[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

snmpconf interim meeting script mib discussion26-27, 2000




>>>>> Dan Romascanu writes:

Dan> 1. Script MIB Discussion

[...]

I have read through the minutes and it looks like the discussion
somehow missed the point.

As I said before: I believe the granularity of the model you are using
is flawed. Modeling individual filters and actions as individual MIB
objects (that have to fit into SNMP message size constraints) will not
scale. My experience is that control programs tend to become complex
and lengthy over time. The benefits of being able to share some action
code between filters will soon be eaten up by the amount of additional
data required to implement the fine granular MIB objects. Furthermore,
once you have to deal with hundreds or even thousends of filters and
actions, you will experience typical SNMP efficiency problems when
installing/removing large sets of policies. In addition, you impose a
non-trivial burden on the manager (and the agent) to do lots of row
creation/deletion operations and the appropriate handling of error
cases.

In contrast, bundling associated filters and actions into a single
script which is transferred to the policy execution engine avoids all
these problems. And in this case, the script MIB becomes the natural
way to transfer, activate and control the execution of policy scripts.

Dan> Dave - trying to summarize - there is value in the model as
Dan> created by this WG, the generic Script MIB does not meet the
Dan> requirements of this model

I can agree that the Script MIB does not fit the requirements imposed
by the current model. But I also believe that the model itself is
flawed (see above) since it won't scale and it will be difficult to
implement and use.

/js

-- 
Juergen Schoenwaelder      Technical University Braunschweig
<schoenw@ibr.cs.tu-bs.de>  Dept. Operating Systems & Computer Networks
Phone: +49 531 391 3289    Bueltenweg 74/75, 38106 Braunschweig, Germany
Fax:   +49 531 391 5936    <URL:http://www.ibr.cs.tu-bs.de/~schoenw/>