Cameron Laird's personal notes on SNMPy

Flash! In spring 2001, Yves Perrenoud released 0.7 of YAPSNMP.

I use both Python and SNMP, so it's natural for me to have a stake in SNMPy, an SNMP-aware Python extension.

Most of what I know about SNMPy's state as of 1999 appears in this Usenet posting:

As spring 2001 begins, Anthony has opened up a SourceForge project, and is actively coding against Python 2.0 and net-snmp 4.2.

Why am I not launching a SIG and maintaining SNMPy? I'm already overcommitted on public works. SNMPy would need to show up on the for-fee side of my business for me to invest much more time in it during 1999.

Good news! There's an alternative; in October 1999, Ilya Etingof has released a preliminary pysnmp module (patience; it appears to be a distant server) written in pure Python. In April 2000, it has BER {en,de}coders and transport methods sufficient for SNMPv1 requests: GET (in a v1 sense), GETNEXT, and SET. However, as Ilya has not yet had time to teach it about MIBs, all OIDs must be specified numerically. In May 2000, it's hoped that SNMPY will appear in a more complete form at SourceForge.

My advice to Pythoneers who need to get real network management accomplishment in finite time is generally to exploit Scotty. It's far more mature than snmpy is likely to be in the foreseeable future, and it's possible to develop quite useful applications with even a small amount of Tcl knowledge. Moreover, it's possible to imagine incremental migration from a Tcl base to a Python one. Scotty's inventor, who has wide knowledge of network management issues, is aware of Python, and has considered its adoption. While he is in no position to start new development, I expect he'd be as generous in reviewing others' designs and implementations as he has been in similar circumstances in the past. He even has specific advice: he urges anyone launching an SNMP-for-Python project to rely on libsmi for MIB parsing ...


Cameron Laird's personal notes on SNMPy/claird@phaseit.net