POSTMASTER@SUNRISE.BITNET (Edward A. Lyon) (02/13/90)
These nodes were in the LINKS002 file but NOT in the NETINIT file:
DBNVB12, HROEUR1A, HROEUR1B
These nodes were in the NETINIT file but NOT in the LINKS002 file:
AWIIEZ44, DS0IFU57, FIRIEN, FRNSAE11, HEARNVAX,
HEITUE52, NOBSCVM, TNEARN, TRHUN, VERS9002
Aside from the comments and the node order, those were the only differences.mwh@IVORY.EDUCOM.EDU (Michael Hrybyk) (02/13/90)
> > These nodes were in the LINKS002 file but NOT in the NETINIT file: > > DBNVB12, HROEUR1A, HROEUR1B > > These nodes were in the NETINIT file but NOT in the LINKS002 file: > > AWIIEZ44, DS0IFU57, FIRIEN, FRNSAE11, HEARNVAX, > HEITUE52, NOBSCVM, TNEARN, TRHUN, VERS9002 > > > Aside from the comments and the node order, those were the only differences. > The LINKS file is generated without benefit of the full set of EARN updates for the version in question. The differences you see are a result of this. The deletes/adds listed above are those that EARN initiated before BITEARN NODES was finalized but after updates were submitted to BITNIC. This scheme has obvious problems. Note that the LINKS file was generated based on routing produces by UCLA. The file is invalid for the topology generated by GENROUTS or FastRout/pathalias, as paths from node to node are not represented solely by the :via tag. Since LINKS shows the same picture as seen by the UCLA routing tables and the :via tag, it should be based on the same data. This explains why it is slightly out of sync with BITEARN NODES. Data sent to Chris at UCLA represents the state of EARN a month earlier. Luckily, after 9003, both the :via tag and the LINKS file lose usefulness, as the entire net topology (well, almost) will be set by GR or pathalias. Any summary report can then be generated correctly from BITEARN NODES alone. This may be a good time to discuss the type of summary report that would be useful. What fields should it contain? Who or what applications would use it? Should we simply beef up the NODES INFO* files? Should the LINKS file be discontinued? Send your suggestions to me (this addr or HRYBYK@BITNIC), and I will summarize to the net. Mike Hrybyk BITNIC HRYBYK@BITNIC mwh@educom.edu