Measure 6LoWPAN addresses and the fragment gate

Steps

Screens captured against Contiki-NG Cooja Contiki-NG release/v4.9 c20b12cd, Cooja 4.8 (2022071901), Temurin 17.0.20.1, image sha256:b14d5afb0a2a852d9e72bbed607ead03f6ccc5f4d4dee44f0799110be2f4b609 on 2026-09-10; the tool may have moved on — the text steps are the contract.

1. 1 Step 1

   Do

   Run the primary CSC until 125 s. Filter Mote output for ^-- (fd00|fe80)::202|^-- Prefix:. Record the receiver addresses and root prefix.

   You will see

   Receiver 2 has fd00::202:2:2:2 and fe80::202:2:2:2; the root advertises fd00::/64.

   Why it matters

   The routed lab prefix and link-local scope have different owners and reach. This run covers the radio mesh only.

2. 2 Step 2

   Do

   Filter for output: header len 56|uncompression: UDP length: 60. Compare the sender’s frame budget with the receiver’s rebuilt UDP header.

   You will see

   Headers shrink from 56 to 15 bytes; the 108-byte IPv6 packet becomes 67 bytes against a 104-byte MAC payload budget. frag_needed is 0; UDP length is 60.

   Why it matters

   The 56 bytes include IPv6, UDP and the RPL extension. Compression is measured for this exact context and path, not a universal fixed saving.

3. 3 Step 3

   Do

   Filter for ^Data .*length 52|^-- MOP:. Match the accepted length to the original 52-byte report.

   You will see

   Data 1 received length 52 appears; the root reports Non-storing.

   Why it matters

   Receiver delivery is the second side of the trace pair. This upstream size-test fixture has arbitrary payload bytes, not encoded environmental sensor readings.

4. 4 Step 4

   Do

   Open cooja-6lowpan-addressing-large.csc from the packet. After it stops, filter for output: header len 56|output: fragmentation needed.

   You will see

   The 300-byte payload makes a 356-byte IPv6 packet, compressed to 315 bytes. Four fragments are required.

   Why it matters

   Growing the same message contract crosses the fragmentation gate even though addresses and radio range stay fixed.

5. 5 Step 5

   Do

   In the large run, filter for input: fragment \(tag|^Data .*length 300. Compare the fragment rows with the accepted payload.

   You will see

   Four input fragment rows precede Data 1 received length 300. Offsets are 0, 136, 232 and 328.

   Why it matters

   One delivered datagram depends on all its fragments; offsets refer to the reconstructed datagram, not four separate application messages.

6. 6 Step 6

   Do

   Open cooja-6lowpan-addressing-loss.csc. Filter for received first element|input: fragment \(tag|^Data . Check expected-loss-output.txt and write the README release decision.

   You will see

   Only the first fragment arrives. The script moves receiver 2 out of range at 80.559936 s; no Data delivery occurs through 125 s.

   Why it matters

   The loss case passes only when delivery is absent. Keep routine reports small; name a fallback and retest after size, context, prefix or route changes.