Concurrent Server Connections
Discussions center on challenges and strategies for handling large numbers of concurrent TCP/HTTP connections in web servers, including limits from file descriptors and resources, connection reuse, pooling, keep-alives, and scalability issues like C10k.
Activity Over Time
Top Contributors
Keywords
Sample Comments
Wouldn't it consume the same number of connections on my server?
You'd have to hold a ton of connections open for 24 hours+
Make sure you’re reusing connections. A lot of languages and libraries create a new connection for every request unless you explicitly manage a session or connection pool.
Web servers are limited in the number of concurrent connections they can maintain (tubes aren’t big enough).This doesn't really make sense -- maintaining a TCP connection requires no "tubes"; the limit comes from finite state storage: a finite state table, a finite number of file descriptors, and a finite amount of RAM in which to store the application-level connection contexts. These numbers turn out to be Really Fucking Big, so this is probably not affecting your application unless
their heuristics is probably looking for long time connections.. you're scaping by moving the client around
I think you can use https keep alive to only require one tcp connection.
That's neat. Does it require 1 connection = 1 process to work? I don't see how you can have a long running server with this feature.
You might run into this limitation more quickly if you are receiving connections via a load balancer.
Even load balancers force you to have a frequent heartbeat all the way to the client for each connection.
Thanks for sharing this How do they handle the server load when they have so many concurrent connections