ATM--is it at layer 2 or layer 3?
NetBandi
Can anyone explain in simple terms as to whether ATM corresponds to
layer 2 or layer 3 of OSI model.
Thanks in advance.
Bandi
Francois Labreque
NetBandi wrote:
>
> Hi all,
> Can anyone explain in simple terms as to whether ATM corresponds to
> layer 2 or layer 3 of OSI model.
Layer 2 and below.
However, it does use some layer3-like features such as making routing
decisions when setting up VCs.
--
Francois Labreque | //\\ Wear an ASCII ribbon!
flabreque | || ||
@ | \\// Support the campain
videotron.ca \\ against HTML e-mail
//\\ and news!
NetBandi
Hi all,
Thank you for posting your views on whether "ATm corresponds to layer 2 or layer 3 of
OSI."
I've received varied opinions and obviously very good ones from the intellects and
highly
experienced in the industry.
So, thought I should share this with you all and hence Iam posting all the collective
responses.
Thanks again for sharing..
Regards,
Nagendra Bandi
|| ||
:||: :||:
::||:: ::||::
.:::||:::. .:::||:::.
...::::::||::::::. . . . ..:::::||::::::...
cisco Systems
Empowering the Internet Generation
Nagendra Bandi
nba...@cisco.com
===========================================================================================
ORIGINAL POST:
===========================================================================================
===========================================================================================
From: car...@cisco.com
Organization: Cisco Systems
Bandi,
ATM as a layer 2 or 3 protocol, based on OSI, has been a long time debate.
ATM as a layer 2 can support protocols such as IP or IPX for the connectionless
addressing. In similarity with IP or IPX, ATM has its own addressing scheme that can be
used
by upper layer protocols that supports the ATM reference model. So, in short, ATM can
be positioned as a Layer 2 protocol like Ethernet or FDDI for the Data Link
functionality
or it can also support the addressing functionality that is needed in the layer 3. The
use of the
ATM addressing scheme is not readily practiced because its much too complex for the
average network
operators and takes up CPU and memory resources in the router if the ATM traffic is
either SAR or
tunneled through an IP network. So, IP + ATM are the current implementation or soon to
be
implemented because of the QoS (..and others features) that ATM provides and the ease
of addressing
that IP has provided over the years.
My 2 cents,
Carlos
car...@cisco.com
--------------------------------------------------------------------------------------------
From:djan...@cisco.com (David Janssens)
Organization:Cisco Systems, Inc.
Hi Bandi,
Good question ;-)
ATM doesn't really map very well to the OSI model - it is much
better mapped to the B-ISDN model.
You have to think of ATM as having two "planes": a Control plane
and a Data plane.
Data flows on established VC's and as such the Data plane corresponds
to layer 2 of the OSI model. ATM is thus purely layer 2 when you're
dealing with PVC's.
But ATM also brings you the possibility to establish SVC's based
on a dynamic routing and signalling protocol (PNNI) and it uses
its own 20-byte NSAP addresses. It is the Control plane which
allows you to establish these SVC's and this corresponds to the
layer 3 of the OSI model.
Regards,
/david
David Janssens .. .. Tel: +32 2 704 5621
djan...@cisco.com || || Fax: +32 2 704 5800
Network Engineer || || Mail: Cisco Systems
CCIE# 3151 |||| |||| Pegasus Parc
Escalation Team .:||||||:..:||||||:. De Kleetlaan, 6
EMEA TAC cisco Systems, Inc 1831 Diegem, Belgium
--------------------------------------------------------------------------------------------
From: David
Organization: Cisco Systems
Bandi,
The ATM architecture is based on a logical model, called the ATM reference
model, that describes the functionality it supports. In the ATM reference
model the ATM physical layer corresponds approximately to the physical
layer of the OSI reference model, and the ATM layer and ATM adaptation
layer (AAL) are roughly analogous to the data link layer of the OSI
reference model.
Hope this helps..
<<david>>
David G. Wilson
Solutions Area Manager - SouthEast
PS Solutions Engineering Team
http://wwwin.cisco.com/CustAdv/SM/ps/Set/
CCIE #4078 (Router/Switching)
Cisco Systems
Atlanta, GA.
Office / VM (678) 352-2608
Home Office (770) 489-4562
Mobile (770) 335-7362
Isaiah 6.8
---------------------------------------------------------------------------------------------
From: Bernie <Ber...@weekend.com>
Organization: Airnews.net! at Internet America
Both. An ATM vcc is layer 2, whether or not it is an SVC, PVC, or
SPVC. Now the protocol (Q.2931) that is used to setup and tear down
SVCs and reroute SPVCs is a layer 3 protocol. This is where the
concept of a multi-plane protocol stack comes from. The data plane of
ATM is entirely layer 2, while the control plane is layers 2 through
3. If you don't understand this concept, find a good ATM reference
guide or datacom book that spends some time explaining multi-plane
networking. It is a difficult concept to understand, so I won't
pretend to be able to explain it in the time I have available.
--Bernie
------------------------------------------------------------------------------------------
From:"Peter Mansell" <pman...@mail.ebbs.com.au>
If you read this Cisco article http://www.cisco.com/warp/public/614/12.html then you can
see the issue
is far from clear. I got this article from a www.google.com search for atm iso layer
by the way.
HTH
--------------------------------------------------------------------------------------------
From:"Dooby" <ayan...@home.com>
Organization: Excite@Home - The Leader in Broadband http://home.com/faster
Figure 8-5: Relationship of ATM functional layers to the OSI reference
model.
Layer 1 and parts of Layer 2, the article is here. Hope this helps.
http://www.cisco.com/univercd/cc/td/doc/cisintwk/idg4/nd2008.htm
-------------------------------------------------------------------------------------------
From: "Andrew" <raid...@usa.net>
Organization: Road Runner - NC
ATM has nothing to do with layer 1. It's a layer 2 protocol whose control functions
reside in layers 2 and 3.
------------------------------------------------------------------------------------------
From: Francois Labreque <flab...@videotron.ca>
Organization:Me? organized! You've got to be kidding!
Layer 2 and below.
However, it does use some layer3-like features such as making routing
decisions when setting up VCs.
--
Francois Labreque | //\\ Wear an ASCII ribbon!
flabreque | || ||
@ | \\// Support the campain
videotron.ca \\ against HTML e-mail
//\\ and news!
------------------------------------------------------------------------------------------
From: Udo Konstantin <Udo.Kon...@t-online.de>
Organization:T-Online
Hi Bandi,
ATM corresponds to the physical layer and part of the data link layer
of the OSI refernece model.
ATM refernece model is composed of the following ATM layers.
-physical layer
-ATM layer (L2 OSI) ->combined with the adapation layer
it is responsible for etablishing connections and passing cells through the
ATM network
-ATM adaption layer (L2 OSI)combined with the ATM layer the AAL is roughly analogous
to the data link layer of the OSI referce model.
More infos ??
buy
Udo
-----------------------------------------------------------------------------------------
From:"Dominic Richens" <dominic...@sympatico.ca>
Organization:Nortel
Yes.
It all depends on what you are using ATM for. IP networking people see ATM
as just another layer 2 transport, maybe a smart layer 2 in that it can hide
complex network topology from you and yet guarantee end-to-end QoS just the
same.
Applications that use ATM natively see ATM as both layer 2 and 3, since it
has a networking layer supported by AESA, PNNI, etc...
--------------------------------------------------------------------------------------------
From:me...@cisco.com
Organization: Cisco Systems
did you look at http://www.protocols.com//hot.htm
http://www.cisco.com/univercd/cc/td/doc/product/atm/c8540/12_0/13_19/atg/index.htm
Meir
--------------------------------------------------------------------------------------------
From: Afroz Lateef <afroz....@fnc.fujitsu.com>
There is not a true one-to-one mapping between ATM and OSI.
ATM can correspond to Layer 2, and somewhat layer 3 too.
Afroz Lateef
-------------------------------------------------------------------------------------------
From:Veronica Tebbe <vte...@cisco.com>
2.
-------------------------------------------------------------------------------------------
From:ch...@cisco.com
Organization: Cisco Systems
Bandi,
ATM as a technology really more relates to layer 1 and 2. ATM does not deal
with L3.
That's what things like MPOA and MPLS can give you over an ATM transport.
*** Cisco Confidential and Proprietary ***
Chris L. McGugan Phone: +1 919.392.3226
Cisco Systems, Inc. Pager: +1 800.365.4578
Sr. Manager, Marketing
IS BU FAX: +1 508.632.6450
7025 Kit Creek Road Cell: +1 919.349.0618
RTP, NC 27709 GSM: 011.44.8025.76052
Email: ch...@cisco.com E-page: ch...@epage.cisco.com
Administrative Assistant: Jessica Loiacono (jmonroe) +1 919.392.2130
------------------------------------------------------------------------------------------
From: Jerald Shetty <jsh...@cisco.com>
Organization: Cisco Systems
Believe me, it is not Layer-3. You can say it is Layer-2, because it defines
access mechanism in a layer below 3. RFC 1483 and 1577 simply define ways of
mapping your Layer-3 address to Layer-2 ATM address(NSAPs and VCs). In the
case of LANE ATM maps the MAC address to the NSAP address.
Could you explain why you are confused about whether ATM is Layer-2 or 3?
Jerry
Jerald Shetty Phone: 408 527-9720
Cisco Systems Fax: 408 526-6951
Customer Support Engineer Pager: 800-365-4578
San Jose TAC Email: jsh...@cisco.com
------------------------------------------------------------------------------------------
From: Prince A Jose <pj...@cisco.com>
layer 2
------------------------------------------------------------------------------------------
Hi,
ATM is a layer 2 protocol (like Frame Relay).
Over ATM you can carry IP (third layer protocol), or voice.
Enjoy the attached doc.
http://www.cisco.com/univercd/cc/td/doc/product/atm/c8540/12_0/13_19/atg/
Zvi Shacham
Manager, Product Marketing
CSE BU, Cisco Israel
Hamelacha 32, Tel:+972-9-8927181
Sapir Industrial Zone Mobile: +972-54-971610
Netania South, 42504, Israel Fax: +972-9-8927223
-------------------------------------------------------------------------------------------
===========================================================================================
From:NetBandi
Organization: Cisco Systems
So what do you think now ;-)
|| ||
:||: :||:
::||:: ::||::
.:::||:::. .:::||:::.
...::::::||::::::. . . . ..:::::||::::::...
cisco Systems
Empowering the Internet Generation
Nagendra Bandi
nba...@cisco.com
=============================================================================================
THANK YOU ALL
=============================================================================================
Hansang Bae
>Thank you for posting your views on whether "ATm corresponds to layer 2 or layer 3 of
>OSI."
>I've received varied opinions and obviously very good ones from the intellects and
>highly
>experienced in the industry.
>So, thought I should share this with you all and hence Iam posting all the collective
>responses.
>Thanks again for sharing..
[snip: good answers that were summarized (like a good OSPF networks
oughtta be]
But in the end, ISO's OSI is dead dead dead... so it's only important
for academic reasons. But methinks you asked for academic reasons so
you got your answers.
Thanks again for the summary.
hsb
"Somehow I imagined this experience would be more rewarding" Calvin
********************************************************************
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account. Please post a followup instead.
********************************************************************
p.zhong
I think it dependes on how you use the ATM in IP network.
Vincent C Jones
Hansang Bae <hbae_@_primenet.com.REMOVE_> wrote:
>
>But in the end, ISO's OSI is dead dead dead... so it's only important
>for academic reasons. But methinks you asked for academic reasons so
>you got your answers.
>
I fear you are guilty of "discarding the baby with the bath water." I
agree that ISO CLNS is dead, but...
* ISO OSI protocols such as ASN.1, IEEE 802, IS-IS, frame relay,
cell relay, SONET, HDLC, X.400, X.500, LDAP, X, SGML, SCSI, xDSL,
V.everything, standard character sets, and many others we take for
granted are still alive and kicking.
* The OSI Reference Model remains the only useful mechanism for
discussing protocol operations in a vendor neutral and well
understood manner.
* IPv6 could also be called CLNSv2. A little history lesson: CLNS
fixed the defects in IP identified by 1975, but not those recognized
since then. IPv6 fixes those in IP & CLNS identified by 1990. Neither
provides sufficient benefit to the typical end-user to cost justify
the migration away from IPv4.
* The OSI movement was critical to pushing TCP/IP to its current
position as the basis for commercial productivity. Remember,
outside of a few individuals on ego trips, the goal of OSI was
open systems unencumbered by proprietary boundaries. We have that
now and the next step is to evolve from legacy TCP/IP to protocol
independent applications running over IPv6. Rather than dead, OSI
has completed its initial task and moved on to the next challenge.
WARNING: The above are purely my own personnal opinions and others
in the community hold equally strong opinions which are different.
We unnecessarily suffered through an extra decade (or more) of closed
systems because individuals were more concerned with "who" defined
open systems than with the open systems that were being defined. And
yes, the IETF community contributed both positively and negatively,
as did the ISO OSI community.
BOTTOM LINE: The old adage still applies: "There is no limit to
what you can accomplish if you don't worry about who gets credit
for it." (My apologies to the original author, whom I've forgotten,
and the mangling of the quotation.)
Vincent C Jones
--
VCJ...@NetworkingUnlimited.com Author of the Addison-Wesley book
Computer Network Consultant High Availability Networking with
http://www.networkingunlimited.com Cisco, see the web link for info.
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 80,000 Newsgroups - 16 Different Servers! =-----
Hansang Bae
>In article <97hrta$fp3$1...@nnrp1.phx.gblx.net>,
>Hansang Bae <hbae_@_primenet.com.REMOVE_> wrote:
>>But in the end, ISO's OSI is dead dead dead... so it's only important
>>for academic reasons. But methinks you asked for academic reasons so
>>you got your answers.
vcj...@NetworkingUnlimited.com (Vincent C Jones) writes:
>I fear you are guilty of "discarding the baby with the bath water." I
>agree that ISO CLNS is dead, but...
>* ISO OSI protocols such as ASN.1, IEEE 802, IS-IS, frame relay,
> cell relay, SONET, HDLC, X.400, X.500, LDAP, X, SGML, SCSI, xDSL,
> V.everything, standard character sets, and many others we take for
> granted are still alive and kicking.
Hmm, I guess you're right. I never thought of these as part of the OSI
family. (Didn't Steve Austin and Jamie Sommers work for them! <G>)
>* The OSI Reference Model remains the only useful mechanism for
> discussing protocol operations in a vendor neutral and well
> understood manner.
[snip]
>BOTTOM LINE: The old adage still applies: "There is no limit to
>what you can accomplish if you don't worry about who gets credit
>for it." (My apologies to the original author, whom I've forgotten,
>and the mangling of the quotation.)
Or the other saying..... if you know what you're doing.... three layers
is enough! <G> What I really meant to say is that the 7 layer model is
an academic only discussion at this point in time. But I agree that
it's a good teaching tool.
Thomas Kerrick
> is enough! <G> What I really meant to say is that the 7 layer model is
> an academic only discussion at this point in time. But I agree that
> it's a good teaching tool.
I know you agree with him, but I just gotta toss in my few cents and I
may as well do it on your post. ;-P
As one who's only been out of technical college for 3.5 months, I'll
AGREE 100% that the ISO OSI model was and IS imperitive to not ONLY my
acadimic learning but my practical learning too. Learning about the OSI,
all the protocols themselves (that Vincent mentioned, etc) the 4 layer
TCP/IP model, etc helps one to have the capacity to think outside of the
model, outside of the pure.
;)
thomas