Compare RPL mesh evidence with Thread claims

Steps

Screens captured against Contiki-NG Cooja Contiki-NG release/v4.9 c20b12cd, Cooja 4.8 (2022071901), Temurin 17.0.20.1, image sha256:b14d5afb0a2a852d9e72bbed607ead03f6ccc5f4b609 on 2026-10-08; the tool may have moved on — the text steps are the contract.

1. 1 Step 1

   Do

   Open the CSC in Cooja and let its GUI pause at 89.065 simulated seconds. Trace root 1, routers 2 and 3, and end devices 4 and 5 in Network.

   You will see

   The running border router receives the fd00::/64 SLIP prefix at 2.065 s. Network shows mote 4 through relay 3 and mote 5 through relay 2, both toward root 1. The source log records #L parent annotations from the actual motes.

   Why it matters

   This establishes a before state for a controlled relay fault. Cooja runs Contiki-NG RPL over simulated 802.15.4; this is not a Thread role view.

2. 2 Step 2

   Do

   In Mote output, filter for Received ping reply|Preferred parent and press Enter. Record both leaf replies and mote 5’s baseline parent.

   You will see

   Mote 4 receives an ICMP reply from fd00::201:1:1:1 in 112 ms, TTL 63. Mote 5 receives the same root address in 104 ms, TTL 63. At 85.065 s, mote 5 reports preferred parent fe80::202:2:2:2.

   Why it matters

   These real stack replies prove sampled IPv6 reachability before the fault. They do not prove a Thread child attachment or a Matter session.

3. 3 Step 3

   Do

   Bring Radio messages to the front and select a captured row. Read its time, From, To, byte count and hexadecimal packet data.

   You will see

   At 85.863 s the real Cooja radio list shows packet 239, From 3 to [3 d]. Its selected packet data is 5 bytes: 02 00 1C 55 6F.

   Why it matters

   These bytes come from Cooja’s simulated radio during the firmware run. A short 802.15.4 row is protocol evidence, not a decoded Thread MLE or Matter exchange.

4. 4 Step 4

   Do

   Minimize Radio messages and click Start/Pause once to continue past the scripted fault. At the 210-second stop, compare the Network arrows and the change timestamps in expected-output.txt.

   You will see

   The script removes preferred relay 2 at 90.065000 s. Mote 5 logs parent switch fe80::202:2:2:2 → fe80::203:3:3:3 at 101.046728 s. The measured change interval is 10.981728 simulated seconds; both leaves now use relay 3.

   Why it matters

   The new parent is an observed RPL routing event after a real simulated link intervention. It is not evidence of Thread MLE reattachment or physical radio resilience.

5. 5 Step 5

   Do

   Filter Mote output for Received ping reply after the run stops. Compare each post-fault reply against that same mote’s baseline.

   You will see

   Mote 4 replies at 180.126952 s with delay 61 ms and TTL 63. Mote 5 replies at 190.153952 s with delay 88 ms and TTL 63. The four displayed replies confirm two sampled paths before and after the fault.

   Why it matters

   Mote 4 is the unaffected comparison; mote 5 had to change its next hop. Reply times are individual simulated samples, not a service guarantee or loss-free proof.

6. 6 Step 6

   Do

   Filter Mote output for Preferred parent|-- OF|-- MOP|-- Rank. Fill the README comparison worksheet and classify each claim by protocol layer.

   You will see

   Mote 5 reports MOP Non-storing and OF MRHOF in both snapshots. Its preferred parent changes 2 → 3 while Rank changes 422 → 416. The headless ScriptRunner ends TEST OK; the stock UDP client still has no echo service.

   Why it matters

   These are Contiki-NG RPL fields and actual simulator output. A Thread claim needs separate MLE role/parent, dataset and Border Router traces; a Matter claim needs application and commissioning traces.