200-301 Select Two (Multi-Select) Questions

Explanation: Option B is correct because runts—frames smaller than 64 bytes—often result from collisions truncating frames on half-duplex links or a faulty NIC. Option D is correct because duplex mismatch can cause late collisions, which appear as output errors in 'show interface'; a device on one side full-duplex and the other half-duplex leads to collisions and framing errors. Option A is wrong because CRC errors can stem from faulty cabling, interference, or a mismatched NIC, not exclusively a bad switch port. Option C is wrong because giants (frames over maximum MTU) may be forwarded if the interface is configured with jumbo frames or the switch is set to accept oversize frames. Option E is wrong because 'show controllers' displays frame-size errors like runts and giants, including details beyond CRC errors.

Explanation: Option A is correct because supervised learning uses labeled data to classify traffic (e.g., voice, video, data). Option B is correct because anomaly detection often uses unsupervised learning to identify deviations from normal behavior. Option C is wrong because reinforcement learning is not used for spam classification; that task uses supervised learning. Option D is wrong because clustering groups data but cannot predict exact bandwidth usage; prediction requires regression models. Option E is wrong because predictive analytics in network operations leverages machine learning models, not solely static thresholds defined by administrators.

Explanation: Option B is correct because the 'switchport voice vlan' command assigns a dedicated VLAN for voice traffic, allowing the IP phone to tag its packets with the voice VLAN ID while the PC remains in the native (data) VLAN. Option D is correct because 'mls qos trust cos' preserves the Layer 2 Class of Service (CoS) markings from the IP phone, ensuring voice packets receive appropriate QoS treatment across the network. Option A is incorrect because a trunk port is not required—the access port with the voice VLAN command handles both VLANs without trunking. Option C is incorrect because disabling Spanning Tree Protocol (STP) is not a recommended practice and does not prevent voice delays; STP is essential for loop prevention and can be tuned with PortFast instead. Option E is incorrect because the port must remain a Layer 2 access port, not a routed port, to support both the PC and IP phone.

Explanation: Option B is correct because the IPv6 default static route uses the destination prefix ::/0, which matches all IPv6 addresses, similar to 0.0.0.0/0 in IPv4. Option D is correct because a floating static route is configured with a higher administrative distance and only becomes active when the primary route (with a lower AD) is removed or fails. Option A is wrong: a floating static route is configured with a higher administrative distance, not lower. Option C is wrong: a directly connected route has an administrative distance of 0, which is always preferred over a static route (even with AD 1). Option E is wrong: IPv4 static routes use the 'ip route' command, while IPv6 static routes use the 'ipv6 route' command; the syntax is different.

Explanation: Options A, B, C, and D are correct. AI models can predict bandwidth congestion by analyzing historical traffic data, enabling proactive capacity planning. Natural language processing (NLP) automates helpdesk ticket responses by interpreting user intent, reducing manual effort. Reinforcement learning can dynamically adjust firewall rules in response to evolving attack patterns, improving threat response without human intervention. Unsupervised learning can cluster behavior patterns to identify unknown device types on the network. Option E ("AI eliminates the need for baseline performance metrics because it learns in real-time") is incorrect because even AI/ML models require baseline metrics to establish normal behavior and detect anomalies; real-time learning does not remove the need for baselines. Option F ("ML models always require labeled training data to be effective in network operations") is incorrect because many ML techniques, such as unsupervised learning (as shown in option D), operate effectively on unlabeled data by discovering patterns and clusters without predefined labels.