Agilent Labs' RIPE NCC Data Analysis project
Agilent Labs' RIPE NCC Data Analysis project
The objective of Agilent Labs'
RIPE NCC data analysis project is to visualize and characterize Border
Gateway Protocol (BGP4) behavior. The work is based on protocol traffic
collected at exchange points around the world by RIPE NCC's RIS.
The project has two tracks: daily analysis and long-term
pattern discovery. Initial observations indicate an unusally
large amount of duplicate, identical announcements and flaps
(without attribute changes).
The following attributes are taken into consideration when
defining an announcement as "identical":
- ORIGIN
- LOCAL_PREF
- MULTI_EXIT_DISC (MED)
- NEXT_HOP
- AGGREGATOR
- ATOMIC_AGGREGATE
- AS_PATH
- COMMUNITY
The daily metrics that are currently provided:
- Announcements
- Duplicate Announcements
- Withdrawals
- Duplicate Withdrawals
- Flaps (without attribute changes)
Each analysis metric is provided per peer, per hour and per
exchange. They are computed for all peers at RIPE NCC (12 peers)
and full feed peers at LINX (1 peer) and AMSIX (3 peers).
Please send us
feedback.
You can find more information about BGP routing instability in the BGP Cheat Sheet.
RIPE NCC
AMS-IX
LINX
|