Transport protocols and bulk file transfer

[Scott Brim <>]

On 05/05/13 20:58, Ezra Kissel allegedly wrote:
> It really depends on the payload.  We used to have a limit of 64k, but
> that has been lifted with the possibility of one or more large option
> blocks within an XSP message.  The smallest XSP control message is
> simply the header (56 bytes).  The smallest message with payload is 56
> byte header + 6  byte option header + 1 byte payload = 63 bytes.
> 
> - ezra

OK.  You can see why I'm wondering.  If the new crop of wonder
algorithms are biased toward small packets (because in the regular old
Internet large packets generally means "bulk" i.e. "it's okay to take a
while"), then they aren't much help to us and in fact we'll have to
actively campaign to keep the option of turning them OFF.  Any ideas
what to do about this?