RE: [ipcdn] Comments on draft-ietf-ipcdn-pktc-eventmess-04.txt

["Sumanth Channabasappa" <sumanth at cablelabs.com>]

Greg,

Thanks for the thoughts and unless vendors/implementors feel strongly
about local throttling, I would be inclined to leave it out scope for
now (akin to other IPCDN MIBs - as you rightly indicated). 

- S 

-----Original Message-----
From: Nakanishi Greg-MGI8179 [mailto:gnakanishi at motorola.com] 
Sent: Wednesday, October 19, 2005 3:18 PM
To: Sumanth Channabasappa; ipcdn at ietf.org
Subject: RE: [ipcdn] Comments on draft-ietf-ipcdn-pktc-eventmess-04.txt

Sumanth,

My opinion is that throttling should to be supported for the local log.
If the local log is small, there is a potential for the log to get
filled up with a flood of events and earlier events that may be
important for analysis gets overwritten.  

That being said, I can see that historically (in DOCSIS and PacketCable)
local log throttling was not specified.  I don't have particularly
strong feelings about this, but I think local log throttling would be a
desirable feature.

greg

-----Original Message-----
From: ipcdn-bounces at ietf.org [mailto:ipcdn-bounces at ietf.org] On Behalf
Of Sumanth Channabasappa
Sent: Wednesday, October 19, 2005 1:17 PM
To: ipcdn at ietf.org
Subject: RE: [ipcdn] Comments on draft-ietf-ipcdn-pktc-eventmess-04.txt

Randy,

Thanks and no, I did not miss your response. However, the MIB is geared
towards 'management events' (and their management) and the resulting
'notifications' are only a reporting feature. Thus, MTAs are required
to:
- store all relevant event information in the logs (storage allowing)
- send only critical (or desired) notifications, created upon event
generation to the external systems (syslog, SNMP)

Now, implementations could, as you suggested, have notifications
generated agnostic of the destination. 
However, in this case, the 'event generation and storage' is required
(it may overwrite earlier events, but current events are always stored),
with an option to convert them into 'notifications', if configured. 

Further, given the disparity between the Service Provider Network
elements (few) and the clients on which this MIB will be
implemented(many), the throttling requirements for 'notifications sent
to the network' were imposed. Currently, there has been no intention or
requirement to throttle events that can be stored locally.

Thanks for the pointer on RFC3014 - I shall take a look at it and see if
we can borrow any concepts.

Greg,
Barring the max size discussion (I should have clarified further, thanks
Eduardo), do we really see a need for throttling local events (would it
serve any purpose if the latest events are always stored)?
 

- S



-----Original Message-----
From: Randy Presuhn [mailto:randy_presuhn at mindspring.com]
Sent: Tuesday, October 18, 2005 8:36 PM
To: ipcdn at ietf.org
Subject: Re: [ipcdn] Comments on draft-ietf-ipcdn-pktc-eventmess-04.txt

Hi -

> From: "Sumanth Channabasappa" <sumanth at cablelabs.com>
> To: <ipcdn at ietf.org>
> Sent: Tuesday, October 18, 2005 2:58 PM
> Subject: RE: [ipcdn] Comments on
> draft-ietf-ipcdn-pktc-eventmess-04.txt
...
> I haven't received any comments on the clarifications requested (dated

> 10/07, enclosed in this email). Can I assume acceptance (status quo) 
> on
> #1 and #2? 

You must have missed my response from two months ago:

|From: ipcdn-bounces at ietf.org [mailto:ipcdn-bounces at ietf.org] On Behalf 
|Of Randy Presuhn
|Sent: Thursday, August 18, 2005 3:59 PM
|To: ipcdn at ietf.org
|Subject: Re: [ipcdn] Comments on draft-ietf-ipcdn-pktc-eventmess-04.txt
|
|Hi -
|
|> From: "Sumanth Channabasappa" <sumanth at cablelabs.com>
|> To: "Jean-Francois Mule" <jf.mule at cablelabs.com>; "Nakanishi 
|> Greg-MGI8179" <gnakanishi at motorola.com>; <ipcdn at ietf.org>
|> Sent: Thursday, August 18, 2005 11:46 AM
|> Subject: RE: [ipcdn] Comments on
|> draft-ietf-ipcdn-pktc-eventmess-04.txt
|...
|> 8) pktcDevEvThrottleAdminStatus
|> - The description makes mention of throttling events using SNMP and 
|> Syslog.  What about the local log?  Are events written to the local 
|> log not subject to throttling?
|...
|
|I'd suggest that if we don't leave it up to the implementation, that it

|would be best for throttling to affect all destinations equally.  
|Swamped logs are the same kind of problem, whether on the local system
or a remote one.
|
|Randy

The code issuing a notification, which is where the DESCRIPTION clause
would lead implementors to put the throttling logic, would in a
reasonably modular environment have no knowledge of whether it is going
to a local log, delivery over some protocol, or whatever.

Randy


_______________________________________________
IPCDN mailing list
IPCDN at ietf.org
https://www1.ietf.org/mailman/listinfo/ipcdn



_______________________________________________
IPCDN mailing list
IPCDN at ietf.org
https://www1.ietf.org/mailman/listinfo/ipcdn


_______________________________________________
IPCDN mailing list
IPCDN at ietf.org
https://www1.ietf.org/mailman/listinfo/ipcdn