[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Comments on architecture doc, Sect. 5
- To: "Mark Day" <markday@cisco.com>, <cdn@ops.ietf.org>
- Subject: Re: Comments on architecture doc, Sect. 5
- From: Patrik Fältström <paf@cisco.com>
- Date: Sat, 4 Nov 2000 12:19:23 -0500
- Delivery-date: Sat, 04 Nov 2000 09:22:55 -0800
- Envelope-to: cdn-data@psg.com
At 15.03 -0500 00-11-03, Mark Day wrote:
> > ACCOUNTING CPGs retrieve the data from
> > SURROGATEs that collect and store the data locally. This interior
> > data may be collected from the SURROGATEs by ACCOUNTING CPGs using
> > SNMP or FTP, for example.
>
>Is it always a "pull" by the CPG? It seems reasonable that the data could
>also be deposited directly to a CPG by the interior ACCOUNTING SYSTEM.
In general, if you are able to let the decision whether this is push
or pull being something the people operating the system chooses, that
is a good idea.
You can _also_ allow a push/pull combination like what we used in
Whois++ and what the notify extension in DNS uses. The server which
have new data tell the other party that new data exists, and closes
the connection (or equivalent). It is then up to the other party to
pull the data itself.
This to minimize the risk of denial of service attacks on the
receiving party in a push system.
paf