[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lwip-users] disconnect - connect delay
From: |
brak brak2 |
Subject: |
Re: [lwip-users] disconnect - connect delay |
Date: |
Mon, 4 Apr 2011 11:10:26 +0200 |
> For most browser, you don't need to save the HTTP request in RAM because most
> of the time, the request will come in in one packet. I know that's not really
> conforming to the standard, but it does save a lot of RAM.
> If you implement it like that, you do have to keep connection states for 2
> (or more) connections, but not the request.
I just did it as you wrote. I save only pBuf pointer of second connection.
> BTW, if you are that limited in RAM, you might be better off using uIP
> instead of lwIP.
lwIP is fine ;)
> Oh, and another thought regarding speed: you might trick the browser into
> using only one connection by correctly implementing HTTP/1.1 persistent
> connections: the browser might think it's faster to re-use one connection if
> it's a persistent one.
In HTTP 1.1 browser should use persistent connection as default, but I
send connection close in HTTP header of the response. I'll try to take
care of this later.
- Re: [lwip-users] disconnect - connect delay, (continued)
- Re: [lwip-users] disconnect - connect delay, Kieran Mansley, 2011/04/01
- Re: [lwip-users] disconnect - connect delay, brak brak2, 2011/04/01
- Re: [lwip-users] disconnect - connect delay, Kieran Mansley, 2011/04/01
- Re: [lwip-users] disconnect - connect delay, brak brak2, 2011/04/01
- Re: [lwip-users] disconnect - connect delay, Kieran Mansley, 2011/04/01
- Re: [lwip-users] disconnect - connect delay, brak brak2, 2011/04/04
- Re: [lwip-users] disconnect - connect delay, brak brak2, 2011/04/04
- Re: [lwip-users] disconnect - connect delay, Simon Goldschmidt, 2011/04/04
- Re: [lwip-users] disconnect - connect delay,
brak brak2 <=
RE: [lwip-users] disconnect - connect delay, Bill Auerbach, 2011/04/01