Participant Relationship Protocol

Archived criterion

C08-GN: Yes

Archived previous-definition assessment — revision 4.72. Retired in 5.73; excluded from the active matrix and counts. This historical verdict is not a current assessment.

Rationale

Select legacy UDP TRANSPORT/CORE/CADET/R5N with proportional ATS, heap dhtcache, DISABLE_BF_RC=YES, peerinfo NO_IO=YES and logging NONE without forced overrides or optional traffic archives. NO_IO leaves the peer filename directory unset, so get_host_filename returns NULL and the HELLO write path is skipped. The heap cache and in-memory Bloom mode avoid the identified DHT disk artifacts. The inspected service/plugin statistics updates request nonpersistent entries; shutdown serializes only entries marked persistent. Start with fresh statistics state and no unrelated services contributing persistent records. The inspected legacy routing services retain operational state in memory rather than requiring a final-peer activity archive. This is a documentary capability witness with loss of learned state across restart, not the default configuration or a claim of zero transient state. Reload locally supplied contacts to resume service; the local node key/configuration are not final-participant communication history. TNG peerstore requirements are not imported into this legacy profile.

Assessment method

Pinned source and explicit configuration composition; no daemon execution or filesystem tracing

Source context

GNUnet 0.20.0 peerinfo, datacache, legacy transport, CORE, CADET, DHT, ATS, statistics and logging sources

Outstanding qualification

Launch only the pinned selected services with the stated effective settings and no forced logging override. Supply bootstrap contacts locally, relay CADET traffic, inspect writes during operation, clean shutdown and crash, then restart and restore contacts to demonstrate service without retained final-peer history. Include plugin loading, supervisor logs, dumps/swap and unselected services in the deployment audit; do not infer their configuration from daemon options. Confirm statistics records contain no persistent peer history. No runtime, crash or restart experiment performed.

Public sources

Editorial evidence references

The following identifiers refer to research records not included in this public package; they are not public links.