Re: [Idr] BGP MIB v2 input

In message <200703262016.l2QKGEVi013401@bright.research.att.com>
Bill Fenner writes:
>  
>  
> >Just wondering: the current BGP4 MIB already supports the routing 
> >databse (BGP4-MIB::bgp4PathAttrBest at least).  Are you proposing 
> >dropping the equivalent functionality from the new MIB or just some 
> >more exciting new variants?
>  
> That was my proposal, since getting a pathattr table that
> nicely handles IP and non-IP NLRI is ... confusing.  It also
> can result in less than ideal behavior when walking the whole
> MIB on a device that has hundreds of thousands of prefixes.

I completely missed the non-IP issue here with the need for a MIB
entry and sorting order covering both IP and non-IP.

> >An example: I currently use both IP-FORWARD-MIB and 
> >BGP4-MIB::bgp4PathAttrBest in SNMP scripts by parsing the config files 
> >(interface-related routing protocol config to be exact) and checking 
> >them against operational policy (e.g., certain kinds of interface 
> >addresses must be in OSPF; others must not be in OSPF; some others 
> >must be in BGP; static routes must be in BGP; etc.).
>  
> Interesting.  I was looking it from the point of view of not knowing
> exactly what prefix you were looking for (e.g., "show ip bgp 12.106.35.1"
> without knowing that it's in a /22).  The obvious thing to do is to grab
> the whole table (probably inadvisable); there are more subtle ways but
> none are very efficient.

More likely do GET on a 12.106.35.1/32, and decrement prefix length to
/31, etc, until something comes up.  Not terribly efficient either.
The route covering this could be 0/0.

To handle IP and non-IP cleanly you almost need an object oriented MIB
entry, where one order is used for IP and other types can define other
orders.  No such thing in SNMP AFAIK.

>   Bill

Curtis

_______________________________________________
Idr mailing list
Idr@ietf.org
https://www1.ietf.org/mailman/listinfo/idr