Which of the following switching techniques is most suitable for interactive…
2011
Which of the following switching techniques is most suitable for interactive traffic?
Answer: C. Packet switching — Concept: A switching technique defines how a network allocates its links to a conversation and how a unit of data is moved from one node to the next. Circuit…
- A.
Circuit switching
- B.
Message switching
- C.
Packet switching
- D.
All of the above
Attempted by 376 students.
Show answer & explanation
Correct answer: C
Concept: A switching technique defines how a network allocates its links to a conversation and how a unit of data is moved from one node to the next. Circuit switching reserves a path end to end for the lifetime of a connection, message switching stores a complete message at each node before forwarding it, and packet switching breaks data into small units that each node forwards as soon as the unit has arrived. Interactive traffic - short request-response exchanges separated by idle think-time gaps, in which the user experiences the round-trip delay directly - is therefore served best by whichever technique keeps the per-exchange delay small without holding capacity reserved during those idle gaps.
Application - apply that criterion to each offered technique:
Characterise the traffic first. An interactive session (remote login, a database query and its response, a chat exchange) sends a few hundred bytes at a time, in bursts, and the idle periods between bursts are orders of magnitude longer than the transfers themselves.
Circuit switching: a setup phase must complete before the first bit travels, and once the circuit exists the reserved capacity stays committed to it through every idle gap. A conversation that is idle for most of its lifetime therefore pays for capacity it is not using, and if the circuit is instead released and re-established for each exchange, the setup latency is paid again every time.
Message switching: each node must receive the whole message before it begins forwarding it, so the end-to-end delay is the sum over all hops of the full message transmission time, plus queueing at each node. That delay grows with both message size and hop count and is highly variable.
Packet switching: when a longer message is divided into multiple small packets, transmissions of successive packets can overlap across consecutive links after the pipeline fills. For a message of m packets over h equal-rate store-and-forward links, ignoring propagation, headers and queueing, serialization takes about (m + h - 1) packet-transmission times; a one-packet exchange gains no inter-packet pipeline advantage. The decisive advantage over message switching for interactive traffic is packetization: small bounded units can be interleaved with other flows and forwarded before a whole long message is received. Statistical sharing also avoids the dedicated idle-capacity reservation of circuit switching.
Cross-check - the three families side by side:
Technique | Resource allocation | Delay for one short exchange |
|---|---|---|
Circuit switching | Path reserved end to end for the whole call | Setup phase first, then a very low transfer delay |
Message switching | One hop at a time; the whole message is buffered at every node | Sum of the full message transmission times over all hops |
Packet switching | Links statistically shared; nothing reserved between bursts | Multi-packet transfers can pipeline across hops; one-packet transfers still incur per-hop store-and-forward delay. |
The stem asks which technique is most suitable - a superlative that has to resolve to a single technique, so a combined choice covering all three cannot answer it. Packet switching is the technique most suitable for interactive traffic: its small, independently forwarded units keep the response delay low, and statistical multiplexing means the bursty, mostly-idle nature of interactive traffic costs the network nothing between bursts. This is why interactive services - remote login, web request-response, and messaging - run over packet-switched networks.