NVIDIA NCP-AIN Exam - Topic 2 Question 12 Discussion

Actual exam question for NVIDIA's NCP-AIN exam

Question #: 12
Topic #: 2

[All NCP-AIN Questions]

[AI Network Architecture]

Which of the following statements are true about AI workloads and adaptive routing?

Pick the 2 correct responses below.

AAI workloads are made of a small number of volumetric flows called elephant flows.

BAI workloads have very high entropy that helps spread traffic evenly without congestion.

CFlow-based load balancing mechanisms increase congestion risk.

DECMP-based load balancing works best for AI workloads.

Show Suggested Answer

Suggested Answer: A, C

AI workloads, particularly in large-scale training scenarios, are characterized by a small number of high-bandwidth, long-lived flows known as 'elephant flows.' These flows can dominate network traffic and are prone to causing congestion if not managed effectively.

Traditional flow-based load balancing mechanisms, such as Equal-Cost Multipath (ECMP), distribute traffic based on flow hashes. However, in AI workloads with low entropy (i.e., limited variability in flow characteristics), ECMP can lead to uneven traffic distribution and congestion on certain paths.

Adaptive routing techniques, which dynamically adjust paths based on real-time network conditions, are more effective in managing AI traffic patterns and mitigating congestion risks.

by Isadora at Nov 20, 2025, 07:46 PM

Limited Time Offer

25%

Off

Get Premium NCP-AIN Questions as Interactive Web-Based Practice Test or PDF

Contribute your Thoughts:

Submit Cancel

Leonora

2 days ago

I’m leaning towards A and C. Elephant flows can lead to congestion.

upvoted 0 times

...

Miesha

7 days ago

C seems true too. Flow-based balancing can cause issues.

upvoted 0 times

...

Eliseo

13 days ago

I agree with A and D. ECMP is efficient for AI traffic.

upvoted 0 times

...

Rebbecca

18 days ago

I feel like B is misleading. High entropy doesn't always mean no congestion.

upvoted 0 times

...

Francine

23 days ago

Wait, are we really saying AI workloads have high entropy? That sounds off.

upvoted 0 times

...

Tiara

1 month ago

Not so sure about B, high entropy doesn't always mean no congestion.

upvoted 0 times

...

Ceola

2 months ago

I think D is spot on too, ECMP is great for AI.

upvoted 0 times

...

Hyman

2 months ago

A is definitely true, elephant flows are a thing.

upvoted 0 times

...

Beckie

2 months ago

I'll just flip a coin. Heads for A and B, tails for C and D. Easy peasy.

upvoted 0 times

...

Long

2 months ago

Elephant flows? More like mammoth flows! This AI stuff is getting out of hand.

upvoted 0 times

...

Roosevelt

2 months ago

C is true. Flow-based load balancing can increase the risk of congestion.

upvoted 0 times

...

Abel

2 months ago

D is the right choice. ECMP-based load balancing is the best fit for AI workloads.

upvoted 0 times

...

Ammie

3 months ago

I'm a bit confused about flow-based load balancing; I thought it was supposed to help, but now I'm questioning if it could actually increase congestion like option C suggests.

upvoted 0 times

...

Temeka

3 months ago

I practiced a question similar to this, and I feel like ECMP-based load balancing is often mentioned as effective for AI workloads, so I might lean towards option D.

upvoted 0 times

...

Fletcher

3 months ago

I think option B sounds familiar because we discussed how high entropy can help with traffic distribution, but I can't recall the specifics.

upvoted 0 times

...

Dorathy

3 months ago

I remember studying that elephant flows are significant in AI workloads, but I'm not sure if they are the only type of flow we should consider.

upvoted 0 times

...

Adelina

3 months ago

I've got a pretty good handle on this topic. I know AI workloads tend to have a mix of large and small flows, so A is likely true. And flow-based load balancing can actually increase congestion risk, so C is also correct. I'll select A and C.

upvoted 0 times

...

Stanford

3 months ago

I think A and D are correct. Elephant flows are common in AI.

upvoted 0 times

...

Kenneth

4 months ago

I think B is the correct answer. AI workloads have high entropy, which helps distribute traffic evenly.

upvoted 0 times

...

Domitila

4 months ago

Okay, let me think this through. I remember learning that AI workloads can be quite bursty, so B seems plausible. And I think ECMP is a common technique used for load balancing, so D makes sense. I'll go with those two.

upvoted 0 times

...

Martin

4 months ago

I'm going with A and D. AI workloads have elephant flows, and ECMP is the way to go.

upvoted 0 times

...

Felicitas

4 months ago

Totally agree with A and D, they make the most sense!

upvoted 0 times

...

Polly

4 months ago

Hmm, this is a tricky one. I'm a bit confused about the relationship between AI workloads and adaptive routing. I'm leaning towards A and D, since I know AI workloads can have some large flows, and ECMP might be a good approach. But I'm not 100% confident.

upvoted 0 times

...

Alona

5 months ago

I'm not totally sure about the differences between elephant flows and AI workloads, but I think B and D might be the right answers. The high entropy part sounds right, and ECMP seems like it could work well for this type of traffic.

upvoted 0 times

...