[ih] UDP Length Field?

[Joseph Touch]

Hi, all,

FYI - Vint Carf is on this list and I’ve already reached out off-list to Bob Kahn. I’ll also reach out to others that might know (notably Carl Sunshine).

> On Nov 28, 2020, at 3:42 PM, Timothy J. Salo via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> Hi,
> 
> Can anyone provide some [historical] insight into why the UDP header
> contains a length field?  TCP manages to ascertain the length of data in
> a packet just fine without a length field, so why couldn't UDP?

It can; IP includes its own length field so it can travel over links that don’t necessarily include length delimiters. 

> Several people have noted that the UDP length field is redundant,
> including for example, the current Internet Draft "Transport Options for
> UDP",
> <https://www.ietf.org/archive/id/draft-ietf-tsvwg-udp-options-09.txt>.

As author of that doc, I’d point out RFC 6081 as the other primary example, which came with a defensive patent. However, that document seems to try to use the UDP length to point longer than the IP length would indicate, rather than shorter (as currently proposed for UDP options).

The idea of using the UDP length field for multiple UDP segments in the same IP packet makes the most sense to me, but isn’t mentioned in the two precursor versions (IEN 71, IEN 88). It’s also counter to IEN 72, in which Postel introduces a “multiplexing protocol” that provides its own protocol/length header that can be repeated to allow for multiple UDP, TCP, or (presumably) other transport protocols to be included in a single IP datagram. The example in that document shows the UDP length field and it being redundant with the multiplexing protocol even at that time.

Joe