From keshav at uwaterloo.ca Mon Mar 1 12:27:29 2010 From: keshav at uwaterloo.ca (S. Keshav) Date: Mon, 1 Mar 2010 15:27:29 -0500 Subject: [e2e] end2end-interest Digest, Vol 72, Issue 1 In-Reply-To: References: Message-ID: <67B70A60-ECF5-491B-B345-54CB8DB3E0CA@uwaterloo.ca> I see an editorial in the making! Can I have it for the July issue (sorry, too late for the April issue) thanks keshav On 2010-03-01, at 3:00 PM, end2end-interest-request at postel.org wrote: > Send end2end-interest mailing list submissions to > end2end-interest at postel.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://mailman.postel.org/mailman/listinfo/end2end-interest > or, via email, send a message with subject or body 'help' to > end2end-interest-request at postel.org > > You can reach the person managing the list at > end2end-interest-owner at postel.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of end2end-interest digest..." > > > Today's Topics: > > 1. Fie on future internet (Jon Crowcroft) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sun, 28 Feb 2010 15:27:42 +0000 > From: Jon Crowcroft > Subject: [e2e] Fie on future internet > To: end2end-interest at postel.org > Message-ID: > Content-Type: text/plain; charset="us-ascii" > > I spend a lot of time reviewing > stuff about the future internet > then its raining on a sunday afternoon > so i wrote down what i really think:- > http://www.cl.cam.ac.uk/~jac22/out/fie.pdf > > then i played the guitar to clam down > http://www.youtube.com/watch?v=UsJD5_nX5RE > > j. > > > ------------------------------ > > _______________________________________________ > end2end-interest mailing list > end2end-interest at postel.org > http://mailman.postel.org/mailman/listinfo/end2end-interest > > > End of end2end-interest Digest, Vol 72, Issue 1 > *********************************************** From dpreed at reed.com Mon Mar 1 14:45:15 2010 From: dpreed at reed.com (David P. Reed) Date: Mon, 01 Mar 2010 17:45:15 -0500 Subject: [e2e] Future Internet Engineering In-Reply-To: <67B70A60-ECF5-491B-B345-54CB8DB3E0CA@uwaterloo.ca> References: <67B70A60-ECF5-491B-B345-54CB8DB3E0CA@uwaterloo.ca> Message-ID: <4B8C437B.5080604@reed.com> Jon - excellent piece. I agree that the mania around the Future Internet (especially the idea that we MUST have a Clean Slate) seems to have been a temporary reaction by a research community to a funding structure problem, or perhaps a process of Bandwagon Joining by people who saw a new bandwagon that seemed to have political traction. Bandwagons and Clean Slates are rhetorical moves that can have enormous value in shaking a clogged bureaucratic process - so there's not necessarily anything wrong with them, but they have indeed become unlikely to produce "good science" or "good engineering" or anything else, now that they are no longer shaking anything up. That said, the idea that one might be able to solve problems without Waiting for Obama or Waiting for IETF or Waiting for Copenhagen is one of the inspiring things about the Internet in comparison to the idea that one must Wait for DARPA or Wait for NSF to fund a zillion CS and EE departments. We might be able to start a small skunk works to think about decentralized solutions to managing the energy flux that impinges on the geosystem, to both avoid absorbing too much energy and disrupting the planet's inhabitants, while at the same time providing support for 10 Billion future inhabitants. That might be based on a few principles of the sort you describe, and an approach that doesn't require top-down agreement by all the stakeholders who want to control it. Hmmm.... From jamel at cin.ufpe.br Tue Mar 2 03:40:22 2010 From: jamel at cin.ufpe.br (Djamel Sadok) Date: Tue, 2 Mar 2010 08:40:22 -0300 Subject: [e2e] Fie on future internet In-Reply-To: References: Message-ID: SMALL Change --> Big leap! The future internet is nothing but the same as the current one with few twinks - small changes that may have big (business or technological) impacts. No one is expected to bank or coordinate worldwide big changes. It is about being disruptive. Just like the human brain, it still maintains the basic old parts while adding new (smarter) layers. There will be no day that one will say, today we have upgraded to the FI. Someone once said, too much government is bad government. That is the case of the Internet and a lesson that governments may learn from it. People will always come up with new ideas and if they work, others may adopt them. Current FI R&D projects will never be "the" solution. CENTRALIZATION What worries me about the Internet is its actual centralization. Take away few search engines and most users won?t get their work done. Hierarchy as Jon talked about is good, but I guess he meant it at the structure level. This hierarchy is missing where it equally matters: the content and service levels. May be the notion of federations, contexts, turfs, societies, or whatever you call them, is the way ahead to break away from such centralized business model. If like nature, the internet favors the fittest, then we are doomed with a centralized structure. CONTROL Governments, transport systems, banking systems, "telecom systems", education systems, and others, like to think they have control and achieve predictability. We are lucky they did not design the Internet because they would have used to it to gain more control and kept to themselves as a secret advantage or tool. The internet design community saw design ideas as more important than control and today the internet may be seen as the first real autonomic system. BIG PLAYERS Many big players are today involved. Some are making money and others see their shares dissipate. The big ones seem to dictate the changes. They facilitate what they want users to have (make ebook readers, take fiber to your home to sell you content, ..) it would be interesting to evaluate to what extent their silent decisions impact the FI. My opinion is they are the ones who are designing it right now. The internet borrowed a great deal from areas such as economy, transport, etc. to build its inner working mechanisms. Those systems have been around longer than the internet. I would suggest that if those areas are going to benefit from the Internet it won?t be at the structural level (routing, buffering etc..), but rather at the community, cooperation, security, information, levels. The internet shows that: 1) a simple trust model may work and give its participant many benefits; 2) even when there are no guarantees by the members, the overall system may still deliver; 3) when each of us contributes with a small resource, the overall system may outperform any big system; 4) people like free things even if they do not work well; etc..... To complement Jon?s thoughts, "higher level" Internet lessons may be best to use in other areas rather than its underlying mechanisms. DJamel On Sun, Feb 28, 2010 at 12:27 PM, Jon Crowcroft wrote: > I spend a lot of time reviewing > stuff about the future internet > then its raining on a sunday afternoon > so i wrote down what i really think:- > http://www.cl.cam.ac.uk/~jac22/out/fie.pdf > > then i played the guitar to clam down > http://www.youtube.com/watch?v=UsJD5_nX5RE > > j. > Crucial public systems deliver food, energy, transport, housing and so on. Three systems seem like low-hanging fruit when it comes to re-application of the ideas behind Internet: -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.postel.org/pipermail/end2end-interest/attachments/20100302/9d208d6a/attachment.html From ashish.makani at gmail.com Tue Mar 2 10:32:58 2010 From: ashish.makani at gmail.com (ashish makani) Date: Tue, 2 Mar 2010 13:32:58 -0500 Subject: [e2e] [OT ?] NYTimes article on the structure & shape(topology) of the internet Message-ID: <5fa6bfbc1003021032g65c7287bj32f8ece8f9a23a2b@mail.gmail.com> An NYT article on peering, May not be very relevant to e2e, but seemed interesting to me. http://www.nytimes.com/2010/03/02/science/02topo.html I learnt about several concepts i was unaware of... David Alderson's HOT model [http://portal.acm.org/citation.cfm?id=1115527] & some new developments like SFMIX [http://www.sfmix.org/] On Mon, Mar 1, 2010 at 5:45 PM, David P. Reed wrote: > Jon - excellent piece. > > I agree that the mania around the Future Internet (especially the idea that > we MUST have a Clean Slate) seems to have been a temporary reaction by a > research community to a funding structure problem, or perhaps a process of > Bandwagon Joining by people who saw a new bandwagon that seemed to have > political traction. > > Bandwagons and Clean Slates are rhetorical moves that can have enormous > value in shaking a clogged bureaucratic process - so there's not necessarily > anything wrong with them, but they have indeed become unlikely to produce > "good science" or "good engineering" or anything else, now that they are no > longer shaking anything up. > > That said, the idea that one might be able to solve problems without > Waiting for Obama or Waiting for IETF or Waiting for Copenhagen is one of > the inspiring things about the Internet in comparison to the idea that one > must Wait for DARPA or Wait for NSF to fund a zillion CS and EE departments. > > We might be able to start a small skunk works to think about decentralized > solutions to managing the energy flux that impinges on the geosystem, to > both avoid absorbing too much energy and disrupting the planet's > inhabitants, while at the same time providing support for 10 Billion future > inhabitants. > > That might be based on a few principles of the sort you describe, and an > approach that doesn't require top-down agreement by all the stakeholders who > want to control it. > > Hmmm.... > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.postel.org/pipermail/end2end-interest/attachments/20100302/9255237a/attachment.html From jnc at mercury.lcs.mit.edu Tue Mar 2 11:51:25 2010 From: jnc at mercury.lcs.mit.edu (Noel Chiappa) Date: Tue, 2 Mar 2010 14:51:25 -0500 (EST) Subject: [e2e] [OT ?] NYTimes article on the structure & shape(topology) of the internet Message-ID: <20100302195125.CE5726BE5BB@mercury.lcs.mit.edu> > From: ashish makani > An NYT article on peering, > May not be very relevant to e2e, but seemed interesting to me. It was basically a pretty good/interesting article. It did have one significant howler in it (he attributed the Internet's ability to change traffic patterns to packet switching, rather than dynamic routing); I sent him a message explaining the difference, let's see if I hear anything back... :-) But other than that it was quite educational. Noel From day at std.com Wed Mar 3 06:04:17 2010 From: day at std.com (John Day) Date: Wed, 3 Mar 2010 09:04:17 -0500 Subject: [e2e] [OT ?] NYTimes article on the structure & shape(topology) of the internet In-Reply-To: <20100302195125.CE5726BE5BB@mercury.lcs.mit.edu> References: <20100302195125.CE5726BE5BB@mercury.lcs.mit.edu> Message-ID: Actually the biggest error in the article is that he didn't notice that this nothing more than backbone providers owned by non-telecomm companies. It does improve the connectivity of the network in general but architecturally there is no difference. The fact that they might focus on a certain type of traffic is coincidental. At 14:51 -0500 2010/03/02, Noel Chiappa wrote: > > From: ashish makani > > > An NYT article on peering, > > May not be very relevant to e2e, but seemed interesting to me. > >It was basically a pretty good/interesting article. > >It did have one significant howler in it (he attributed the Internet's >ability to change traffic patterns to packet switching, rather than dynamic >routing); I sent him a message explaining the difference, let's see if I hear >anything back... :-) > >But other than that it was quite educational. > > Noel From rsofia at inescporto.pt Tue Mar 2 22:56:43 2010 From: rsofia at inescporto.pt (rsofia@inescporto.pt) Date: Wed, 03 Mar 2010 07:56:43 +0100 Subject: [e2e] Fie on future internet Message-ID: <20100303075643.152438c4knfqf417@horde.inescporto.pt> Hello Jon, my 2cents on the "Future Internet" buzz... In addition to what has been discussed for the past 5 or 6 years, during ARCADIA, FIRE, etc., there are in fact some paradigm changes which are worth to be explored. Some initiatives (like EIFFEL) are trying to do that in a constructive way, by bringing together people :) that can ignite new topics and - the key point to mention - by asking the community to comment on such topics. Topics such as virtualization impact, SDR, control by the operators and the way this is being changed by user-empowered are IMO worth discussing. They are relevant and not well explored. As for FI stuff that may serve as an example to other fields...but hasn't this been done before? I recall from my previous affiliation (access vendor) that FI was the chicken of the golden eggs, in particular for access vendors which, in the beginning of the current crisis had to try to re-invent themselves and their focus. So most of them turned to "megacities", to energy-awareness, etc., and tried to apply networking and the Internet to these new goals...where exactly is the difference to what triggered this thread ? So some hints on how to solve this FI "issue"/fashion: 1. Differentiate (provide a categorization) between what may really be a teaser topic, and what for sure is not...Energy-awareness that is not green, or let's say energy awareness that just looks into making a box green for sure is a turn-off. 2. Stop calling it "future". The future is now (and it seems it has been since 55 :-D ). It's the Internet evolution anyway and we are already living it whether we like it or not. Why not call it disruptive Internet architectures/services? Isn't this what FI is all about if we consider a non clean slate approach? the fact is that we have a few services and architectures which disrupt Internet's operation, Internet stakeholders, as well as Internet basic principles... 3. try to understand which topics may be true paradigm shifts. What is here being named as "decentralization" and which seems to me to be "user-empowerment" (decentralization is simply a way to say that some control functionality is moving from common providers to users) will no doubt provide paradigm shifts, from an architectural, e2e perspective BR, Rute On 03/02/2010 11:40 AM, Djamel Sadok wrote: > > > SMALL Change --> Big leap! > The future internet is nothing but the same as the current one with few > twinks - small changes that may have big (business or technological) > impacts. No one is expected to bank or coordinate worldwide big changes. > It is about being disruptive. Just like the human brain, it still > maintains the basic old parts while adding new (smarter) layers. > There will be no day that one will say, today we have upgraded to the FI. > > Someone once said, too much government is bad government. That is the > case of the Internet and a lesson that governments may learn from it. > People will always come up with new ideas and if they work, others may > adopt them. Current FI R&D projects will never be "the" solution. > > CENTRALIZATION > What worries me about the Internet is its actual centralization. Take > away few search engines and most users won?t get their work done. > Hierarchy as Jon talked about is good, but I guess he meant it at the > structure level. This hierarchy is missing where it equally matters: the > content and service levels. > > May be the notion of federations, contexts, turfs, societies, or > whatever you call them, is the way ahead to break away from such > centralized business model. If like nature, the internet favors the > fittest, then we are doomed with a centralized structure. > > CONTROL > Governments, transport systems, banking systems, "telecom systems", > education systems, and others, like to think they have control and > achieve predictability. We are lucky they did not design the Internet > because they would have used to it to gain more control and kept to > themselves as a secret advantage or tool. > > The internet design community saw design ideas as more important than > control and today the internet may be seen as the first real autonomic > system. > > BIG PLAYERS > Many big players are today involved. Some are making money and others > see their shares dissipate. The big ones seem to dictate the changes. > They facilitate what they want users to have (make ebook readers, take > fiber to your home to sell you content, ..) it would be interesting to > evaluate to what extent their silent decisions impact the FI. My opinion > is they are the ones who are designing it right now. > > > The internet borrowed a great deal from areas such as economy, > transport, etc. to build its inner working mechanisms. Those systems > have been around longer than the internet. I would suggest that if those > areas are going to benefit from the Internet > it won?t be at the structural level (routing, buffering etc..), but > rather at the community, cooperation, security, information, levels. The > internet shows that: > > 1) a simple trust model may work and give its participant many benefits; > 2) even when there are no guarantees by the members, the overall system > may still deliver; > 3) when each of us contributes with a small resource, the overall system > may outperform any big system; > 4) people like free things even if they do not work well; > etc..... > > To complement Jon?s thoughts, "higher level" Internet lessons may be > best to use in other areas rather than its underlying mechanisms. > > DJamel > > > On Sun, Feb 28, 2010 at 12:27 PM, Jon Crowcroft > > wrote: > > I spend a lot of time reviewing > stuff about the future internet > then its raining on a sunday afternoon > so i wrote down what i really think:- > http://www.cl.cam.ac.uk/~jac22/out/fie.pdf > > > then i played the guitar to clam down > http://www.youtube.com/watch?v=UsJD5_nX5RE > > j. > > > Crucial public systems deliver food, energy, transport, > housing and so on. Three systems seem like low-hanging > fruit when it comes to re-application of the ideas behind > Internet: ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. From michawe at ifi.uio.no Wed Mar 3 01:04:49 2010 From: michawe at ifi.uio.no (Michael Welzl) Date: Wed, 3 Mar 2010 10:04:49 +0100 Subject: [e2e] Fie on future internet In-Reply-To: References: Message-ID: Hi Jon, all, Shouldn't we, as researchers, ideally be left alone to do whatever we want (as long as we produce some good, measurable output for those who pay us), in the hope that good stuff happens? Then, isn't the "future Internet" an entry ticket into the "do whatever you want" world that just makes "do whatever you want" look good and interesting to funding bodies? Do we really want funding bodies restrain what we do by announcing much more restrictive schemes? You cite Trilogy as a project doing good stuff. From their webpage: "Trilogy is a collaborative research project within the ICT theme of the 7th Framework Programme of the European Commission that contributes to the objective "Network of the Future" of the Work Programme." - shouldn't we be glad that such a funding scheme exists, allowing them to do their thing? I'd say yes, and Jon, I thought that you would be the first person to also say so... thus, while I understand that one can get bored from reading "clean slate yum-di-dum" again and again, I'm quite surprised that you would make a statement like that. (let alone Keshav suggesting to use it as an editorial!) On a side note, mentioning the architectural principle of "Abstraction" in conjunction with the Internet is a joke. If I'd have to pick one major thing that's clearly wrong about the Internet, then it's the *lack of* abstraction. Abstraction in the stack would allow you to exchange the actual elements in lower (that is, below the application) layers - an application opening a TCP connection or sending a UDP packet to a specified IP address isn't a very abstract concept, is it? Cheers, Michael On Feb 28, 2010, at 4:27 PM, Jon Crowcroft wrote: > I spend a lot of time reviewing > stuff about the future internet > then its raining on a sunday afternoon > so i wrote down what i really think:- > http://www.cl.cam.ac.uk/~jac22/out/fie.pdf > > then i played the guitar to clam down > http://www.youtube.com/watch?v=UsJD5_nX5RE > > j. From hgs at cs.columbia.edu Thu Mar 4 09:54:22 2010 From: hgs at cs.columbia.edu (Henning Schulzrinne) Date: Thu, 4 Mar 2010 12:54:22 -0500 Subject: [e2e] Fie on future internet In-Reply-To: References: Message-ID: <4145FF92-8805-4200-BAFA-8936171EDDF7@cs.columbia.edu> I'll just pick up on one analogy: "Government Governments are big and centralised. This is an inefficient way to organise what is essentially an in- formation brokering service. Perhaps the Internet and the web offer an extreme decentralisation model, with abstract interfaces and protocols that would scale gov- ernment to the small, personal level and to the large, far more effectively." That's generally not true in countries with a federal style of government (Germany and the US are examples I'm familiar with, but there are many others). However, living in New Jersey, this has distinct downsides. In NJ, we have hundreds (566, to be precise) of tiny municipalities, each with their own school district (616 for 8 million inhabitants, i.e., smaller than NYC), police force, fire department, mayor, council, department of health and their own foreign policy. Among many problems, there is lots of duplication of effort, lots of management overhead (each school district needs a superintendent and every 25-member police force a police chief) and a lack of sufficient local expertise and media supervision. The problem is not the routine administration of forms, records and payments, but rather decision making and avoiding "capture". New Jersey has exported many useful things, but I doubt that anybody would recommend its governance model... Thus, I'd be rather wary of blindly trying to translate networking concepts to other domains. For transport, modern GPS systems already provide mid-course re-routes around traffic jams, but I don't see how the BGP model would be an improvement over more centralized (if still regional) traffic information models. Henning On Feb 28, 2010, at 10:27 AM, Jon Crowcroft wrote: > I spend a lot of time reviewing > stuff about the future internet > then its raining on a sunday afternoon > so i wrote down what i really think:- > http://www.cl.cam.ac.uk/~jac22/out/fie.pdf > From ccie15672 at yahoo.com Fri Mar 12 17:59:02 2010 From: ccie15672 at yahoo.com (Derick Winkworth) Date: Fri, 12 Mar 2010 17:59:02 -0800 (PST) Subject: [e2e] 1993 RED paper equation... Message-ID: <482410.75954.qm@web57202.mail.re3.yahoo.com> In the 1993 RED paper, equation #3 on page 13 gives you a way to calculate what the weight should be if you want to allow a burst of L packets to pass without dropping any packets. Essentially, per the equation, you want to ensure the outcome is less than minth. Which is great, but is there an alternate version of this equation if minth and L are bytes rather than packets? Thanks.. Derick -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.postel.org/pipermail/end2end-interest/attachments/20100312/09b86a5a/attachment.html From amer at cis.udel.edu Tue Mar 16 17:43:25 2010 From: amer at cis.udel.edu (Paul D. Amer) Date: Tue, 16 Mar 2010 20:43:25 -0400 Subject: [e2e] Internet packet loss rates Message-ID: <4BA025AD.9040501@cis.udel.edu> Does anyone know of recent studies empirically measuring Internet packet loss rates/distributions? Thanks, Paul Amer Univ of Delaware