Low-BW) Even UCMP 3-tier would also be available. But latency is added always here/ I would not say HIGH/ that is the problem, just INCREASED=some Latency. but also I would not say LOW Bw. Just REDUCED BW, on uplinks. I wonder if they think Ross scales better here, and it is actually C-> Scalability. More Spines, more Leafs, don't add more latency, just about the same. More spine/leafs actually may help BW, but there is not enough info. This is one of the worst Cisco CCDE questions ever. I would go for Low BW, because the uplink cables aggregation factor, but even today with mlag... still on traditional 3-tier... East West traffic, where it matters is when devices are in spread Campus (also not mentioned), including cores and spread distros. Both then latency and BW suffer. Also scales badly.

I agree that Latency and Bandwidth are both correct, however if i think about which one is worse, then Bandwidth wins. I would rather have more BW and more delay than small BW and less delay. Answer A

Its about east-west traffic, latency would not be main concern, its the the bandwidth. 3-tier design inherently enforces to have less links on the core / distribution layer, & as the east-west traffic rapidly grows, we may ran out of ports on core/distri switches(being only few switches, can't be added on the go)

According to this article both A and D are correct in an Three-layer hierarchical architecture. Both make sense in very particular scenarios. Here are two important disadvantages: Limited bandwidth: vPCs solve the STP problem that we can only use one active link, but vPCs are limited to two active uplinks. Latency: The server-to-server latency could be high, depending on the traffic path. East-west traffic between pods has to go through the distribution and core layers. https://networklessons.com/cisco/ccna-200-301/spine-and-leaf-architecture#Three-layer_hierarchical_architecture

https://www.leviton.com/en/solutions/industries/data-centers/architectures/threetier-network-architecture#:~:text=Three%2Dtier%20architecture%20comes%20with,is%20connected%20to%20the%20server.

D is the ans

as there will be more hops added to the path

I think C，a large Layer 2 network is more advantage Comparing with traditional network. The Endpoint can be moved to different pods.

high latency