1. 16
  1.  

  2. 13

    This is a thin wrapper from Hackaday over this submission.

    1. 2

      Good point, usually I try to get the actual original sources for Hackaday submissions but got lazy this time. Sorry about that.

      1. 2

        No worries buddy. :)

    2. 3

      I had the pleasure of going through these steps at my last job. In that time, I followed a course lead by Iljitsch van Beijnum, author of the O’Reilly book on BGP. Had some fun poisoning the routing table of a fellow student, but they were on to me quick enough. ;)

      I’m a software engineer professionally, but the knowledge and experience I gained from the whole process are very valuable. Many things are networked these days, and a basic understanding of how the lower layers of networking and a routing table work (even on your personal machine or servers) is super helpful in debugging issues.

      I’ve found that a good amount of people in the field are hardware enthusiasts, much like portrayed in the blog post of the original author. Certainly everyone has their preferences for and bad experiences with specific brands and/or product lines. My 2 cents is I can’t stand the Cisco CLI, and all imitations / derivations of it (like the popular Quagga.)

      But our setup was small, and besides a hardware switch, we actually ran plain Debian on Dell servers, with BIRD. Despite running lots of VoIP traffic through it, we never had any issues with routing performance at the scale.

      There’s also stories of people running Mac Minis for the job. It all depends on what kind of performance you need, and what kind of peering sessions you’re setting up. It’s very possible to do BGP at a small scale, with just a couple of ISPs. The routing tables will be small, and memory requirements low, in that case.