[Fwd: Connections to switches]
Peter Jones <peter@atm-net.com.au> Mon, 28 July 1997 02:30 UTC
Received: from cnri by ietf.org id aa18563; 27 Jul 97 22:30 EDT
Received: from thumper.bellcore.com (thumper.bellcore.com [128.96.41.1]) by cnri.reston.va.us (8.8.5/8.7.3) with ESMTPid WAA14861 for <ietf-archive@cnri.reston.va.us>; Sun, 27 Jul 1997 22:28:50 -0400 (EDT)
Received: from info.curtin.edu.au (root@info.curtin.edu.au [134.7.70.222]) by thumper.bellcore.com (8.8.6/8.8.6) with ESMTP id VAA02506 for <atommib@thumper.bellcore.com>; Sun, 27 Jul 1997 21:57:14 -0400 (EDT)
Received: from atm-net.com.au (phillipe.curtin.edu.au [134.7.130.54]) by info.curtin.edu.au (8.8.5/8.8.5) with SMTP id JAA16496; Mon, 28 Jul 1997 09:56:59 +0800 (WST)
Received: from josef by atm-net.com.au (SMI-8.6/SMI-SVR4) id JAA00732; Mon, 28 Jul 1997 09:59:43 +0800
Sender: peter@phillipe
Message-ID: <33DBFD0E.514@atm-net.com.au>
Date: Mon, 28 Jul 1997 09:59:42 +0800
From: Peter Jones <peter@atm-net.com.au>
Organization: Atmosphere Networks
X-Mailer: Mozilla 3.0 (X11; I; SunOS 5.5.1 sun4m)
MIME-Version: 1.0
To: atommib@thumper.bellcore.com
CC: kaj@cc.bellcore.com, John Neil <jn@atm-net.com.au>
Subject: [Fwd: Connections to switches]
Content-Type: multipart/mixed; boundary="------------76C662B57D85"
Hi There, We have found that <draft-ietf-atommib-atm1ng-03.txt> is unclear about the usage of TX and RX when defining the traffic descriptor usage for connections via the Proprietary Virtual Interface to the AAL5 entity in a switch. I requested clarification from Kaj in the attached mail. His response was that he thought the directions should be the same as any other interface to the switch fabric, i.e. case 1) as defined in the mail. I would like this clarified in the RFC by adding some text like: Connections to the Proprietary Virtual Interface are treated like a normal (e.g. UNI) interface. The TX and RX indexes in the VPL or VCL Entry, (e.g. atmVclReceiveTrafficDescrIndex and atmVclTransmitTrafficDescrIndex) apply from the point of view of the ATM switching fabric (i.e. the ATM entity in Figure 7), i.e. the atmVclReceiveTrafficDescrIndex defines the traffic parameters that apply to the flow from the AAL5 entity to the ATM Entity, not the traffic flow from ATM Entity to the AAL5 entity. RxTrafficDescrIndex ATM Entity <--- AAL5 Entity TxTrafficDescrIndex ATM Entity ---> AAL5 Entity What this means is that the traffic parameters appear to be normal from the switch's "switch fabric" point of view, and reversed from the switch's AAL5 Entity (e.g. Management CPU) point of view. Regards Peter Jones -- ______________________________________________________________________ Peter Jones Atmosphere Networks Project Manager - NMS GPO Box U 1987 Direct:+61 (8) 9266 3581 Perth WA 6845, Australia Mobile: Tel: +61 (8) 9266 2922 Email: peter@atm-net.com.au Fax: +61 (8) 9266 3244 _______________________________________________________________________
--- Begin Message ---Hi Kaj, We have a question about the details of building connections to switches. Reference the following picture from atm1ng-03. ___________________________ | | | ============= | | | AAL5 | | | | Entity | | | ============= | | | | | -----Prop. Virtual Interface | | | | ============= | | | ATM | | | | Entity | | | ============= | |_____|__|__|__|__|_______| | | | | | ---------------- ATM UNIs | | | | | | | | | | v v v v v Figure 7: Model of an AAL5 Entity in a Switch What is the interpretation of the "direction" on the VCL or VPL table entries that relate to the "Prop. Virtual Interface"? There are two possibilities: 1) Treat the "Prop. Virtual Interface" like a normal UNI interface. In this case, the TX and RX indexes, e.g. atmVclReceiveTrafficDescrIndex and atmVclTransmitTrafficDescrIndex effectively apply from the point of view of the ATM switching fabric (i.e. the ATM entity in the figure above), i.e. the atmVclReceiveTrafficDescrIndex defines the traffic parameters that apply to the flow from the AAL5 entity to the ATM Entity, not the traffic flow from ATM Entity to the AAL5 entity. RxTrafficDescrIndex ATM Entity <--- AAL5 Entity TxTrafficDescrIndex ATM Entity ---> AAL5 Entity 2) Treat the "Prop. Virtual Interface" as a special case switch interface. In this case, the TX and RX indexes, e.g. atmVclReceiveTrafficDescrIndex and atmVclTransmitTrafficDescrIndex effectively apply from the point of view of the AAL5 Entity, i.e. the atmVclReceiveTrafficDescrIndex defines the traffic parameters that apply to the flow from the ATM Entity to the AAL5 entity, not the traffic flow from AAL5 Entity to the ATM entity. RxTrafficDescrIndex ATM Entity ---> AAL5 Entity TxTrafficDescrIndex ATM Entity <--- AAL5 Entity I can come up with arguments to justify either approach. What I want to know is what the "standard" approach is (which should be very clear from the RFC's). I have had a look at the drafts, but I can't see anywhere where this is defined. Regards Peter Jones -- ______________________________________________________________________ Peter Jones Atmosphere Networks Project Manager - NMS GPO Box U 1987 Direct:+61 (8) 9266 3581 Perth WA 6845, Australia Mobile: Tel: +61 (8) 9266 2922 Email: peter@atm-net.com.au Fax: +61 (8) 9266 3244 _______________________________________________________________________--- End Message ---
- [Fwd: Connections to switches] Peter Jones