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

RE: [EFM] OAM loop back / echo server function, ping.

We all have been using ping for years to debug ethernet networks. Is there
any chance that this would suffice as a remote loopback command? I have a
hard time envisioning any kind of a remote loopback on a PON that would not
protocol layers higher than L2, or some very complex hardware.

For a PON a true tx-to-rx loopback would require a complete shutdown of
upstream traffic except for the ONU being tested. This could be useful
during installation of a new network, or restoring service after a
catastrophic failure, but would be unacceptable for routine debugging. Ping
is a mature protocol, network technicians are very familiar with it, and it
is very flexible, and most important - it doesn't add any cost or hardware
complexity to support.

Any comments from the service providers??

-----Original Message-----
From: Carlos Ribeiro [mailto:cribeiro@xxxxxxxxxxxxxxxx]
Sent: Thursday, August 30, 2001 4:57 PM
To: bob.barrett@xxxxxxxxxxxxxxx; Roy Bynum; Rick Li; Charles Cook; Faye
Cc: glen.kramer@xxxxxxxxxxxx; zhangxu72@xxxxxxxxx; RHirth@xxxxxxxxxxxx;
Subject: RE: [EFM] OAM loop back / echo server function

At 17:29 30/08/01 +0100, Bob Barrett wrote:
>Remote loop back of Ethernet packets / 802.3 frames is a really bad idea. 
>No mater how well intentioned it will go wrong sometimes and when it does 
>it is really bad news.

I also was late on this discussion, but I have the same opinion. In 
general, thinking as a service provider perspective, each and every user 
should have a logically protected circuit to the 
headend/olt/switch/whatever. Any chance of having packets exchanged 
directly between the users opens a wide door for exploits (or even 
unintentional) problems. In this sense, I believe that EFM should allow

- direct broadcast on the downstream, limited by VLAN, where the service 
provider can control what packets get broadcasted;
- direct point-to-point communication upstream between the ONU and the OLT.

For some specific applications, bridging may be a simple solution for the 
direct communication between the endpoints. As an access solution for 
residential users, it is a bad idea.

Carlos Ribeiro
CTBC Telecom