Comments on: On implementing Chord http://www.emilsit.net/blog/archives/on-implementing-chord/ Wed, 10 Feb 2010 17:08:52 -0500 http://wordpress.org/?v=2.8.4 hourly 1 By: Princeton S* Network Systems» Blog Archive » “Systems Researcher Interview” http://www.emilsit.net/blog/archives/on-implementing-chord/comment-page-1/#comment-27127 Princeton S* Network Systems» Blog Archive » “Systems Researcher Interview” Wed, 25 Mar 2009 14:28:11 +0000 http://www.emilsit.net/blog/?p=171#comment-27127 <p>[...] from colleagues about their experience building distributed systems. He started this series after describing some implementation issues with building Chord, which was one of the first and remains the [...]</p> [...] from colleagues about their experience building distributed systems. He started this series after describing some implementation issues with building Chord, which was one of the first and remains the [...]

]]> By: Systems Researchers Interview Series - Emil Sit: Thoughts on Systems http://www.emilsit.net/blog/archives/on-implementing-chord/comment-page-1/#comment-27119 Systems Researchers Interview Series - Emil Sit: Thoughts on Systems Sun, 15 Mar 2009 01:41:10 +0000 http://www.emilsit.net/blog/?p=171#comment-27119 <p>[...] Emil Sit: Thoughts on Systems Skip to content Archives « On implementing Chord [...]</p> [...] Emil Sit: Thoughts on Systems Skip to content Archives « On implementing Chord [...]

]]> By: Emil Sit http://www.emilsit.net/blog/archives/on-implementing-chord/comment-page-1/#comment-27118 Emil Sit Sat, 14 Mar 2009 13:49:57 +0000 http://www.emilsit.net/blog/?p=171#comment-27118 <p>@ctl: Yeah, we don't do that...<br /> @Derrick: STP is simply implemented over UDP, but it's a bit hairy.</p> @ctl: Yeah, we don’t do that…
@Derrick: STP is simply implemented over UDP, but it’s a bit hairy.

]]> By: Derrick Coetzee http://www.emilsit.net/blog/archives/on-implementing-chord/comment-page-1/#comment-27116 Derrick Coetzee Sat, 07 Mar 2009 01:33:51 +0000 http://www.emilsit.net/blog/?p=171#comment-27116 <p>Hi Mendel, David Hilley pointed me to this interesting post after I recently (coincidentally) wrote a post summarizing and discussing the original Chord paper at:</p> <p>http://papersincomputerscience.org/2009/03/04/chord-a-scalable-peer-to-peer-lookup-service-for-internet-applications/</p> <p>I'm most intrigued by your use of a custom transport protocol - I'm surprised you have enough control over your network stack to pull that off. Thanks for taking the time to share your practical insights.</p> Hi Mendel, David Hilley pointed me to this interesting post after I recently (coincidentally) wrote a post summarizing and discussing the original Chord paper at:

http://papersincomputerscience.org/2009/03/04/chord-a-scalable-peer-to-peer-lookup-service-for-internet-applications/

I’m most intrigued by your use of a custom transport protocol – I’m surprised you have enough control over your network stack to pull that off. Thanks for taking the time to share your practical insights.

]]> By: Chris Lesniewski-Laas http://www.emilsit.net/blog/archives/on-implementing-chord/comment-page-1/#comment-27057 Chris Lesniewski-Laas Mon, 02 Mar 2009 16:33:39 +0000 http://www.emilsit.net/blog/?p=171#comment-27057 <p>A good DHT ought to be able to deal with asymmetrically reachable nodes by routing over the DHT peers, a la RON. If A mentions B to C, then A either can contact B itself, or A knows another node D which can contact B directly.</p> A good DHT ought to be able to deal with asymmetrically reachable nodes by routing over the DHT peers, a la RON. If A mentions B to C, then A either can contact B itself, or A knows another node D which can contact B directly.

]]>