[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [RRG] A data point on transit MTU size
>From: Stephen Sprunk [mailto:email@example.com]
>Sent: Wednesday, September 24, 2008 4:15 PM
>Cc: Routing RG
>Subject: Re: [RRG] A data point on transit MTU size
>Tony Li wrote:
>> |The code for reassembly isn't hard, it's allocating the buffers to
>> |store the arrived fragments. That's nearly impossible in hardware.
>> Patently nonsense. Packet reassembly in hardware is very
>much straightforward if one is willing to burn large buffers
>to do so. Imagine taking an implemention in C and simply
>converting it to Verilog. Not at all out of the question.
>> In fact, a subset of this problem has been solved for a very
>long time. Remember ATM? Remember a SAR chip? That's
>reassembly of fixed size packet fragments.
>I remember building ATM networks (not voluntarily) and that
>whole movement seemed to burn out when it was discovered that
>SAR chips couldn't keep up above OC-12 speeds -- and that was
>with easy fixed-size cells. I also remember months of my life
>around the same time eaten up by tuning a certain vendor's
>routers for optimum use of various-sized IP buffers so that
>they wouldn't drop packets even without reassembly being an issue.
>Now, I'm not a hardware guy, but I'd like to see some
>assurance from those that are that things have changed enough
>that we can ask routers now and in the foreseeable future will
>be able to reassemble the 50% of packets that will be at the
>Ethernet MTU (whatever that may be) + LISP header size at line
>rate -- including for customers who have GE, 10GE, and (in a
>few years) 100GE uplinks to their ISP(s).
But, that's just it. We're not talking about steady-state
line-rate reassembly for the long term. We are talking about
very short term identification and eradication of marginal
links in the interdomain core that probably should have
been replaced years ago anyway - and for relatively low
cost. As the obsolete links go away, reassembly at ETRs
dissipates and the Internet eases into a comfort zone that
includes all-jumboframe links everywhere. IMHO, it is well
worth the (very) short-term transition.
>to unsubscribe send a message to firstname.lastname@example.org 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 email@example.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