Skip to main content


Showing posts from December, 2012

Speeding up HTTP with minimal protocol changes

As SPDY works its way through IETF ratification I began wondering whether it was really necessary to add a complex, binary protocol to the HTTP suite to improve HTTP performance. One of the main things that SPDY sets out to fix is defined in the opening paragraph of the SPDY proposal: One of the bottlenecks of HTTP implementations is that HTTP relies on multiple connections for concurrency.  This causes several problems, including additional round trips for connection setup, slow-start delays, and connection rationing by the client, where it tries to avoid opening too many connections to any single server.  HTTP pipelining helps some, but only achieves partial multiplexing.  In addition, pipelining has proven non-deployable in existing browsers due to intermediary interference. The solution to this problem (as currently proposed) is SPDY. But I couldn't help thinking that solving the multiplexing problem could be done in a simpler manner within HTTP itself. And so here is a p

Listen on a UDP port and dump received lines of data

I needed to quickly fake up a syslog server for some debugging and wrote a small Perl program to listen for messages (lines of text) on a UDP port and dump them to the console. The program listens on a port specified on the command-line and simply prints out whatever it receives. It is only suitable for line oriented protocols since it uses the Perl <FN>  operator to read data. Here it is: #!/usr/bin/perl # # - listen on UDP port and dump out whatever has been received use strict; use warnings; use Socket; die "Usage: <port>" if (!defined($ARGV[0])); socket(UDP, PF_INET, SOCK_DGRAM, getprotobyname("udp")); bind(UDP, sockaddr_in($ARGV[0], INADDR_ANY)); print $_ while (<UDP>); PS Some people have asked why

How I ended up with so much Hacker News karma

On the Hacker News leaderboard I'm currently in position #12 with 32,360 points. I was curious to find out how I ended up in that position, so I used the HNSearch API  to pull down my complete submission and comment history. Some headline numbers: I've been on Hacker News for well over 5 years (I joined on July 2, 2007) and have made 2,985 comments or submissions. That's 1.5 comments or submissions every day for over 5 years (I've actually commented far more than I've submitted: 670 submissions; 2,315 comments). And for all those items I've received on average 10.8 points. So, the first conclusion is: consistency over a long period. Looking at comments and submissions separately, I've received 15,985 points for 2,315 comments and 18,119 points for 670 submissions. That means my average comment received 6.9 points, but my average submission received 27 points. Second conclusion: good submissions earn way more points than comments (which was probably