Stateful Ethernet Bridging and it's effect on overall Internet topology.
Gregory Taylor
greg at xwb.com
Tue Mar 16 20:51:43 UTC 2004
I agree, however there are some implementations of this type of bridging that 'routing' would not be a good substitute for. Say mangling traffic going outbound for compression purposes (A La Redline (Yes I know redline does proxying and not bridging)). I guess my best question would be, is there a solution to the problem. Maybe a possible way of bridging the traffic without polluting the world with unnecessary broadcasts of MAC addresses and over-head ethernet frames. (Is there a way to strip that garbage from the outbound traffic generated by the bridge).
Greg
---------- Original Message ----------------------------------
From: "Wayne E. Bouchard" <web at typo.org>
Date: Tue, 16 Mar 2004 12:49:38 -0700
>This goes back to traditional bridging issues.
>
>The problems include:
>
>loops and ineffective or broken STP implementations
>
>arp and broadcast storms
>
>mac address collisions
>
>which version of bridging to use and their associated advantages and
>disatvantages.
>
>I can't see that adding the capacity to do traffic shaping or
>filtering changes any of these issues. It just adds to the complexity.
>It still holds that, generally speaking, if you can route instead of
>bridging, it's a better option.
>
>On Tue, Mar 16, 2004 at 01:36:48PM -0600, Gregory Taylor wrote:
>>
>> I have a question and would like all of your opinions on this matter, as I research heavily into stateful ethernet bridging, packet mangling and their advantages and disadvantages to local and wide area network topologies.
>>
>> Deployed in large volumes, what negative effects, if any, would ethernet and fiber bridges have on the Internet as a whole.
>>
>> Lets say I was to build a bridge designed to intercept and manipulate traffic coming in from an outside network into my 'colo site' to do traffic shaping, packet filtering, and ethernet frames manipulation. And I deployed 100s of these into the facility as a means to control overall traffic. Would these transparent bridges be detrimental in any way to the rest of the internet. I understand that since they are re-transmitting data that the possibility of their MAC addresses popping up every time a machine behind it pops up could be an issue when doing network monitoring. But I'd just like to know what everyone thinks about such products.
>>
>> (Excuse me if my statements seem a little incoherent, I just woke up)
>>
>> Greg
>
>---
>Wayne Bouchard
>web at typo.org
>Network Dude
>http://www.typo.org/~web/
>
More information about the NANOG
mailing list