Participant Relationship Protocol

Back to matrix

G05-ND: Pending

Rationale

The retained G01 construction uses NFD 24.07 multicast v5 over a tree, with / listing all neighboring faces and only local producer registrations. Its source forwards a new Interest to every eligible next hop. For distinct globally scoped names absent from caches and pending state, with leaf producers, this floods each tree edge once. Thus a leaf interface receives requests originated by all other leaves, even when it is not their destination. At ten billion active leaves and 1000 fresh requests per leaf/day, this is about 115.74 million incoming Interests/second per leaf. Counting only four bytes per Interest already gives 3.70 Gbit/s, before names, other fields, framing or Data. This is a conditional lower bound for that construction and workload, not a measured NFD rate.

Assessment method

Pinned forwarding semantics and conditional traffic lower bound; no runtime measurement

Source context

NFD 24.07 multicast v5; existing G01 witness; common workload arithmetic

Outstanding qualification

G05-ND remains Pending: this defeats using the existing flooding witness alone as a capacity argument, not every NDN routing policy. Fresh names and unsatisfied distinct requests specify the no-hit case rather than assigning an arbitrary average cache-hit rate. The common workload includes long-tail/churn sensitivity; the calculation is not a claim that all real traffic always misses caches. A positive witness needs selective prefix routing or another documented strategy on the same regional graph, with forwarding-state, update and placement costs. No new directory, NLSR, IP underlay or physical links are silently introduced. G01 reachability remains Yes. No daemon, global experiment or E04 measurement is claimed.

Public sources

Editorial evidence references

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