Re: Clock skew and the importance of clock synchronization in HTTP servers.

• Messages sorted by: [ date ][ thread ][ subject ][ author ]
• Next message: Marc Slemko: "Re: indefinite server-push (was 'Last-Modified in chunked footer')"
• Previous message: David W. Morris: "Re: indefinite server-push (was 'Last-Modified in chunked footer')"
• In reply to: Roy T. Fielding: "Re: Clock skew and the importance of clock synchronization in HTTP servers."
• Next in thread: Joel N. Weber II: "Re: Clock skew and the importance of clock synchronization in HTTP servers."

Roy T. Fielding:
>
[...]
>It is still a good idea to recommend installation of time synchronization
>software along with the server docs, but things are not as bad as they seem.

I agree, things are not as bad as they seem.

I may be repeating the obvious here, but the 1.1 Age header will do
away with incorrectness due to unsynchronised clocks, at least in a
pure 1.1 chain.  Once everything is 1.1, unsynchronised clocks will no
longer threaten correctness.

They will still threaten efficiency though, due to some unfortunate
properties of the (current) HTTP/1.1 age calculation algorithm.

>.....Roy

Koen.

• Next message: Marc Slemko: "Re: indefinite server-push (was 'Last-Modified in chunked footer')"
• Previous message: David W. Morris: "Re: indefinite server-push (was 'Last-Modified in chunked footer')"
• In reply to: Roy T. Fielding: "Re: Clock skew and the importance of clock synchronization in HTTP servers."
• Next in thread: Joel N. Weber II: "Re: Clock skew and the importance of clock synchronization in HTTP servers."