Re: Lingering Close

Zhong Yu wrote:
> Thanks Willy, I think I get what you mean by now. FIN should be
> initiated by server to avoid the TIME_WAIT problem. Therefore the
> half-close step is important.
> 
> The current text makes perfect sense to me now.

That's great.  I knew about the RST problem (and waiting for data-ACK
isn't enough to solve that), but I didn't think about the client-side
TIME_WAIT problem before now.  Thanks Willy!

> Unfortunately, this lingering close process cannot be implemented on
> top of some APIs; these APIs don't do transparent lingering close upon
> close() either. But it is their fault, not the spec's.

Unfortunately because the old HTTP specs didn't mention this, it seems
to be widely believed that ordinary close is fine for HTTP, and APIs
were created with HTTP in mind to run over them with this limitation.
It's unfortunate that lingering close wasn't made clear in the earlier
specs.

-- Jamie

Received on Wednesday, 28 November 2012 19:36:38 UTC