Several unrelated jobs slow down at once and nothing is dropping packets. How do you confirm a pause cascade and find where it started?

Follow pause requests against the direction of data, estimate queue growth and distinguish cascade evidence from its possible causes.

Updated Oct 2026 · Learn how AI infrastructure works through explanations, worked examples and diagrams, then practise applying it to interview questions.

Follow pause requests against the direction of data, estimate queue growth and distinguish cascade evidence from its possible causes.

The full answer is part of Premium. A free account includes more answers in each topic, but does not unlock this one.

20 answers per topic instead of 10, plus saved progress and bookmarks · no cardor unlock all 316 remaining answers · ₹2,000 / $25

The concepts behind this question

Use these concept explanations to review the ideas behind this question.

Core

🔌 Networking & StorageSign in

RDMA, InfiniBand and RoCEv2RDMA lets network adapters transfer data directly between registered memory regions. InfiniBand and RoCE provide different network and operating environments for those transfers.

Advanced

🔌 Networking & Storage🔒 Premium

Congestion Control for AI FabricsDiagnose RoCE congestion using queue growth, ECN feedback and per-priority pause duration. Compare matched traffic patterns before changing routing or flow control.

Foundational

🔌 Networking & Storage

NCCL and Collective AlgorithmsNCCL implements GPU collectives using algorithms, protocols and channels over available hardware paths. Learn what each choice changes, how to read bandwidth results, and which evidence separates a communication problem from late application work.

Advanced

🔌 Networking & Storage🔒 Premium

Rail-Optimized and Fat-Tree FabricsRail-optimized and Clos fabrics organize traffic between GPU nodes. Port counts, placement and oversubscription determine which paths are available to collectives.

Explore the full curriculum →

UP NEXT ON YOUR JOURNEY

Next in this track8,192 GPUs exceeds what a two-tier Clos of 64-port switches supports. Compare adding a tier against using higher-radix switches.Next in this trackHow would you know GPUDirect RDMA is not being used, given that nothing reports an error when it is disabled?Next in this trackDesign storage for a multi-tenant GPU cloud. How do you keep one customer's checkpoint burst from slowing another's training run?

FEDITOR'S NOTE

Trace data and pause directions explicitly. Use queue growth and feedback delay to explain why ECN thresholds alone cannot establish a safe configuration.

DISCUSSION · 0

No comments yet — be the first to share your approach.