Incorrect Base Assumptions About Network Management [OpEd]Comments on: http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/ coherent thoughts on diverse topics Sun, 21 Mar 2010 03:24:23 -0700 http://wordpress.org/?v=2.9.2 hourly 1 By: Kyle Brady: Blog - An Open Letter to Senator Diane Feinstein http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3671 Kyle Brady: Blog - An Open Letter to Senator Diane Feinstein Fri, 30 Oct 2009 14:17:29 +0000 http://www.kyle-brady.com/?p=3107#comment-3671 [...] this process of overloading, they employ tactics of filtering, shaping, and blocking traffic, often citing bogus claims of “network overload” or “unacceptable traffic”. Finally, an underlying issue that will [...] [...] this process of overloading, they employ tactics of filtering, shaping, and blocking traffic, often citing bogus claims of “network overload” or “unacceptable traffic”. Finally, an underlying issue that will [...]

]]> By: Robb Topolski http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3093 Robb Topolski Wed, 15 Jul 2009 19:39:22 +0000 http://www.kyle-brady.com/?p=3107#comment-3093 Dr. Robert's is still wrong -- not confused, but purposefully wrong. <br><br>If I have a 1 Mbps upload connection and you have a 1 Mbps upload connection, and I'm uploading using 100 connections and you're uploading using 1, then how can I possibly be using more bandwidth than you? <br><br>Robb Topolski Dr. Robert's is still wrong — not confused, but purposefully wrong.

If I have a 1 Mbps upload connection and you have a 1 Mbps upload connection, and I'm uploading using 100 connections and you're uploading using 1, then how can I possibly be using more bandwidth than you?

Robb Topolski

]]> By: Tom Koltai http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-2827 Tom Koltai Fri, 10 Jul 2009 07:59:20 +0000 http://www.kyle-brady.com/?p=3107#comment-2827 The comments actually now show up the weakness of where I was about to make a comment.<br>P2P by itself is not actually the main culprit. video streaming or any srtreaming client ---> server is more of a culprit. In fact P2P utilised at the local ISP level relieves congestion significantly.<br>Unfrotunately most P2P clients actively look for client/peers at 5+ hops which of course negates the practicality of P2P as a network decongestant.<br><br>Should ISP's be permitted to install Squid P2P Cahce servers at a local level and throttle all other P2P traffic, then I dare say, a great deal of the current P2P congestion would disappear. <br><br>I guess it comes down to the regulatory environent. If Governments are more interested in kowtowing to copyright industries than fostering open access and intelectual growth providing a much needed stimulus for the economy of every nation, then I'm afraid P2P will continue to be part of the problem instead of the "missing link". The comments actually now show up the weakness of where I was about to make a comment.
P2P by itself is not actually the main culprit. video streaming or any srtreaming client —> server is more of a culprit. In fact P2P utilised at the local ISP level relieves congestion significantly.
Unfrotunately most P2P clients actively look for client/peers at 5+ hops which of course negates the practicality of P2P as a network decongestant.

Should ISP's be permitted to install Squid P2P Cahce servers at a local level and throttle all other P2P traffic, then I dare say, a great deal of the current P2P congestion would disappear.

I guess it comes down to the regulatory environent. If Governments are more interested in kowtowing to copyright industries than fostering open access and intelectual growth providing a much needed stimulus for the economy of every nation, then I'm afraid P2P will continue to be part of the problem instead of the “missing link”.

]]> By: Tom Koltai http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3092 Tom Koltai Fri, 10 Jul 2009 07:59:20 +0000 http://www.kyle-brady.com/?p=3107#comment-3092 The comments actually now show up the weakness of where I was about to make a comment.<br>P2P by itself is not actually the main culprit. video streaming or any srtreaming client ---> server is more of a culprit. In fact P2P utilised at the local ISP level relieves congestion significantly.<br>Unfrotunately most P2P clients actively look for client/peers at 5+ hops which of course negates the practicality of P2P as a network decongestant.<br><br>Should ISP's be permitted to install Squid P2P Cahce servers at a local level and throttle all other P2P traffic, then I dare say, a great deal of the current P2P congestion would disappear. <br><br>I guess it comes down to the regulatory environent. If Governments are more interested in kowtowing to copyright industries than fostering open access and intelectual growth providing a much needed stimulus for the economy of every nation, then I'm afraid P2P will continue to be part of the problem instead of the "missing link". The comments actually now show up the weakness of where I was about to make a comment.
P2P by itself is not actually the main culprit. video streaming or any srtreaming client —> server is more of a culprit. In fact P2P utilised at the local ISP level relieves congestion significantly.
Unfrotunately most P2P clients actively look for client/peers at 5+ hops which of course negates the practicality of P2P as a network decongestant.

Should ISP's be permitted to install Squid P2P Cahce servers at a local level and throttle all other P2P traffic, then I dare say, a great deal of the current P2P congestion would disappear.

I guess it comes down to the regulatory environent. If Governments are more interested in kowtowing to copyright industries than fostering open access and intelectual growth providing a much needed stimulus for the economy of every nation, then I'm afraid P2P will continue to be part of the problem instead of the “missing link”.

]]> By: Chris Brandt http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-2824 Chris Brandt Thu, 09 Jul 2009 17:42:38 +0000 http://www.kyle-brady.com/?p=3107#comment-2824 Web browsers have a very different bandwidth profile, but your point is valid in that other types of services can saturate a connection. Certainly any UDP service can do this.<br><br>We used to take high speed circuits (100Mb, OC3, etc) out for a test spin with a well seeded torrent as that was one of the few things that would really hit it hard.<br><br>All of that aside, the point of your article is spot on in that the nature of the traffic is changing and the telcos need to adjust their assumptions and figure out how to continue to make it profitable. Web browsers have a very different bandwidth profile, but your point is valid in that other types of services can saturate a connection. Certainly any UDP service can do this.

We used to take high speed circuits (100Mb, OC3, etc) out for a test spin with a well seeded torrent as that was one of the few things that would really hit it hard.

All of that aside, the point of your article is spot on in that the nature of the traffic is changing and the telcos need to adjust their assumptions and figure out how to continue to make it profitable.

]]> By: Chris Brandt http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3089 Chris Brandt Thu, 09 Jul 2009 17:42:38 +0000 http://www.kyle-brady.com/?p=3107#comment-3089 Web browsers have a very different bandwidth profile, but your point is valid in that other types of services can saturate a connection. Certainly any UDP service can do this.<br><br>We used to take high speed circuits (100Mb, OC3, etc) out for a test spin with a well seeded torrent as that was one of the few things that would really hit it hard.<br><br>All of that aside, the point of your article is spot on in that the nature of the traffic is changing and the telcos need to adjust their assumptions and figure out how to continue to make it profitable. Web browsers have a very different bandwidth profile, but your point is valid in that other types of services can saturate a connection. Certainly any UDP service can do this.

We used to take high speed circuits (100Mb, OC3, etc) out for a test spin with a well seeded torrent as that was one of the few things that would really hit it hard.

All of that aside, the point of your article is spot on in that the nature of the traffic is changing and the telcos need to adjust their assumptions and figure out how to continue to make it profitable.

]]> By: Kyle Brady http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-2823 Kyle Brady Thu, 09 Jul 2009 17:12:56 +0000 http://www.kyle-brady.com/?p=3107#comment-2823 Hey Sean,<br><br>I also agree with you - the oversubscription for low-cost purposes is an issue I didn't touch here, but may in the future.<br><br>Good insight though.<br><br>--Kyle Hey Sean,

I also agree with you – the oversubscription for low-cost purposes is an issue I didn't touch here, but may in the future.

Good insight though.

–Kyle

]]> By: Kyle Brady http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3091 Kyle Brady Thu, 09 Jul 2009 17:12:56 +0000 http://www.kyle-brady.com/?p=3107#comment-3091 Hey Sean,<br><br>I also agree with you - the oversubscription for low-cost purposes is an issue I didn't touch here, but may in the future.<br><br>Good insight though.<br><br>--Kyle Hey Sean,

I also agree with you – the oversubscription for low-cost purposes is an issue I didn't touch here, but may in the future.

Good insight though.

–Kyle

]]> By: Kyle Brady http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-2822 Kyle Brady Thu, 09 Jul 2009 17:10:51 +0000 http://www.kyle-brady.com/?p=3107#comment-2822 Hey Chris,<br><br>Thanks.<br><br>I'll agree with too many streams saturating a connection... this is a known problem. Depending on, say, the Torrent client you use, you can decide whether or not you want to be friendly to your network by not going overboard. But if I recall, browsers have the same problem as they just open new connections whenever they feel the need to do so.<br><br>So it's hardly limited to P2P.<br><br>--Kyle Hey Chris,

Thanks.

I'll agree with too many streams saturating a connection… this is a known problem. Depending on, say, the Torrent client you use, you can decide whether or not you want to be friendly to your network by not going overboard. But if I recall, browsers have the same problem as they just open new connections whenever they feel the need to do so.

So it's hardly limited to P2P.

–Kyle

]]> By: Kyle Brady http://www.kyle-brady.com/2009/07/09/incorrect-base-assumptions-about-network-management/#comment-3088 Kyle Brady Thu, 09 Jul 2009 17:10:51 +0000 http://www.kyle-brady.com/?p=3107#comment-3088 Hey Chris,<br><br>Thanks.<br><br>I'll agree with too many streams saturating a connection... this is a known problem. Depending on, say, the Torrent client you use, you can decide whether or not you want to be friendly to your network by not going overboard. But if I recall, browsers have the same problem as they just open new connections whenever they feel the need to do so.<br><br>So it's hardly limited to P2P.<br><br>--Kyle Hey Chris,

Thanks.

I'll agree with too many streams saturating a connection… this is a known problem. Depending on, say, the Torrent client you use, you can decide whether or not you want to be friendly to your network by not going overboard. But if I recall, browsers have the same problem as they just open new connections whenever they feel the need to do so.

So it's hardly limited to P2P.

–Kyle

]]>