Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: [RPRWG] RPR and 1+1 linear APS like configuration.




Hello All,
	Perhaps I've missed something here - I understood that one of the
goals of RPR was to avoid the redundancy and reserved extra bandwidth that's
typically required when using 1+1 APR, BLSR or UPSR. Even with BLSR with
protection access, you use the extra bandwidth under normal conditions, but
still lose it under a fault condition. Under RPR, we could still use the
fast fault detection capability that is inherent in SONET, but we need a
protection scheme that re-routes packets the other way around the ring
without having to reserve specific protection bandwidth. In a fault
condition, a resulting potential overload on some of the links around the
ring would be handled by the fairness process and/or priority queuing, etc,
at the layer 2/packet level.
	So, what have I missed here?

Bob Solosko

-----Original Message-----
From: Mike Takefman [mailto:tak@xxxxxxxxx]
Sent: Friday, August 03, 2001 3:15 PM
To: Kannan Rangarajan
Cc: thanh.nguyen@xxxxxxxxxxxxx; Siamack Ayandeh;
owner-stds-802-17@xxxxxxxxxxxxxxxxxx; stds-802-17@xxxxxxxx; Ajay
Kamalvanshi
Subject: Re: [RPRWG] RPR and 1+1 linear APS like configuration.



Kannan, 

Speaking as myself and not the chair.

I think you are reading too much into the objectives, which is both
their power and weakness.

To me the ring protection and restoration in less than 50 msec
does not in any way require support for BLSR or UPSR especially 
condsidering Ethernet PHY layers. What it means is that the 
stations on the ring will do a protection switch (by means to 
be defined by the group) in under 50 msec. How the group decides
to do this is up for discussion. 

Support SONET/SDH phy layers does not require support for 
1+1 linear protection, although clearly a ring with a span
missing ends up being a linear topology. To me, SONET/SDH phy layers 
means that some carriers prefer SONET framing and the ability to 
carry RPR channels through existing infrastructure they might have.
This does imply using SONET framers that can do the following

1) support for OC-Nc/STM-N/4c with POS framing since 
this is both the most widely deployed / only standard for carrying
packets over SONET/SDH links (I seem to recall there is some other
way of doing it :) (forgive me I am on vacation). 
2) Support for smaller channels around traditional SONET/SDH ring, but 
the RPR layer operating independantly (mostly) of the SONET/SDH layer.

mike


Kannan Rangarajan wrote:
> 
> The RPR objective talks about the following:
> 1. RPR MAC - Ring protection and restoration in less than 50 msec.  This
means
> support of BLSR and UPSR (Check out Telcordia GR 253 for details).
> 2. Support SONET/SDH phy layers - This can mean requirement to support 1+1
> linear APS also.
> 
> Rajan.
> 
> -----Original Message-----
> From: owner-stds-802-17@xxxxxxxxxxxxxxxxxx
> [mailto:owner-stds-802-17@xxxxxxxxxxxxxxxxxx]On Behalf Of
> thanh.nguyen@xxxxxxxxxxxxx
> Sent: Friday, August 03, 2001 11:27 AM
> To: Siamack Ayandeh
> Cc: owner-stds-802-17@xxxxxxxxxxxxxxxxxx; stds-802-17@xxxxxxxx; Ajay
> Kamalvanshi
> Subject: Re: [RPRWG] RPR and 1+1 linear APS like configuration.
> 
> I thought 1+1 APS Linear is for point-to-point configuration not for ring
> configuration as in RPR. You would need to take a look at Ring APS since
> RPR's IPS is very similar to Ring APS. The difference is that Ring APS is
a
> layer 1 protocol and IPS is a layer 2 protocol. So the simple answer is
> still a no.
> 
> Thanh
> 
>                     Siamack Ayandeh
>                     <sayandeh@xxxxxxxxxx>             To:     Ajay
Kamalvanshi
> <vanshi@xxxxxxxxxxx>
>                     Sent by:                          cc:
> stds-802-17@xxxxxxxx, owner-stds-802-17@xxxxxxxxxxxxxxxxxx
>                     owner-stds-802-17@majordom        Subject:     Re:
[RPRWG]
> RPR and 1+1 linear APS like configuration.
>                     o.ieee.org
> 
>                     08/03/2001 08:56 AM
> 
> Ajay,
> 
> RPR is a frame based technology.  Not sure what you have in mind in terms
> of 1+1.  Could you please elaborate.  The simple answer seems to be no.
> 
> Thakns, Siamack
> 
> Ajay Kamalvanshi wrote:
> Hi All,    I have read few of the documents that are published/posted on
> web and I have this very basic question regarding 802.17. Can RPR used to
> implement something like 1+1 linear non-revertive APS equivalent. One
> simple way is to look at it as a ring with two nodes and four links. If a
> link breaks other link can be reused. Since I did not see this mentioned
> specifically I was wondering if it is a possible configuration for
> RPR. Thanks,vanshi
> 
> Get your FREE download of MSN Explorer at http://explorer.msn.com
> 
> << Attachment Removed : sayandeh.vcf >>

-- 
Michael Takefman              tak@xxxxxxxxx
Manager of Engineering,       Cisco Systems
Chair IEEE 802.17 Stds WG
2000 Innovation Dr, Ottawa, Canada, K2K 3E8
voice: 613-271-3399       fax: 613-271-4867