JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD B.Tech III DEVOPS Previous Question Paper (March - 2024 - R18)

Identifying network bottlenecks during performance testing involves monitoring specific metrics and analyzing traffic patterns to pinpoint where data flow is restricted. The process typically includes:

• Monitoring Throughput and Bandwidth Utilization: Compare the actual data transfer rate against the maximum available network capacity. If utilization consistently hits 90-100% of the link capacity, the network is likely saturated.
• Analyzing Latency and Round-Trip Time (RTT): High RTT values often indicate congestion or inefficient routing. By measuring the time taken for a packet to travel from the client to the server and back, testers can distinguish between application processing time and network transit time.
• Packet Loss Analysis: Use diagnostic tools like ping, traceroute, or specialized network sniffers (e.g., Wireshark) to detect dropped packets. Significant packet loss is a primary indicator of buffer overflows in routers or switches.
• TCP Retransmission Rates: A high frequency of TCP retransmissions suggests that the network is unreliable or congested, forcing the protocol to resend data segments that failed to arrive or were corrupted.
• Queue Depth Monitoring: Check the interface queues on network devices. If queues are consistently full, it indicates that the device cannot process incoming traffic as fast as it is arriving, leading to delays and potential packet drops.

Tools such as Netstat, Iperf, and cloud-native monitoring services (e.g., AWS CloudWatch or Azure Monitor) are essential for capturing these metrics in real-time during load tests.

• Scrum is a structured, time-boxed framework that organizes work into fixed-length iterations called Sprints (typically 2-4 weeks), whereas Kanban is a continuous flow model that focuses on visualizing work and limiting Work-in-Progress (WIP) without predefined time-boxed iterations.
• In Scrum, roles are strictly defined, including the Scrum Master, Product Owner, and the Development Team. Kanban does not mandate specific roles, allowing teams to maintain their existing organizational structure.
• Scrum teams commit to a specific set of backlog items for the duration of the Sprint, and changes are generally discouraged once the Sprint begins. Conversely, Kanban allows for continuous delivery where items can be added or prioritized as soon as capacity becomes available.
• Performance metrics differ significantly: Scrum primarily uses Velocity (amount of work completed per sprint) to plan future iterations, while Kanban utilizes Cycle Time and Lead Time to measure the efficiency and speed of the delivery pipeline.
• Scrum requires specific ceremonies such as Sprint Planning, Daily Stand-up, Sprint Review, and Sprint Retrospective. Kanban is more flexible, with meetings often held on an as-needed basis or triggered by specific events in the workflow.