cert
cert-1
cert-2

Pass Cisco 300-610 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
300-610 Exam - Verified By Experts
300-610 Premium File

300-610 Premium File

$69.99
$76.99
  • Premium File 350 Questions & Answers. Last Update: Nov 13, 2026

Whats Included:

  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
 
$76.99
$69.99
accept 10 downloads in the last 7 days
block-screenshots
Cisco 300-610 Practice Test 16

Question 150

At a dual-spine infrastructure project, after correction, the team wants to prevent recurrence of: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. The team needs a defensible design decision rather than a vendor feature checklist. Which continuing design control best prevents the same problem from recurring?

  1. Retune thresholds on hardware or link-speed changes.
  2. Alert on prolonged PFC duration and remediate misbehaving NIC firmware.
  3. Document which queues truly require no-drop behavior.
  4. Recheck DSCP and 802.1p mappings when new services are added.

Explanation

Correct Answer

A:

The most relevant ongoing safeguard is: Retune thresholds on hardware or link-speed changes. The risk it addresses is: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. This check is tied to the technical principle: ECN can signal endpoints earlier than PFC; correctly placed ECN and PFC thresholds may improve stable no-drop transport. The immediate correction remains: Configure supported ECN thresholds below PFC emergency thresholds and verify endpoint feedback. Periodically validate the safeguard through two different observations. Observe ECN-marked packet counts and congestion notification Compare queue fill and PFC frame occurrence under controlled incast. The continuing acceptance target is: Congestion is relieved primarily by rate feedback rather than persistent PFC deadlock.

Incorrect Answers

B:

This is a continuing control for pfc watchdog failure handling; its underlying principle is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. The recurring risk here is different: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The intended long-term acceptance condition remains: Congestion is relieved primarily by rate feedback rather than persistent PFC deadlock.

C:

This is a continuing control for avoiding global pause for ai traffic; its underlying principle is: Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback. The recurring risk here is different: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. The intended long-term acceptance condition remains: Congestion is relieved primarily by rate feedback rather than persistent PFC deadlock.

D:

This is a continuing control for pfc head-of-line blocking; its underlying principle is: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. The recurring risk here is different: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. The intended long-term acceptance condition remains: Congestion is relieved primarily by rate feedback rather than persistent PFC deadlock.

Question 151

At a low-latency GPU workload fabric, a design review reports the following issue: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The maintenance plan must not disrupt unrelated tenants. Which design correction most directly addresses this specific failure or requirement?

  1. Set class bandwidth shares and admission assumptions alongside PFC and ECN policy.
  2. Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class.
  3. Trace the packet's markings and class mapping across every hop and align supported QoS policy.
  4. Adopt class-specific supported PFC/ECN design instead of indiscriminate link-level pause.

Explanation

Correct Answer

B:

The relevant failure is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The technical distinction is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. The targeted correction is: Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class. Before changing the design, check two independent observations. Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab. Accept the change only if: A pause storm cannot indefinitely block the rest of the fabric.

Incorrect Answers

A:

This action targets ets bandwidth for rdma and storage, where the mechanism is: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. That is a separate condition from the problem in this question: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. Even a successful repair of the alternative would not establish the required result: A pause storm cannot indefinitely block the rest of the fabric. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

C:

This action targets qos mapping drift across fabric, where the mechanism is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path. That is a separate condition from the problem in this question: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. Even a successful repair of the alternative would not establish the required result: A pause storm cannot indefinitely block the rest of the fabric. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

D:

This action targets avoiding global pause for ai traffic, where the mechanism is: Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback. That is a separate condition from the problem in this question: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. Even a successful repair of the alternative would not establish the required result: A pause storm cannot indefinitely block the rest of the fabric. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

Question 152

At a two-region enterprise VXLAN fabric, two engineers propose competing diagnoses for this condition: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. Acceptance requires evidence from the running system, not just the saved design. Which pair of observations most directly tests the proposed diagnosis before a change?

  1. Inspect per-priority pause frame counters and queue assignments; correlate paused priority events with control-plane latency spikes.
  2. Inspect watchdog trigger and blocked queue counters; verify traffic resumes after isolating the misbehaving sender in a lab.
  3. Observe ECN-marked packet counts and congestion notification; compare queue fill and PFC frame occurrence under controlled incast.
  4. Inspect class scheduler utilization and offered load; compare RDMA throughput and tail latency under mixed high-priority traffic.

Explanation

Correct Answer

B:

The first observation is: Inspect watchdog trigger and blocked queue counters The independent second observation is: Verify traffic resumes after isolating the misbehaving sender in a lab. Together they test the proposed diagnosis of this incident: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The underlying technical principle is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. If those checks confirm it, the appropriate next action is: Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class. The expected measurable result is: A pause storm cannot indefinitely block the rest of the fabric.

Incorrect Answers

A:

These observations investigate pfc head-of-line blocking; they would be useful if the symptoms were: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. That alternative rests on the principle: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. The question instead asks for evidence about: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The discriminating observations here are: Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab.

C:

These observations investigate ecn marking before pfc; they would be useful if the symptoms were: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. That alternative rests on the principle: ECN can signal endpoints earlier than PFC; correctly placed ECN and PFC thresholds may improve stable no-drop transport. The question instead asks for evidence about: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The discriminating observations here are: Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab.

D:

These observations investigate ets bandwidth for rdma and storage; they would be useful if the symptoms were: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. That alternative rests on the principle: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. The question instead asks for evidence about: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The discriminating observations here are: Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab.

Question 153

At a 60-rack compute cluster, the design group challenges an assumption behind this issue: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. One component outage must remain inside the agreed availability budget. Which engineering principle best explains the observed behavior?

  1. Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback.
  2. PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking.
  3. Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses.
  4. Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class.

Explanation

Correct Answer

C:

The controlling technical principle is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. This matters because the reported design failure is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The first diagnostic check is: Inspect watchdog trigger and blocked queue counters The independent check is: Verify traffic resumes after isolating the misbehaving sender in a lab. These facts support the targeted intervention: Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class. Confirm that the resulting design meets this condition: A pause storm cannot indefinitely block the rest of the fabric.

Incorrect Answers

A:

The offered principle describes avoiding global pause for ai traffic: Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback. That is useful for a distinct incident, such as: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. It does not account for the reported failure: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The governing principle in this situation is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses.

B:

The offered principle describes pfc head-of-line blocking: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. That is useful for a distinct incident, such as: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. It does not account for the reported failure: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The governing principle in this situation is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses.

D:

The offered principle describes ets bandwidth for rdma and storage: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. That is useful for a distinct incident, such as: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. It does not account for the reported failure: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The governing principle in this situation is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses.

Question 154

At a virtualized application data center, a change has been proposed for the following condition: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The architect must preserve the intended failure domains. Which measured result most directly proves the original requirement is satisfied?

  1. RoCE loss goals are met without unnecessarily suspending all traffic on the link.
  2. A pause storm cannot indefinitely block the rest of the fabric.
  3. Loss-sensitive service obtains its needed bandwidth without starving other queues.
  4. RDMA packets traverse the expected protected class on every transit device.

Explanation

Correct Answer

B:

The relevant measurable success criterion is: A pause storm cannot indefinitely block the rest of the fabric. It addresses the original reported defect: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. A saved design or configuration command alone does not establish that outcome. Observe the affected system in two independent ways: Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab. The correction being evaluated is: Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class. The underlying reason that correction matters is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses.

Incorrect Answers

A:

This metric would help validate a correction to avoiding global pause for ai traffic, rather than the original reported issue. The other issue is: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. Here the defect to be resolved is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The expected measurable outcome for the actual incident is: A pause storm cannot indefinitely block the rest of the fabric. A positive result for the other service would not prove this requirement is met.

C:

This metric would help validate a correction to ets bandwidth for rdma and storage, rather than the original reported issue. The other issue is: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. Here the defect to be resolved is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The expected measurable outcome for the actual incident is: A pause storm cannot indefinitely block the rest of the fabric. A positive result for the other service would not prove this requirement is met.

D:

This metric would help validate a correction to qos mapping drift across fabric, rather than the original reported issue. The other issue is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. Here the defect to be resolved is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The expected measurable outcome for the actual incident is: A pause storm cannot indefinitely block the rest of the fabric. A positive result for the other service would not prove this requirement is met.

Question 155

At an AI-training data-center pod, a post-incident design review focuses on preventing: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The team needs a defensible design decision rather than a vendor feature checklist. Which continuing design control best prevents the same problem from recurring?

  1. Review QoS shares when application proportions change.
  2. Alert on prolonged PFC duration and remediate misbehaving NIC firmware.
  3. Recheck DSCP and 802.1p mappings when new services are added.
  4. Document which queues truly require no-drop behavior.

Explanation

Correct Answer

B:

The most relevant ongoing safeguard is: Alert on prolonged PFC duration and remediate misbehaving NIC firmware. The risk it addresses is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. This check is tied to the technical principle: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. The immediate correction remains: Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class. Periodically validate the safeguard through two different observations. Inspect watchdog trigger and blocked queue counters Verify traffic resumes after isolating the misbehaving sender in a lab. The continuing acceptance target is: A pause storm cannot indefinitely block the rest of the fabric.

Incorrect Answers

A:

This is a continuing control for ets bandwidth for rdma and storage; its underlying principle is: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. The recurring risk here is different: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. The intended long-term acceptance condition remains: A pause storm cannot indefinitely block the rest of the fabric.

C:

This is a continuing control for pfc head-of-line blocking; its underlying principle is: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. The recurring risk here is different: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. The intended long-term acceptance condition remains: A pause storm cannot indefinitely block the rest of the fabric.

D:

This is a continuing control for avoiding global pause for ai traffic; its underlying principle is: Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback. The recurring risk here is different: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. The intended long-term acceptance condition remains: A pause storm cannot indefinitely block the rest of the fabric.

Question 156

At a regulated private-cloud campus, during a controlled design review, the team discovers: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The maintenance plan must not disrupt unrelated tenants. Which design correction most directly addresses this specific failure or requirement?

  1. Set class bandwidth shares and admission assumptions alongside PFC and ECN policy.
  2. Limit PFC to required RDMA classes and retain appropriate lossy queues for unrelated traffic.
  3. Trace the packet's markings and class mapping across every hop and align supported QoS policy.
  4. Enable documented PFC watchdog handling with a planned drop and recovery policy for the affected class.

Explanation

Correct Answer

C:

The relevant failure is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The technical distinction is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path. The targeted correction is: Trace the packet's markings and class mapping across every hop and align supported QoS policy. Before changing the design, check two independent observations. Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class. Accept the change only if: RDMA packets traverse the expected protected class on every transit device.

Incorrect Answers

A:

This action targets ets bandwidth for rdma and storage, where the mechanism is: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. That is a separate condition from the problem in this question: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. Even a successful repair of the alternative would not establish the required result: RDMA packets traverse the expected protected class on every transit device. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

B:

This action targets pfc head-of-line blocking, where the mechanism is: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. That is a separate condition from the problem in this question: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. Even a successful repair of the alternative would not establish the required result: RDMA packets traverse the expected protected class on every transit device. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

D:

This action targets pfc watchdog failure handling, where the mechanism is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. That is a separate condition from the problem in this question: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. Even a successful repair of the alternative would not establish the required result: RDMA packets traverse the expected protected class on every transit device. The applicable design change must address the diagnosed mechanism of the stated case, not the neighboring design problem.

Question 157

At a dual-spine infrastructure project, before changing the fabric, the team observes: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. Acceptance requires evidence from the running system, not just the saved design. Which pair of observations most directly tests the proposed diagnosis before a change?

  1. Review whether global pause or PFC is actually negotiated; measure impact on ordinary TCP and control flows during a test burst.
  2. Inspect class scheduler utilization and offered load; compare RDMA throughput and tail latency under mixed high-priority traffic.
  3. Capture DSCP and CoS before and after the suspect hop; compare egress queue and drop counters to the intended class.
  4. Inspect per-priority pause frame counters and queue assignments; correlate paused priority events with control-plane latency spikes.

Explanation

Correct Answer

C:

The first observation is: Capture DSCP and CoS before and after the suspect hop The independent second observation is: Compare egress queue and drop counters to the intended class. Together they test the proposed diagnosis of this incident: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The underlying technical principle is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path. If those checks confirm it, the appropriate next action is: Trace the packet's markings and class mapping across every hop and align supported QoS policy. The expected measurable result is: RDMA packets traverse the expected protected class on every transit device.

Incorrect Answers

A:

These observations investigate avoiding global pause for ai traffic; they would be useful if the symptoms were: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. That alternative rests on the principle: Link-wide pause can throttle all traffic classes and is not a substitute for class-scoped priority flow control and congestion feedback. The question instead asks for evidence about: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The discriminating observations here are: Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class.

B:

These observations investigate ets bandwidth for rdma and storage; they would be useful if the symptoms were: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. That alternative rests on the principle: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. The question instead asks for evidence about: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The discriminating observations here are: Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class.

D:

These observations investigate pfc head-of-line blocking; they would be useful if the symptoms were: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. That alternative rests on the principle: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. The question instead asks for evidence about: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The discriminating observations here are: Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class.

Question 158

At a low-latency GPU workload fabric, the data-center team needs to explain this behavior: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. One component outage must remain inside the agreed availability budget. Which engineering principle best explains the observed behavior?

  1. End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path.
  2. PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking.
  3. ECN can signal endpoints earlier than PFC; correctly placed ECN and PFC thresholds may improve stable no-drop transport.
  4. Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class.

Explanation

Correct Answer

A:

The controlling technical principle is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path. This matters because the reported design failure is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The first diagnostic check is: Capture DSCP and CoS before and after the suspect hop The independent check is: Compare egress queue and drop counters to the intended class. These facts support the targeted intervention: Trace the packet's markings and class mapping across every hop and align supported QoS policy. Confirm that the resulting design meets this condition: RDMA packets traverse the expected protected class on every transit device.

Incorrect Answers

B:

The offered principle describes pfc head-of-line blocking: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. That is useful for a distinct incident, such as: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. It does not account for the reported failure: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The governing principle in this situation is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path.

C:

The offered principle describes ecn marking before pfc: ECN can signal endpoints earlier than PFC; correctly placed ECN and PFC thresholds may improve stable no-drop transport. That is useful for a distinct incident, such as: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. It does not account for the reported failure: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The governing principle in this situation is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path.

D:

The offered principle describes ets bandwidth for rdma and storage: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. That is useful for a distinct incident, such as: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. It does not account for the reported failure: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The governing principle in this situation is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path.

Question 159

At a two-region enterprise VXLAN fabric, the team completed an authorized repair for: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The architect must preserve the intended failure domains. Which measured result most directly proves the original requirement is satisfied?

  1. RoCE loss goals are met without unnecessarily suspending all traffic on the link.
  2. A pause storm cannot indefinitely block the rest of the fabric.
  3. RDMA packets traverse the expected protected class on every transit device.
  4. Congestion is relieved primarily by rate feedback rather than persistent PFC deadlock.

Explanation

Correct Answer

C:

The relevant measurable success criterion is: RDMA packets traverse the expected protected class on every transit device. It addresses the original reported defect: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. A saved design or configuration command alone does not establish that outcome. Observe the affected system in two independent ways: Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class. The correction being evaluated is: Trace the packet's markings and class mapping across every hop and align supported QoS policy. The underlying reason that correction matters is: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path.

Incorrect Answers

A:

This metric would help validate a correction to avoiding global pause for ai traffic, rather than the original reported issue. The other issue is: Operators configure IEEE 802.3x global pause for all leaf interfaces to protect one RoCEv2 traffic class. Here the defect to be resolved is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The expected measurable outcome for the actual incident is: RDMA packets traverse the expected protected class on every transit device. A positive result for the other service would not prove this requirement is met.

B:

This metric would help validate a correction to pfc watchdog failure handling, rather than the original reported issue. The other issue is: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. Here the defect to be resolved is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The expected measurable outcome for the actual incident is: RDMA packets traverse the expected protected class on every transit device. A positive result for the other service would not prove this requirement is met.

D:

This metric would help validate a correction to ecn marking before pfc, rather than the original reported issue. The other issue is: A RoCEv2 cluster responds to congestion only after repeated PFC pause storms occur. Here the defect to be resolved is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The expected measurable outcome for the actual incident is: RDMA packets traverse the expected protected class on every transit device. A positive result for the other service would not prove this requirement is met.

Question 160

At a 60-rack compute cluster, the architecture standards group considers the recurring risk: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The team needs a defensible design decision rather than a vendor feature checklist. Which continuing design control best prevents the same problem from recurring?

  1. Alert on prolonged PFC duration and remediate misbehaving NIC firmware.
  2. Review QoS shares when application proportions change.
  3. Recheck DSCP and 802.1p mappings when new services are added.
  4. Validate end-to-end QoS maps as part of leaf software upgrades.

Explanation

Correct Answer

D:

The most relevant ongoing safeguard is: Validate end-to-end QoS maps as part of leaf software upgrades. The risk it addresses is: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. This check is tied to the technical principle: End-to-end QoS needs consistent DSCP/CoS trust and queue mappings; a local correct mark does not ensure a lossless path. The immediate correction remains: Trace the packet's markings and class mapping across every hop and align supported QoS policy. Periodically validate the safeguard through two different observations. Capture DSCP and CoS before and after the suspect hop Compare egress queue and drop counters to the intended class. The continuing acceptance target is: RDMA packets traverse the expected protected class on every transit device.

Incorrect Answers

A:

This is a continuing control for pfc watchdog failure handling; its underlying principle is: Prolonged priority pause can cause PFC deadlock; a supported watchdog can detect and break pathological pauses. The recurring risk here is different: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A host repeatedly sends pause frames on one no-drop class and the entire storage network eventually stalls. The intended long-term acceptance condition remains: RDMA packets traverse the expected protected class on every transit device.

B:

This is a continuing control for ets bandwidth for rdma and storage; its underlying principle is: Priority-based flow control does not guarantee sufficient scheduler share or sustained throughput for a congested class. The recurring risk here is different: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A no-drop queue receives almost no scheduler capacity during peak traffic despite correct PFC setup. The intended long-term acceptance condition remains: RDMA packets traverse the expected protected class on every transit device.

C:

This is a continuing control for pfc head-of-line blocking; its underlying principle is: PFC pauses at priority level and poorly scoped no-drop classes can propagate congestion and head-of-line blocking. The recurring risk here is different: An application marks RoCE packets correctly at the host but a transit leaf remaps them to the default lossy queue. The appropriate ongoing control must prevent recurrence of that particular failure, not just the separate issue of: A lossless Ethernet design enables PFC on all priorities and a paused storage queue blocks latency-sensitive control traffic. The intended long-term acceptance condition remains: RDMA packets traverse the expected protected class on every transit device.

Designing Cisco Data Center Infrastructure for Traditional and AI Workloads (300-610) Practice Tests