Compare RPL objective functions and candidate paths

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 MRHOF CSC to 180 s. Trace mote 5’s arrows in Network; predict whether it will choose relay 2 or 3.

   You will see

   Mote 5 uses 2 → 1 in this seed-1 run. Both candidate relays are within its 50 m range.

   Why it matters

   A visible path is an output; the candidate records explain why this output is defensible.

2. 2 Step 2

   Do

   Filter Mote output for ^fe80::20[23].*=> and compare the rpl-nbr snapshot at 140.065 s.

   You will see

   For mote 5, relay 2 shows 310, 159 => 469; relay 3 shows 286, 249 => 535.

   Why it matters

   The columns are neighbor Rank, link metric and resulting Rank. MRHOF uses accumulated link cost plus eligibility, freshness and hysteresis rules.

3. 3 Step 3

   Do

   Filter for ^-- OF:|^-- Preferred parent|^-- Rank:|^-- Hop rank increment. Check the chosen function and state at 141.065 s.

   You will see

   MRHOF reports hop-rank increment 128, parent fe80::202:2:2:2 and Rank 469.

   Why it matters

   A claimed objective function must match the running firmware. The project header compiles both supported functions and explicitly selects MRHOF.

4. 4 Step 4

   Do

   Open cooja-rpl-objective-functions-of0.csc. After it stops, use the candidate filter from step 2.

   You will see

   OF0 shows relay 2: 512, 187 => 1024; relay 3: 768, 181 => 1280.

   Why it matters

   The topology, seed and 0.80 receive-success setting are held fixed. Different control behavior changes the observed link samples; this is not a promise of identical packet histories.

5. 5 Step 5

   Do

   Use the state filter from step 3 in the OF0 run. Compare the selected parent and Rank with MRHOF.

   You will see

   OF0 reports increment 256 and Rank 1024, also through parent 2. Both headless runs deliver 78 requests and receive replies at all five clients.

   Why it matters

   Do not compare raw Rank numbers as a universal score across objective functions. These functions agree on the parent here; the measurements do not show a delivery winner.

6. 6 Step 6

   Do

   Filter for Received ping reply, then complete the README’s prefix, Dijkstra and distance-vector trace tables, including their link-removal cases.

   You will see

   OF0’s root ping returns ttl 63 and delay 94 ms; MRHOF’s log reports 100 ms. In the separate worksheet, Dijkstra improves R3 from cost 5 to 4 via R2/R4.

   Why it matters

   Packet success checks the selected RPL path. The paper traces preserve the original algorithm exercises without pretending that this firmware runs OSPF, RIP or DSDV.