Past exam question: 2078 Bhadra, Q9(c)
Open in the Padhaiko appWrite short note on: Packet switching versus message switching
Solution
Architectural Concepts: Both packet switching and message switching operate on the store-and-forward principle, where intermediate network nodes receive, buffer, inspect, and route data without establishing a dedicated end-to-end physical circuit:
- Message Switching: The source transmits the entire block of user data as a single monolithic unit (a "message"). Intermediate switches must buffer the entire message before forwarding it to the next link.
- Packet Switching: The source decomposes large messages into small, standardized chunks called packets (typically bytes), prepending headers with sequence numbers and addresses. Packets are routed independently or via virtual circuits.
Comparison of Key Engineering Characteristics:
| Dimension | Message Switching | Packet Switching |
|---|---|---|
| Data Unit Size | Variable, arbitrarily large (full message) | Fixed maximum packet size (e.g., MTU ) |
| Pipelining Effect | None. Intermediate switch waits for entire message | High. Node 2 forwards packet 1 while receiving packet 2 |
| Switch Buffer Requirements | Extremely high (requires magnetic disks / large RAM) | Low (small internal RAM buffers) |
| End-to-End Latency | High (cumulative store-and-forward delays) | Significantly lower due to parallel link pipelining |
| Error Handling | If 1 bit is corrupted, the entire massive message is resent | Only the corrupted small packet is retransmitted |
| Suitability | Bursty, non-real-time messages (telegrams, early email) | Modern interactive internet traffic (web, voice, video) |
In message switching, an entire data file is forwarded as a single undivided message, causing high latency, large switch memory requirements, and inefficient error recovery. In packet switching, data is broken into small, bounded packets that pipeline concurrently across successive links, providing drastically reduced latency, lower switch memory demands, and fast retransmission of isolated corrupted packets.