[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [RRG] Six/One Router Design Clarifications



I have to agree with Dino about this, and IMHO the
performance question is a red herring. My code has
been done in the linux kernel, which I realize is far
different than optimized fast-path router hardware,
but based on what I see there parsing headers is a
miniscule component of what goes on in terms of the
bigger picture of forwarding a packet.

Fred
fred.l.templin@boeing.com 

>-----Original Message-----
>From: Dino Farinacci [mailto:dino@cisco.com] 
>Sent: Thursday, July 24, 2008 6:20 PM
>To: Christian Vogt
>Cc: Drake, John E; Routing Research Group Mailing List
>Subject: Re: [RRG] Six/One Router Design Clarifications
>
>Not true, if you have fixed headers with well known code-points, you  
>can get to the data you need to parse. It's been done that way for  
>decades at very high speeds. What hasn't been done at 
>high-speeds is a  
>translator.
>
>With enough thrust anything can fly, but it's easier to do a  
>decapsulator than a translator.
>
>Dino
>
>On Jul 24, 2008, at 2:14 AM, Christian Vogt wrote:
>
>>
>> On Jul 23, 2008, Drake, John E wrote:
>>
>>>> Talking about performance:  An ACL that can limit its looks to
>>>> a single place in the IP header (i.e., with translation) can
>>>> likely be more efficient that an ACL that needs to look into
>>>> an inner IP header behind a pair of LISP and UDP headers.
>>>
>>> JD:  In any map & encap scheme, won't a transit space router have to
>>> deal with a multiplicity of packet formats (e.g., both 
>encapsulated  
>>> and
>>> non-encapsulated packets) and won't it be a performance hit 
>to figure
>>> out a packet's format and then look for the correct fields 
>within it?
>>
>> Yep, exactly.
>>
>> And you avoid this performance hit with translation-based schemes,  
>> where
>> it is always sufficient to look at the addresses in the IP header.
>>
>> - Christian
>>
>>
>>
>> --
>> to unsubscribe send a message to rrg-request@psg.com with the
>> word 'unsubscribe' in a single line as the message text body.
>> archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg
>
>
>--
>to unsubscribe send a message to rrg-request@psg.com with the
>word 'unsubscribe' in a single line as the message text body.
>archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg
>

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg