Re:Bounce your http requests around with bouncy
by substack on Tuesday October 11, 2011 @06:09AM
(#3097958)
I guess what I was saying there isn't the whole story, since bouncy uses node's http parser directly to handle splitting up keep-alive requests. What I meant was more that bouncy sends data over a tcp stream instead of simply passing the request object through to an http client request.
Re:Bounce your http requests around with bouncy
by MartinMond on Tuesday October 11, 2011 @06:18AM
(#3097987)
Ok, then this is awesome. Also, does it maintain the upstream connections?
E.g. do two requests on one kept-alive connection get translated into two separate requests for the upstream server or will the upstream connection also be kept-alive/reused?
Re:Bounce your http requests around with bouncy
by icebraining on Tuesday October 11, 2011 @09:36AM
(#3098469)
For 'dumb' (not protocol aware) redirection you can just use iptables, it'll probably be much faster.
For protocol aware redirection, there are various reverse proxies with that ability. TCPMUX was probably one of the firsts, but it requires explicit support by the client. Delegate[1], on the other hand, can identify the protocol and multiplex based on that, IINM.
Re:Bounce your http requests around with bouncy
by qw on Tuesday October 11, 2011 @03:43PM
(#3100007)
It's a cool hack, but if you are going to have a server in front of a web server why not use something like Varnish instead and get caching as well as load balancing.
Re:Bounce your http requests around with bouncy
by substack on Tuesday October 11, 2011 @06:04AM
(#3097946)
This software fills the same niche that something like HAProxy or apache virtual hosts fills right now. Bouncy is great for when you have services on your internal network or on different ports and you want to map them to subdomains.
Re:Bounce your http requests around with bouncy
by darklajid on Tuesday October 11, 2011 @06:11AM
(#3097966)
Different thing, no?
This seems to redirect locally. It seems to be more of a dynamic port redirection (the article could've been a little clearer..).
You wouldn't use SSH to redirect stuff from port x on localhost to port y and you couldn't do it dynamically (unless you'd like to spawn ssh for each connection, setting up -L according to header information in the request).
E.g. do two requests on one kept-alive connection get translated into two separate requests for the upstream server or will the upstream connection also be kept-alive/reused?
Maybe could even add the ability to peek into the stream to determine which host to forward to based on protocol (think FTP etc.).
For protocol aware redirection, there are various reverse proxies with that ability. TCPMUX was probably one of the firsts, but it requires explicit support by the client. Delegate[1], on the other hand, can identify the protocol and multiplex based on that, IINM.
[1]: http://www.delegate.org/delegate/
This seems to redirect locally. It seems to be more of a dynamic port redirection (the article could've been a little clearer..).
You wouldn't use SSH to redirect stuff from port x on localhost to port y and you couldn't do it dynamically (unless you'd like to spawn ssh for each connection, setting up -L according to header information in the request).