Re: [hybi] Framing take IV [why fragment]

[Patrick McManus <[e-mail]>]

On Tue, 2010-08-03 at 18:20 -0700, Maciej Stachowiak wrote:

>  people think fragmentation is simpler than a variable size length
> field, they should explain why, 


My previous mail
( http://www.ietf.org/mail-archive/web/hybi/current/msg02886.html )
provided some arguments other than simplicity in favor of chunking.

But I also believe fragments with fixed lengths are simpler than a BER
because properly aligned lengths are always correct data - there are no
error cases as all the bits are defined to be some quantity regardless
of their arrangement. BER has error cases and determining the expressed
quantity requires multiple indirections and a significant algorithm to
use it. 

That contrasts with a fixed length 16 bit unsigned integer that requires
only a load instruction and perhaps a built in byte order swap to use
it. The algorithm is literally already in all hardware.

The cost of chunking vs BER is going to be 1 header (~4 bytes) every
64KB and only occurs at all on messages over 64KB (which excludes lots
of use cases). That's cheap!

I will say this though - sometimes the receiver would like to know how
big the total message is upfront in order to do non fragmented
preallocation for example. If the sender knows that information (i.e. it
is not really streaming) but is sending a > 64KB message using fragment
framing there is no way to convey that useful info to the recvr up
front. An option, unrelated to the framing, might be used to carry that
info in a variable length format (ber or maybe even ascii hex..
whatever)