Jev Networking

JEV-IDS is an open experimental network intrusion detector that serializes one flow per request and asks Jev for a binary attack probability and a finite-choice traffic category. On a 300-flow NSL-KDD pilot it reports F1 0.859 and novel-attack recall 0.838.

A first empirical study of Jev on application classification for network flows, using ten packet-header features from CESNET-QUICEXT-25. Forty labeled examples raise accuracy from 9.80% to 28.42%, while tree ensembles trained on 8,000 records reach about 67-70%.

A comparison of Jev-1.13.0 and two other decision models with LLMs for intent interpretation in 6G Open RAN. Jev meets the 1 s near-real-time budget on 99.8% of calls while two hosted LLMs meet it on 17.9% and 0%, with median interpretation taking 0.286 to 2.35 s.

An integration of Jev's decision API into edge service orchestration that extracts four to eight bounded intent fields per request. Across 33 test conditions Jev cuts median decision latency by 22.7-64.5% versus the fastest LLM, and on a live admission path it keeps 0.91-0.95 of requests exact and on time.