back to article Net neutrality: How to spot an arts graduate in a tech debate

Arts and humanities graduates are schooled for years in metaphor and analogy - and these are very useful skills for understanding the world. But what happens when an approach based on metaphor and analogy meets hard science and engineering reality? And what happens when the chosen metaphor doesn't fit? While you can choose …

  1. Tomato42
    Boffin

    "Furthermore, the assumption that traffic management is the cause of service differentiation is itself a narrow and misleading assumption. If you take away traffic management from a network, the network wouldn't suddenly become a Garden of Eden-like paradise. It probably wouldn't work at all."

    Wow, so much BS I need to put on protective goggles.

    It is possible to have network infrastructure that has more bandwidth than the consumers can use. Every Ethernet switch worth the box it came in manages it just fine.

    1. DavCrav

      "It is possible to have network infrastructure that has more bandwidth than the consumers can use. Every Ethernet switch worth the box it came in manages it just fine."

      Since we talking about networks by analogy, how about this as a arts-graduate friendly analogy: take the road network for the UK. It definitely had traffic management. How about we make that network neutral, by removing all rules for right of way, so that all road users are treated equally all the time. You want that lane and there's somebody already there? Not fair! Just move over anyway.

      OK, so we've decided that *some* traffic management is essential. The ethernet switch that you mention engages in traffic management. The question is, does more traffic management make things better or worse?

      The thing is, this is not about neutrality or management or anything. What this is about is people worried that ISPs will charge differential pricing to access certain parts of the net, or deliberately degrade access to those websites that don't cough up. A solution to this is to demand that all networks treat all packets equally, a legal solution to a technical problem that would make network switches illegal, since they clearly engage in packet management, holding on to one packet until it has cleared the last, etc.

      I don't know what the legal or technical solution to this is, but it helps to know why people are bothered about net neutrality.

      1. Joe Gurman

        Poor analogy

        A better one would be, "Imagine a UK motorway system where only vehicles belonging to major, international corporations could use the highways your taxes paid for, and you were forced to use one-lane roads with no passing places."

        1. TheOtherHobbes

          Re: Poor analogy

          Indeed. Claiming that net neutrality is the creation of twinkie know-nothing lawyers is pure FUD.

          The problem isn't traffic management. It means things like:

          The prospect of selective surcharges for certain kinds of content

          (Want to host video or audio? Pay the networks extra or they won't let your customers see it.)

          Creating selective bandwidth caps for some content providers but not for others

          Setting up walled-gardens with limited or nonexistent outside access

          It basically means breaking everything that makes the Internet so useful, and providing favourable treatment for large corporates at the expense of start-ups and individual content providers.

          Twinkie lawyers have nothing to do with it. Twinkie psychotic bean counters and corporate predators are much more the problem.

        2. T. F. M. Reader

          Re: Poor analogy

          If you want to go down that road (sorry...), how about the fastest motorways being toll roads (same price for everyone, and no one is excluded, but you have to pay it on top of the taxes if you wish to use high bandwidth, low latency pipes)? And if you pay a moderate fee in advance your packets get a special flag in the headers so that they don't sit in buffers at ingress or egress points and are switched using separate high priority queues with lower latency? Will that still be considered neutral?

        3. ipghod

          Re: Poor analogy

          maybe not better, but more accurate..

          imagine every place has a private driveway you pay to maintain, and every intersection has a toll, with your toll based on how often you go through.

          certain areas decide to flat rate the intersections, to save time and effort, with everyone understanding they are to attempt to maintain equitable usage and not hog the roads.

          and then a Walmart buys the land next to the intersection, connects it's parking lot to a single road at the intersection, and floods that intersection with traffic. the people running the intersection have to install lights to keep traffic moving through the now overloaded intersection.

          in the political 'net neutrality' world, Walmart is fully justified in claiming the maintainer of the intersection should add more roads, widen all the feeders, and do away with the traffic lights, because deep inside one of the neighborhoods that helps fund the intersection, there is a farmers market, a butcher , and an auto parts store, so clearly, installing traffic lights was designed to hurt walmart's business.

          asking Walmart to pay for widening the intersection, or bypassing it completely, and extending their parking lot directly to the neighborhood is 'unfair', because people want acces to Walmart, and how dare you stop them from getting there! freedom to travel!

          net neutrality as a movement is well intentioned, but needs to stay grounded in the reality that any 'anti neutral' behavior requires actual proof of an intention to harm someone, not a conspiracy theory based on a metaphor.

          1. SImon Hobson Silver badge

            Re: Poor analogy

            > ... and then a Walmart buys the land next to the intersection ...

            Another broken metaphor.

            Every person with a drive is paying for access to the road networks - specifically a per-month fee based on the capacity they require. A big chunk of that fee goes to the organisation that's running the local roads, but another chunk of it goes to the people running the bigger roads your local road network connects to - by analogy, the former group is your ISP, the latter group are the tier1 and tier2 carriers they buy their connectivity from (fairly loose analogy, but it's good enough to debunk yours).

            So there you all are paying your fairly modest monthly fees which allow you up to a certain number of movements per day - typically no-one other than a small group will get that many car movements unless they're having a party and lots of people come by car. Yes, a Walmart or ASDA or Tesco or ... can come and set up - but they'll find that if they try that stunt then they'll have few customers once they've used their daily allowance within 10 minutes of opening time.

            So when they want to build their store they will do one of two things :

            They'll either go and connect to one of those road networks where they can buy a high capacity connection, or they'll work with the local roads provider who (in return for a wodge of cash) will upgrade their network so as to support the extra traffic.

            So your analogy is exactly what would happen if (say) a Netflix decided to start up and rather than paying for a decent connection just ordered a couple of domestic ADSL lines ! They don't do that, they shell out a lot to connect to the networks with the capacity to handle their traffic.

            Where net neutrality comes in is that there you are with your driveway connected to the local roads, and rather than driving to Walmart ... you go online and order for delivery. The people who run your local network don't like Walmart because it competes with the store they run - so they setup checkpoints and deliberately pull in all the Walmart vans trying to deliver your shopping unless Walmart pays them extra (or you pay extra).

            Don't forget, you have already paid for having access to the roads, Walmart have already paid for their access to the road system at their end, but someone in the middle wants extra just because it's Walmart on the side of the van.

          2. Dave Howe

            Re: Poor analogy

            That is a poor analogy; let me explain.

            Lets say your local road network is owned by a company - you have to pay to connect *your* driveway to the network, and they charge you based on how much weight comes into or out of your driveway (in most cases, they expect the weight into your driveway to be so much larger that they only bother measuring that) - but you are paying based on usage. You reasonably expect that, once you drive off your driveway onto the network, that the speed restrictions are imposed in a equal and fair way to make sure you get your fair share of the benefits of driving - but do does everyone else. Usually, you don't have much choice on which road network to join (after all, it's outside your house!) but on the whole, they don't want the Government stepping in and imposing price restrictions, so they keep their usury within reasonable bounds.

            Walmart opens a store near a road network owned by a *different* company - that company is thrilled - the amount Walmart needs to pay to connect their parking lot to the road network is a *lot* more than they would get from the average home driveway, and Walmart expect that as part of the cost of doing business. So, all well and good - when you go to Walmart, that trip has been paid for twice; once by you, to get onto the network, and once by Walmart, to get off the network at their Store (and of course, the reverse to come back).

            The problem arises when your company looks at how much money the *other* company is making from Walmart, and starts thinking "how can I get some of that? I want more money". So they try charging the other company to let their customers go from their own road network to the Walmart provider's network. This starts a major battle, where everyone suffers - your company either won't let you get to the Walmart provider's company, or forces you to go via a third or even fourth company's roads, taking much longer and adding traffic to roads that shouldn't need to carry it, simply as leverage to try and extort money out of the other road company.

            After a while, the dust on that one settles down - a few smaller road companies are now paying, but the larger companies are in a mexican standoff, lots of money has been spent (and continues to be spent) on lawyers, but the end result is that that money is leaving all the road companies (and going to lawyers) and no new money is coming into the system, so it's a loss for everyone (and yes, there really was a peering war, and it is currently in a standoff)

            Your company is now upset. They tried charging another company, and on the whole, made a loss on the whole thing. They are already charging *their* customers as much as they dare, and even charging them more than agreed making them "service charges" and hoping that they don't notice - a lot like surcharges on holiday packages. Then, they have a bright idea. Claiming how unfair it is that so many of their customers drive to Walmart, they start deliberately slowing down or stopping their *own* customers if they are driving to Walmart - then tell Walmart that they will only stop doing that if Walmart gives them money to stop doing so. Walmart says "WTF? you are already being paid for that traffic by your Customers, what business is it of yours if they are coming here! We pay a LOT to our road company for access to that customer base, so take it up with them" and the road company says "well look at it this way. We will charge any of your competitors too; that means you aren't in a worse position, but any small businesses that get caught up in this will be unable to compete, so that extra business will come to you". At this point, the Government looks at the situation and says "WTF guys? you are already being paid for this, and you are getting to the point that you are deliberately making the market anti-competitive by degrading your own service for purposes of extortion. Stop it or we will make you stop it" - to which of course the companies reply "lol, you can't, you have no authority to do that"... which brings us to the new laws on Net Neutrality. NN isn't about if filtering and control should happen, but about motive - filtering should be for legitimate service-improvement purposes, not extortion.

      2. Graham Marsden
        Boffin

        @DavCrav - take the road network for the UK.

        > How about we make that network neutral, by removing all rules for right of way, so that all road users are treated equally all the time.

        You miss the point.

        Imagine you're on a three lane Motorway and find that Lane 3 has been bought by the Ford Motor Company

        That Lane has a 100mph speed limit but can *only* be used by people driving Ford Cars.

        Meanwhile Lanes 1 and 2 have to take all the other traffic and have had their speed limits reduced to 60mph and even then you'll be lucky if you can do that because you've got "crawler races" with one HGV is trying to overtake another at 1mph faster.

        *THAT* is what Net Neutrality is about, minus all the petty sneers about Arts Graduates and other such BS.

        1. DavCrav

          Re: @DavCrav - take the road network for the UK.

          "You miss the point."

          Did you read past my analogy, to where I make the same point as you?

        2. jeremyjh

          Re: @DavCrav - take the road network for the UK.

          As an arts graduate (the horror!), I'm glad you said that.

          That said, even the fact that I'm typing this here probably marks me out from other arts grads.

          1. Scorchio!!

            Re: @DavCrav - take the road network for the UK.

            As an arts graduate who moved on to the hard sciences, I find arts graduates have among their number an unusually high percentage of people who appear to think that reality has to change, no matter how impossible this may be. This need not be inherent to arts courses, but due to the kind of individual that applies for arts courses as a 'soft option'. (You may not be one such.)

        3. Anonymous Coward
          Anonymous Coward

          Re: @DavCrav - take the road network for the UK.

          Imagine that you're on the M4 and the outside lane has been given to the bus companies. All other traffic has to crawl along behind HGVs having "crawler races". Except for politicians, who can whizz along in the empty lane because they are more important than you.

          Oh, you don't have to imagine it. It's real.

          Sorry. Having a bad day. Stretching the analogy, people wouldn't mind it so much if they were regular bus users, and occasional car-users. In the same way, if their emails took a second longer to arrive, or a static web-page refreshed a bit slower, but their cat videos or Skype calls were uninterrupted, they would probably be happy. But this would be done for the benefit of users, not companies. So fat chance.

        4. Hollerith 1

          Re: @DavCrav - take the road network for the UK.

          Thank you Mr Marsden and all others who have made this point. It is easier to sneer at 'arts graduates' and their pathetic lack of understanding of 'packets' than to grasp what their argument is about, which is the grabbing of what was a 'fair and open to all' (as much as that can ever exist) system and changing it for the benefit of big business. Because that always benefits everyone, doesn't it? As in, nowhere.

        5. Dr Stephen Jones

          Re: @DavCrav - take the road network for the UK.

          Road metaphors are what arts graduates (like Al Gore) use to describe packet networking. Road is suitable for phone networks but not packet networks. There is no analogy for a road that dynamically reconfigures itself. The capital cost of a new dedicated toll lane has no analogy in multiplexed networks.

          Sorry, but you are ignorant and have just proved the premise of the article.

      3. jonathanb Silver badge

        Taking the road analogy, you can have things like traffic lights which switch between roads on a junction, giving each one a chance to access the junction in turn, but apply equally to all types of traffic on that road. That's net neutrality. You could also have things like bus lanes which give one type of traffic, buses, priority over other types of traffic. That's not net neutrality.

        Today, when driving home, I was delayed for about 20 minutes because a lorry ignored a "no HGVs" sign and went down a narrow road towards a single track bridge that it couldn't get across, when it should have gone about 3 miles upstream to the next bridge. That is an example of where net neutrality would be a bad thing.

      4. Pat 4

        That is an extremely dishonest and weak argument. You equate neutrality with anarchy, which is stupid, and base your rebuttal on that turd... Just like the article does. Utter bovine excrements.

    2. Chris Miller

      It is possible to have network infrastructure that has more bandwidth than the consumers can use.

      You might be able to achieve this on your own premises, though I think it's unlikely if you have more than 100 users. But for an ISP connecting thousands of domestic users (which is the only place where 'net neutrality' has any relevance, as commercial organisations will have their own Ts&Cs) - not a chance.

    3. Paul Shirley

      It's not about removing traffic management though, it's about preventing abuse of traffic management. What we're lacking is any credible definition of what abuse we're trying to stop, the cluelessness of the non technical on how abuse could be detected is pretty unimportant right now - but the idea that deliberate manipulation can't be detected in a torrent of randomness is astonishing and a lot of scientists have clearly wasted their lives doing exactly that in many fields of research.

      1. Neil Davies 1

        That's not precisely what the report says.

        It says that a) if you suspect differential treatment you can't localise it to where in the delivery chain it occurred with the current tools [hence not "fit-for-ofcom's needs" - who do you chase after and how?]; b) there are lots of differential treatments that can't be detected with the current tools; c) The differential treatment is the wrong question - you can test for a quality floor much more easily.

        All those scientists who look for "signals" in randomness have models for the signal they are looking for, there isn't a model to test against at present (and "net neutrality" is an aspiration not a model)

        disclosure: I was one of the reports authors.

        1. Anonymous Coward
          Anonymous Coward

          All is not that easy

          Apart from Orlowski and the large cable and mobile networks, no-one minds if SMTP traffic (all smtp traffic) gets lower priority than http for example.

          As has been said many times, it isn't about the type of traffic you are managing its about management by source. Packet racism (if you like highly emotive analogies) where the management is based on where it comes from (or where its going) rather than the content.

          Where we run into problems is when you have protocols which are source/destination specific or which compete directly with ISP-provided services. In HR it would be called, "indirect discrimination" and would include things such as setting a height requirement for employment so that south-east Asians would never qualify, or specifying only people with blonde or light-brown hair can will be accepted. The ISP equivalent is Facetime. You could say, "well, we treat all facetime packets equally" - they all get treated like torrents and SMTP - but that would be indirect discrimination because effectively you are victimising one particular group. What if someone starts using port 25 for voice traffic - do we then upgrade all SMTP to voice-level priorities? Never underestimate the ability of American lawyers to disregard common sense.

          This is why we have well-known ports. Perhaps IANA will become so much more important than before. Do we make IANA the standard by which all traffic types are categorised? I don't really see a particularly better way. What happens to VPN traffic? Is that deprioritised? Does your voice traffic go horribly wrong because you want to protect it from snooping?

          There may be no easy answers to all of the questions, but this we know for sure - no discrimination based on IP source/destination. If you mess with different protocols too much, you'll get things like Google's HTTP/UDP combined with chunked transfers to do even more obscured torrents. I'm sick of everything trying to run over http. just downgrade the stuff you don't doesn't matter that much, upgrade the stuff know does and stop pretending that streaming prevents piracy. Keep the consumers as the customer, keep a single published price list for data leaving the ISP network if data isn't being hosted there.

    4. theblackhand

      Re: some protective googles

      It is possible to have more bandwidth the the consumers can use, but it generally begins to fall down when:

      - you start adding more boxes and eventually reach a point where there will be bottlenecks once vendor slot or port limits are hit

      - you start adding latency

      - you start adding costs

      - you start adding third parties (and their design choices) to reach the destinations you desire

      We have had (and still have) networks that deliver the bandwidth they promise, they just happen to be significantly more expensive and most people don't want to pay for the extra "quality". i.e BT Infinity 1 costs £27.99/month versus a 10Mbps Internet link in London costing around £500/month (rising to around £850/month for 40Mbps). This may not guarantee you the bandwidth to the destination you want, but working with your ISP (or changing ISP) may allow you to remedy that for more money....

      I realise this isn't a perfect example (i.e. different providers charge differing amounts for differing levels of service - i.e. Cogent/Sprint vs AT&T/Verizon in the US) but it's a significant step up from treating the Internet as a single Ethernet switch.

    5. Anonymous Coward
      Anonymous Coward

      straw fuel

      strawman burnt to a crisp.

    6. Destroy All Monsters Silver badge
      Facepalm

      It is possible to have network infrastructure that has more bandwidth than the consumers can use.

      35 upvotes.

      I though this was a tech site?

      Oh well. Maybe economics is no longer even needed for modern button pushers.

      "If Obama can promise 5 billion of "special pension outlays", why not 500 billion??"

    7. JohnG

      "It is possible to have network infrastructure that has more bandwidth than the consumers can use."

      Yes - but on average, consumers aren't prepared to pay for an Internet which has more bandwidth than they use.

  2. madofo

    Enterprises buy connectivity solutions with 'quality floors'. I think they're paying more than $20/month for that service.

    1. dotdavid

      I think we're more used to "quality ceilings" around here

    2. <shakes head>

      that depends on the level of the "floor" i guess sub basment would be cheaper the penthouse :¬)

  3. Anonymous Coward
    Anonymous Coward

    Flufflepuff!

    Thanks for the gratuitous Flufflepuff /pfudor picture. Brought a smile to my face.

    AC for obvious reasons.

    1. Anonymous Coward
      Anonymous Coward

      Re: Flufflepuff!

      I thought I clicked onto one of my daughter's YouTube favourites by mistake. Now I can't get that damn tune out of my head .... Pink fluffy unicorns dancing on rainbows, pink fluffy unicorns dancing on rainbows, pink fluffy unicorns dancing on rainbows ....

      1. Chika

        Re: Flufflepuff!

        This'll do it.

        https://www.youtube.com/watch?v=QQWhJrkgXII

        (\

        1. This post has been deleted by its author

    2. Anonymous Coward
      Anonymous Coward

      Re: Flufflepuff!

      Now looking for hints of Orlowski's presence on Equestria Daily.

  4. Anonymous Coward
    Anonymous Coward

    "How you spot an arts graduate in a tech debate"

    They're serving the coffee.

    1. TheOtherHobbes

      Re: "How you spot an arts graduate in a tech debate"

      >They're serving the coffee.

      Actually they're probably running the political system or lying to you while pretending to be journalists.

      There's a reason politics is full of lawyers and has almost no engineers in it - it's because too many engineers and developers are shockingly easy to mislead, misdirect, scam, exploit, and manipulate.

      As far as the humanities grads are concerned, knowing how to code or put together a network doesn't make you a master of the universe - it makes you the help.

      1. Anonymous Coward
        Anonymous Coward

        Re: "How you spot an arts graduate in a tech debate"

        I no longer know if I belong in this world!!!... as a BA Behavioural Sciences from UEA I emerged with an excellent understanding of stochastic modelling, group dynamics, kinship networks, shortest spanning trees, anthropology, sociail psychology, FORTRAN and COBOL programming, sociology, system design, business analysis, the historical and philosophical bases of our economic and political systems, linear and non-linear multivariate programming, I-O Analysis, hyperspaces and eigen values, J.S. Mill, Bentham, Descartes, Milton Friedman, Hobbes, Nietzsce.. uncle tom cobbly and all. Amazing what you can learn in 3 years as an 'arts graduate'!

        Its been very useful when working at IBM, speaking at computer conferences on performance modelling and 'glass house network' design and performance to keep those blocks and packets flowing to and from the processors.

        ..and I can agree with most of you all. The issue here is an economic one i.e. behavioural, and yes it is all about grabbing control of the supply chain of a 'basic utility' good so that illegitimate value can be extracted from the assured demand for it. Its a re-enactment of the historical cycle of discovery, application, creation of value, robbery, constraint and ransomed selective provisioning.

        Neil Davies summarised 'his' report well. It defines the real world constraints faced by anyone seeking to create a system of governance for it. From there we have two discussions going on. a) the technical issues around management, measurement and diagnostic analysis of stochastic processes (that's like econometrics.. BA stuff) and b) the 'political economy' issues and deeper philosophical issues as to whose world-view should form the basis of a universal belief system and hence the ethical basis of the system of governance the Internet should have.... or should we smash it up into 'domain-states' each choosing the robber-baron approach all 'good' Aristotleans, or the universalist 'humanitarian' approach of 'good' Socrateans. Do you want to dominate people and extract value to yourself and your elite mates (Aristotle, Tory, Republican, Oligarch, Theocrat, Dictator, Tyrant)? Or, do you want an open and free internet in which you always get the service level you pay for, much like an IT Dept Service Level Contract with the rest of the company has delivered over the past 50 years?

        The first option is easy... it simply extends our dominant, and really quite localised, socio-economic commercial model into the unitary global network domain. The PR guys will call it 'economic freedom', 'e-democracy' and other misnomers and we will al believe it - just like we 'believe' political spin doctors... and if we don't the media will shape our perceptions until we do.

        The second option is more difficult... we first have to decide what we believe and why, how to create the economic model and hence the tools (the network tools in this case) needed to implement and manage the model so that we get the type of (network-) world we actually want to see. That has not been done... and it won't get done until you have all sat down and done it for the real world out there. The one you live and work in, the one many struggle and die in, the one that tells you that more productivity and more efficiency from fewer and fewer people will give every single person higher and higher levels of consumption. The one that tells you we will all have the money to pay for the products we need to consume to fulfil our needs because those of us not producing for us will be serving the immediate well-being needs of others. The ones that DON'T tell you that ultimately you are all paid through the excessive conspicuous consumption of those few controlling the supplies of input materials and input labour. the ones that DON'T tell you that a minimum wage will always be the norm for most because frivolous services will never be demanded in conditions of shortages of the labour to provide them. From time to time those few will vote amongst themselves (by proxy through your media driven hand at the ballot box) to decide the next set of executives in the governance service industry. The democratic model was conceived 2,500 years ago under the same conditions... 10 suppliers of service.. mostly low paid... to every consumer of services. That was Greece 450BC.. 10 slaves to every 'citizen'.... Socrates had to drink hemlock.... while Plato looked on... and went on to teach Aristotle. The battle continues.... What sort of Internet do YOU WANT? Then decide on the the tools and methodology... if you choose right we will have to commission the technology we need, just as we already can see the need to re-invent our socio-political governance technology in the real world.

        Never, never make the mistake of thinking Techies can do without the Humanities... AND VICE VERSA!!

    2. Hollerith 1

      Re: "How you spot an arts graduate in a tech debate"

      Acutally, arts graduates are probably running the company you work for. Check out the CVs of your CEO and so on. They might have an MBA, but some might be analysts or financa/accountant people, or even English majors. It's a fox and hedgehog world and, while hedgehogs can do one thing well, they mostly can only do one thing.

    3. Michael Wojcik Silver badge

      Re: "How you spot an arts graduate in a tech debate"

      Ha. I have degrees in both the humanities and the sciences, which is why I secretly control everything.

      But perhaps I've said too much.

  5. Kubla Cant

    It's easy enough to spot when Stephen Fry offers a ridiculous explanation of a technical subject on QI (as he did here and here and here - just two examples amog many).

    That's what's wrong with arts graduates like Stephen Fry, they don't know that 1 + 1 + 1 = 2.

    1. Naselus

      "That's what's wrong with arts graduates like Stephen Fry, they don't know that 1 + 1 + 1 = 2."

      I thought 1+1+1 = 7?

      1. Vic

        I thought 1+1+1 = 7?

        1+1+1 = 11, or course. Duh...

        Vic.

      2. Groaning Ninny

        I can make 1+1+1=2, and I can make 1+1+1=4, but I'll admit I'm not getting the 1+1+1=7

        1. GrumpenKraut
          Boffin

          Concatenate, then read as binary. If you replace "binary" with "K-ary" you get (K^3-1)/(K-1) = K^2+K+1 for any K >= 2 of your choice.

  6. Confused Vorlon

    Re: Re: Just another attempt

    I have lived in a country where the telephone monopoly deliberately downgraded skype traffic so that you would have to use their higher-priced phone service.

    That's a pretty clear example of a non-neutral network.

    Non-neutrality is like monopolistic behaviour in a business context; Sometimes it is easier to spot - sometimes it is more subtle and harder to spot - but it is still a bad thing, and we do well to make it illegal.

    1. jerehada

      Re: Just another attempt

      Which is the monopoly - Microsoft ?

    2. SImon Hobson Silver badge

      Re: Just another attempt

      > I have lived in a country where the telephone monopoly deliberately downgraded skype traffic so that you would have to use their higher-priced phone service.

      That doesn't narrow down the list of countries much !

      I have been procuring and using WAN/Internet connections for probably around 1/4 century. One thing I have observed is that whenever a new technology comes along, the incumbent (BT in this case) will do all in it's power to hang on to it's cash cows.

      With ISDN they crippled certain functions to protect their leased lines cash cow. With ADSL they "held back" on deployment to protect their leased lines cash cow. When FTTC came along, they 'held back" enabling cabs in primarily business areas to protect their leased lines cash cow.

      In some places they have announced "non availability" of (eg) FTTC to an area (typically, but not exclusively, rural villages) - only to have a change of heart when someone comes along and offers something else. In some of these cases, the "imminent" arrival of FTTC quite coincidentally fails to materialise once the alternative offering is off the table (typically because the competitor withdraws or goes bust when enough customers decide to "wait for the BT option") !

      That is what monopoly or quasi monopoly business do if they are not controlled by regulators. It's why we have regulators.

  7. Anonymous Coward
    Anonymous Coward

    Smug-Mode Enabled

    As commendably smug as this article is, many techies understand net neutrality at the macro level of throttling some sources such as Netflix to be a clear and addressable issue, The Register included:

    http://search.theregister.co.uk/?q=net+neutrality&advanced=1&author=&date=the+dawn+of+time&site=0&results_per_page=20

    So to the extent that lawyers, politicians and the like have formed these opinions, ask yourself where they have come from in the first place, and to what extent, at the macro, policy type level, they are actually wrong.

    1. Dan 55 Silver badge

      Re: Smug-Mode Enabled

      Also worth noting "best effort" is a technical term which can be found in many telecoms engineering manuals and RFCs when referring to QoS. Once you get into the wonderful world of cablecos charging for content and so on then of course they redefine everything as suits them, including QoS.

  8. Zippy's Sausage Factory
    Devil

    Er, but...

    Let's say I run a dodgy ISP. I require you to use my DNS servers, so I block calls to 8.8.8.8 and 8.8.4.4 directly, thus I control your traffic*. So I filter NetFlix through a proxy that checks who you are. Do you pay the extra 5 simoleons a month for faster Netflix? No? Well my proxy limits you to 1mbps then. Do you pay the extra 5 simoleons a month for better Facebook? No? Well then my proxy limits you to 56K a second on Facebook.

    Net neutrality isn't really about detecting magic kit that ISPs aren't using yet. It's about future proofing the law to prevent ISPs getting creative and installing ways to screw punters for another fiver a month for the "upgraded" service so they can watch cat videos on YouTube...

    * Yes, simplified. It's just an example...

    1. Anonymous Coward
      Anonymous Coward

      Re: Er, but...

      That would be a great example of the need for net neutrality.

      Does such an ISP exist?

    2. The Axe

      Re: Er, but...

      What's wrong about letting ISPs charge different amounts for different levels of service? Maybe I should ask for smart phone neutrality. It's not fair that Apple can make crap phones* and charge massive amounts whilst I have to make do with a OK phone.

      * Clickbait for Apple fans. IGMC

      1. auburnman

        Re: Er, but...

        This isn't about different levels of service; this is about stopping ISP's from double dipping (or even arguably triple dipping) into people's pockets and/or crushing competitors. If I pay for internet access, and Netflix pays for internet access, what right does any ISP have to say that we have to pay extra to stream content? Would you feel comfortable if Sky broadband throttled Netflix unless you paid extra but left their own streaming service alone?

      2. Aedile

        Re: Er, but...

        The protest isn't about charging for different levels of service since they already do so. IE $50/month for 1 Mbps up/down or $80/month for 10 Mbps up/down.

        The problem is after I have paid for that level of service they then want to charge again to ensure specific things go the speed I already paid to receive. If I paid for the 10 Mbps plan then I should get Netflix (or any other service) at that rate. The ISP should NOT be slowing it down to 1 Mbps unless Netflix (or I) cough up more money.

        To use your phone example. If you bought a smart phone you expect to receive the same quality connection regardless of what company you call. If it was like the net then you will find you only get clear stable calls to companies that pay extra. If you call companies that don't pay extra most of the time the network will be busy and the times you do get through the calls are staticy and drop regularly. Furthermore, due to the high rates only big established companies can pay so it harms all small local businesses since no one can have a decent call with them.

      3. relmasian

        Re: Er, but...

        The Axe, not much is wrong with charging ISP customers for different levels of service. Indeed, ISPs have been doing that from the beginning, including under Net Neutrality. What they are not allowed to do to give an advantage or disadvantage to data packets coming to their network from outside their customer base. Under Net Neutrality, Internet traffic should be handled the same no matter whether it originated from the East or West Coast, even if one of those places is in the UK and the other in the US.

    3. Anonymous Coward
      Anonymous Coward

      Re: Er, but...

      Can we make that the norm for all (anti-)social media sites. Let's even go one stage further and limit the access to something slower. IMHO, 19.2K is plenty fast enouth got those wastes of space.

      Why am I so anti?

      A very good friend of mine lost his daughter to cyber bullying.

      All sorts of madeup stories were posted about her online including that she'd slept with 20 men in one night. The police (the don't live here) wouldn't do a thing about it. In the end she... well you know the rest.

      1. Hollerith 1

        Re: Er, but...

        @AC - the story you tell us is tragic, and this should never happen, but I don't think throttling access will solve cyber bullying, because creeps are often happy to pay to be able to wreak their havoc. This girl needed protection and help outside of the interent.

      2. Anonymous Coward
        Anonymous Coward

        Re: Er, but...

        This whole debate reminds every other debate which focusses on the rights and obligations of people towards 'economic free goods'.... air, water, land, sunshine, plants to eat, routes to travel along, routes to transmit emf radiation across, routes to transmit electrons over...

        Every one is seen as a single 'something'. But they are not. Each is like a blank sheet of paper. It can be used for innumerable purposes by innumerable people. WHAT IS YOUR INTERNET? A replacement for the 'penny post'? A replacement for the morning newspaper? A replacement for the evening down the pub with your friends? An community policing 'bobby on the beat'? A corporate telephone tie-line network around the world? A TV service? A cinema? A sports stadium? A remote control panel for your factory floor? AN ADVERTISING DELIVERY CHANNEL?

        We have rules and regulations governing almost every aspect of use and implementation of the 'legacy' versions of every one of these 'products' and 'tools'. We often have government departments and ministers dedicated to each one of the areas they serve.... but they are domain specific. We do not have a global 'Department of Transport' and supporting body of legislation for each modality. They are confined to national boundaries. We all use similar vehicles, and can use them cross-border, but have to operate them under different sets of rules as we cross boundaries according to what vehicle we use and for what purpose we are using it.

        We accept toll-roads, bus lanes, car-club lanes, diversions and outright closed off no-go areas. If you are all happy with that model, which also has some international agreement on basics, then look no further for ideas.... But don't be surprised when when Internet equivalent of Egypt, France and the UK choke off 'their' Suez Canal, or the USA and Panama choke off 'their' Panama, or Turkey 'their' Bosphorus... the fighting will continue.

        Or you Techies can learn from the History of transportation and the Political Science face of economic competition and use what you find axiomatically useful and create real solutions for this new real world problem... time to 'stand on the shoulders of giants'... like we always have had to do in the past, and still do. The problem is sociological, the solution is economic and political... its implementation will rely upon our technology. Time to get togetther!!! Solve this one and you will change everything..... A real technological revolution? Viva!

  9. DavCrav

    "How you spot spot an arts graduate in tech?"

    How you spot spot a tech graduate using English?

  10. Anonymous Coward
    Anonymous Coward

    I have always taken "Net Neutrality" to be a statement about the business practices rather than the mechanics of the network. Traffic shaping according to traffic type and network capacity is different from deliberately discriminating for/against a particular endpoint's content.

    1. Anonymous Coward
      Anonymous Coward

      Good summary.

    2. Anonymous Coward
      Anonymous Coward

      This. The problem is differentiating between the two given samples collected at or near the endpoints. Now if we have access to the whole network for our sampling, then the problem is only horribly difficult rather than impossible.

    3. tom dial Silver badge

      But: infrastructure to consistently deliver 99.9% or more of packets with no more than 20 ms latency may be costlier to build and operate than what is required for delivery of email where minutes of delay is likely to be somewhere between unnoticed and quite acceptable. Many of the comments here seem to relate to how the differential should be billed.

  11. Joe Gurman

    Except that....

    As interesting as Mr. Orlowski's presentation of the report, and the principles underlying them, may be, a theoretical net neutrality or its impossibility is obfuscation: what most people mean by the term is not countenancing deep-pocketed players' doing better than stochastic usage would predict, all the time, while the rest of punters always do worse. Would Mr. Orlowski feel less aggrieved I f we pushed our glued-together, Coke-bottle lensed glasses higher up on our noses, adjusted our pocket protectors, and held up banners demanding a statistically fair network, where Netflix, Amazon, and each of has the same expectation of latency and bandwidth as a fraction of advertised "up to" values?

    If nothing else, one of the principal advocates of net eutrality has been Google, er, Alphabet, not a firm known for the liberal arts (as we call it in the States) background of its senior management.

    1. Chris Miller

      Re: Except that....

      The reason that Google advocates net neutrality has little to do with technological but a very great deal to do with money.

    2. theblackhand

      Re: Google

      Google does have a significant investment in this game - they are rumoured to have a larger existing network than any of the other US providers and contribute around 50% of total Internet traffic (i.e. YouTube), but they have a relatively small number of users for their ISP business relative to the big players.

      A costly regulatory fix hurts their potential competitors if they are looking to expand their ISP business.

    3. auburnman

      Re: Except that....

      I can't find the link but I could swear there was an El Reg article about Google reversing their Net Neutrality stance once they started laying fibre cables. Anyone remember it?

    4. tom dial Silver badge

      Re: Except that....

      As a Netflix and occasional Amazon streaming customer I consider it in my interest if the source, intermediate, and delivery communication providers can provide consistent unpixelated video content delivery that is within the buffering capability of my TVs. That may cost more in network investment and operation, and I see no reason that the cost necessarily should be recovered in part from those who are not Netflix or Amazon customers, and it makes no difference to me whether the collection is directly from me or proxied by Netflix and Amazon, who would pass it through to me anyhow.

      My guess is that network neutrality works best when network capacity at all points exceeds traffic by a factor of 5 to 10 and begins to break with increasing frequency as that factor declines from that.

  12. Joe Werner Silver badge
    Boffin

    Stochastiv vs. deterministic

    > "Stochastic" means it evolves over time in a deterministic way.

    Sorry, but no. Just... no. Stochastic is non-deterministic. Chaotic is deterministic, trajectories of chaotic systems can be described by ordinary differential equations, and knowing the initial state exactly means we can (in theory) predict the exact trajectory forward in time. Unfortunately we cannot know the state exactly, and the uncertainties in state space increase exponentially in time. Thus, there is some limit time after which our forecast is completely uncertain and smeared out over the attractor of the system. There are some cases where it does not matter, any chaotic system observed with sufficiently low temporal resolution appears to be stochastic and can be modeled as such.

    1. Neil Davies 1

      Re: Stochastiv vs. deterministic

      Agreed - that was a "typo" due to journalistic deadlines - has been updated.

      Technically stochastic is not really non-deterministic (that means you know nothing!) - it expresses levels of uncertainty in the evolution of the system. Interestingly it does have analogies with chaos, but network protocols can take a stochastic system and make the outcomes become more determined (i.e. keep the system within its predictable region of operation and, for example, make the time to complete of a large file transfer more predictable) or, it the presence of nastier forms of non-determinism (which is actually what happens in today's networks) the emergent properties can vary widely - just as you experience.

      Disclosure: I was one of the report's authors.

  13. teebie

    '"Stochastic" means it evolves over time in a deterministic way'

    No, no, that's not what that word means.

    It's fairly close to the opposite of what the word means though, so well done.

  14. dcluley

    Stochastic is not deterministic. Stochastic phenomena can be analysed statistically eg we can say with absolute certainty that exactly half of a lump of radioactive material will decay in a certain time; we can never determine which atoms will be the ones to decay in that time.

    1. Hollerith 1

      Can analyse but not predict

      Stochastic, unlike chaotic, can't predict the way [whayever it is] will end or evolve, but you can get probabilities. This is not a hard concept.

  15. Julian Bradfield

    Um, "stochastic" means the opposite of "evolves deterministically". It means something that behaves randomly, so that analysis has to be done statistically rather than deterministically.

  16. keithpeter Silver badge
    Windows

    Really?

    "‘Stochastic' means it evolves over time in a deterministic way."

    Really?

    PS: downloading the report now, thanks for highlighting it.

  17. Kristaps
    Headmaster

    Stochastic != deterministic

    "‘Stochastic' means it evolves over time in a deterministic way."

    Ehrm.... no. Stochastic is the opposite of deterministic. Ironic that this is in an article on bad comparisons.

  18. Bob Starling

    Prompted a memory

    Of an arts graduate I knew who was employed by a large arms company as an experiment in providing an alternate view to the engineering types. He soon found that he and his colleagues became known as Snotty Nosed Arts Graduates aka SNAGs

    Hitting a SNAG had a whole new meaning.

    1. lucki bstard

      Re: Prompted a memory

      Yep, and I bet they were paid equal or more than the engineers; whilst they also had more relaxed degrees. Plus they can take the experience from that role and get themselves into middle management... Work smarter not harder.

      How many IT staff would be earning more as a plumber then sat in a cubicle damaging our health in front of a keyboard?

    2. Hollerith 1

      Re: Prompted a memory

      One of many degrees you can take in the Arts is Philosophy, specifically, ethics, and how to intuit moral and immoral outcomes. Or History, to see how things were done in the past and if that worked or not. Or law, which teaches you to argue from evidence, logic and reason. Or Literature, which opens to you the vast ways humans have communicated ideas, stories and beliefs throughout our existence on earth. Or Fine Art, where the craft of the beautiful, the sublime, and the disturbing is studied to see how humans can create something that stops us in our tracks. Or Political Science, which analyses how people use and abuse power.

      Now why the heck would anyone want any of those? Why would that be good for humanity? Why can't it just be engineers, as was so often suggested from about 1880 to the 1960s?

      As an Arts graduate now in IT, I do get really tired of guys who studied computers and only computers all their life and have never actually been exposed to much else, and their curled lip of contempt for anything outside their specialty.

  19. Anonymous Coward
    Windows

    Can we have a Straw man icon?

    In the article Martin Geddes is quoted referring to "Stanford educated lawyers" who don't understand packet networking but no names and no examples of their ignorance.

    I enjoyed reading the article and it made me think about the practicality of the deceptively simple concept (treat all traffic equally) but please, save the exaggerations and hyperbole for Bootnotes.

  20. Mephistro
    Facepalm

    "...rather than clamouring for "neutral" networks, which in reality don't, can't and will never exist."

    Andrew, this is the best example of the Black or White Fallacy I've seen in a long, long time.

    It's akin to saying that, as we can't completely banish crime, we shouldn't bother with cops, courts and such.

    1. teebie

      The analogy I thought of got was "it would be hard to detect evidence of this crime, so we shouldn't make it a crime"

      (Presumably any net neutrality framework would be regulatory or civil law based, so I admit this isn't the most fantastically imaginative analogy)

  21. Mage Silver badge
    Thumb Up

    Not a new problem.

    Great article

    "So If you're doing spectrum policy, it really helps to consult the physicists, so you make rules that are consistent with the constraints of Maxwell's wave equations. If you're doing broadband policy, it really helps to consult the stochasticians, so you make rules that are consistent with the constraints of packet-based statistical multiplexing," added Geddes.

    Stochastic models used to decide how many voice calls at once an exchange can support, no exchange could ever support all users at once. The "engaged" tone might mean connection capacity is used up on the local exchange.

    Most of what is advocated regarding Net Neutrality is bonkers. Especially so on a Mobile connection.

  22. Anonymous Coward
    Anonymous Coward

    Proof

    Lets try this as a means to detect non-neutral throttling:

    1) During a none peak time period download a video from a popular site that may be competing against your ISP's video service.

    2) Set up a VPN and repeat via the VPN.

    3) If the VPN download is significantly quicker then it is likely your ISP throttling certain sites.

    4) Compare notes with other users of your ISP to establish a pattern.

    I suppose someone could write a program to automate the above with a few extras and for various services.

    .

    ref: http://www.theregister.co.uk/2014/07/18/want_to_beat_verizons_slow_netflix_get_a_vpn/

    1. Neil Davies 1

      Re: Proof

      Take a look at the report: Glasnost approach would detect the presence of such differential treatment, but where on the supply chain is the issue?

      There never (rarely) is a single "domain of control" that is responsible for the delivery to you - and that's Ofcom's issue, if you can't isolate you can't enforce.

      In the US it is slightly clearer (and remember this is where the stuff started) but even there there are other issues of responsibility, in the UK our digital supply chain is much more involved.

      Disclosure: I was one of the report's authors.

      1. Anonymous Coward
        Anonymous Coward

        Re: Proof

        The technical side of this debate seems to revolve around 'accountability'. Without it no control or policing of compliance can ever be succesful... just as it actual governance cannot be succesful without effect and meaningfully appropriate sanctions. Ultimately a job for the ICC's jurisdiction.... and now we start to see why the US won't sign up to it!!

        Pardon my rusty IP packet knowledge (c.1998) but it seems to me that we will get nowhere unless IP packet architecture is split into a virtual and a 'real' (can't use 'physical' in this context) layer. A variable-length virtual packet concept (with current sequence numbering rules) mapped across multiple fixed-length packets which are numbered sequentially within the 'domain' of the virtual packet they carry.

        Armed with that we can then infinitely collect a 'hop history' list with arrival and despatch timestamps along the whole route of every packet at every device (including caches). The 'expanded hop count' history needs to list the receiving and the despatch MAC ID's of every bit of kit it hits along the way.

        We then require public access to a human readable, global registration DB of every MAC-ID, its location (down to sub-frame mounting?), its unequivocal accountable ownership and its purpose. The DB could be maintained and distributed co-resident with every DNS... perhaps as a UN ICC-owned and managed resource.. post-Security Council 'Permanent Member veto' era I would think... and you simply don't get on the Internet if your country is not signed up to all forms of international accountability. That will sort out the current business models of Google, M$, Apple.... and hobble the NSA (aka. Globalisation Protection Services Inc.)

        If you cannot establish accountability and sanction the wrong-doers, in terms of what ever rules you settle on, then we are all wasting our time and the robber-baron era of human history will continue for another 10,000 years....

    2. Anonymous Coward
      Anonymous Coward

      Re: Proof

      Lets assume that ISP's don't throttle/restrict traffic in any way other than physical limits.

      Scenario one: path from home to website

      Home -> ISP -> heavily utilised link -> ISP -> web site

      Scenario two: path from home to website via VPN

      Home -> ISP -> underutilised link -> VPN provider -> underutilised link -> ISP -> website

      Wouldn't your test identify throttling in this case even if none existed? Or does over utilisation of links/equipment count as throttling? This is a simple example - in practice there are many more links and potential points for bottlenecks.

    3. graeme leggett Silver badge

      Re: Proof

      5. Repeat with other sources of video and alternate VPN providers

      6. Repeat at different times of day over a period of days

      But most importantly

      0. Estimate what limits of detection of throttling are and what level constitutes throttling etc. I.e. construct your hypothesis (and the procedure for handling data including removing potential outliers) before you process the data.

    4. Mephistro
      Devil

      Re: Proof

      These approaches may work, but there's an easier method. Whistle-blowers. Offer a reward for employees or ex-employees informing the authorities of any illegal throttling, and whenever the occasion rises, send a group of auditors with a warrant and a SWAT team, just in case ;-). At the very least it would improve the work conditions at ISPs.

  23. Aedile

    The reason why the "violation detection tools need to infer from a tiny number of static data points rather than observe from a wide-angle vantage point" is because the ISPs purposely make it difficult to determine what they are doing. If they actually allowed researchers/experts to evaluate/observe their networks then such “inaccurate” tools and methods wouldn’t have to be employed. However, at that point all the lies they’ve been spewing will also be laid bare for all to see so it will never happen.

    1. Paul Shirley

      Which means an essential part of any legislation may be compelling them to allow better monitoring, though not something we necessarily want to happen given the risks more surveillance brings.

      Ultimately though 'today's tools cant do the job' means nothing more than it says, 'today's tools'. Arguably a signal of abuse must be detectable, otherwise no one is actually being abused. Once detected the tool of legal compulsion can be employed to dig out the truth. If the carriers are smart they'll not risk that anal probing level of investigation.

      1. Anonymous Coward
        Anonymous Coward

        It would be interesting to see the dynamics of the received packets in these situations. Fortunately now I'm retired I don't have to pull such rabbits out of hats any more .

  24. Tom 7

    Theres plenty of room for improvement with more sensible routing.

    I live out in the wilds. I'm on BT can2can for my sins. All my traffic seems to go via London. Even the shit for next door. They could nigh on double the capacity if they stopped by packets waving to each other as they go past.

  25. asdf

    MIS just as bad

    Not a lot of arts students in my computer science class back at university but the MIS track business students were required to take a few of the intro classes and they stuck out like a sore thumb. They were the frat daddy types whose projects usually didn't compile. You could tell for them those classes were the brutal weed out classes lol.

    1. Anonymous Coward
      Anonymous Coward

      Re: MIS just as bad

      Reminds me of the early days when there were no Computer Science degree courses in the UK. Companies recruited people with a degree in anything on the assumption that they were above average intelligence and had shown the ability to learn for their degree,

      Our department's job was helping the design engineers commission a new prototype mainframe. That meant we turned our hand to anything from running the operating system, writing test programs, or finding out what the hardware had done wrong when something fell over.

      They gave us a new graduate with a degree in history and philosophy. After three months they transferred him to the contracts department. He had proved unable to master the hex numbering system. He never grasped that A-F were merely symbols for decimal numbers 10 to 15. Don't ask about binary arithmetic or the mantissa and exponent of the Floating Point Unit we were debugging.

  26. Big_Boomer
    Coat

    Packet Pixie

    Am I the only one who imagined a parcel delivery guy wearing a brown tutu (guess the brand?) whilst delivering your parcels?

    Ok, it's just me,... I'll get my coat.

    1. Naselus

      Re: Packet Pixie

      tbh, from now on I'm calling all our network guys 'packet pixies'.

      1. Anonymous Coward
        Anonymous Coward

        Re: Packet Pixie

        from now on I'm calling all our network guys 'packet pixies'

        I wonder if they'd fix some of the routing issues I've been flagging for the past 9 months if I left out a bowl of milk?

  27. Jason Bloomberg Silver badge

    On the other hand...

    Where would the most ardent "everything must be equal" supporter stand when their robo-surgery surgeon was battling a choppy video feed as the network was overwhelmed by users watching dancing cats on YouTube?

    Where does the "net neutrality" argument go if people accept there are cases where it is right and proper that some traffic is given greater priority?

    It is not in my view really about "neutrality" it's about not being deliberately and unreasonably deprived of something which you could otherwise have had. Let the surgeon have as much bandwidth as they need, but don't allow my ISP to throttle rival services so I have no real choice but to accept their offerings.

    1. Naselus

      Re: On the other hand...

      "Where would the most ardent "everything must be equal" supporter stand when their robo-surgery surgeon was battling a choppy video feed as the network was overwhelmed by users watching dancing cats on YouTube?"

      And where would the most ardent 'net neutrality is for schmucks' supporter stand when their robo-surgery surgeon was battling a choppy video feed as the network was throttled by the ISP because they have absolutely no problem extorting hospitals for more cash, despite the fact people might die as a result...?

    2. Dan 55 Silver badge

      Re: On the other hand...

      Net neutrality and QoS are two entirely different things, you can have net neutrality with or without QoS and vice versa. It is important not to confuse the two. Net neutrality (or lack of it) is billing concept, QoS is a technical concept.

      The best solution for the surgeon is a network with QoS so video has priority and with net neutrality so their connection isn't throttled for not paying the Mafia fire insurance.

  28. nijam Silver badge

    Another Orlowski diatribe against - well, this time it's not clear, probably Google because it usually is: "Google support network neutrality, so that must be a *bad* thing". Who knows. I really must stop reading anything with his name on, his wonky perception of the world is becoming tiresome.

    And for all the people supporting him in the comments, wake up, we aren't talking about packet pixies, we're talking about throttling trolls.

  29. strum

    BS

    > The whole 'neutrality' debate has been about the ‘fair' outcomes

    No it isn't. That's the lie at the centre of Orlowski's rant.

    It's about intent, not outcomes.

  30. BitDr

    Hmmmm...

    >"Currently, ISPs manage the floor through a mix of pricing, scheduling, and not taking on too many customers."

    If paying more gets you better performance then there are multiple floors and the ability to screw some while benefiting others is true. Not taking on too many customers is the right way to do it, but greed will win over common sense, and the pipes will be oversold, leaving many wanting.

    > we made commercial steam engines without fully understanding the properties of gases, and experimented with electricity long before Maxwell developed his theories of electro magnetism.

    The lack of understanding of gasses did not impede our ability to use the steam engines, and Maxwell would not have been able to formulate his theories unless experimentation was performed. Packet-switched networks are a creation of humanity, not some natural force that has been harnessed by us. As such they are theory in practice. I would posit that the understanding of their operation would need to be well in-hand from the start in order for them to exist at all.

    When some are favoured and a great many are disfavoured the unfairness is prevalent and most likely intentional. To state that it is a result of emergent behaviour is all well and good, but it smells like a cop-out, something in the relationships is causing the bias; figure out what it is and squash it, unless of course it is deliberate, and they you have to invent another reason for it being the way it is.

    I think network neutrality mean doing nothing special with any of the data passing through the network, (i.e. just keep it moving). Trying to give some data priority over other data is when (IMHO) "emergent behaviour" begins.

    1. Anonymous Coward
      Anonymous Coward

      Re: Hmmmm...

      " I would posit that the understanding of their operation would need to be well in-hand from the start in order for them to exist at all."

      Much of the current networking theory - like all new technology - is based on trying to explain afterwards why things didn't happen as expected. There are several RFCs that show the progression of work-rounds that were devised when each new unexpected operational problem was discovered. Van Jacobsen's and Phil Karn's TCP algorithms are good examples that now seem obvious in hindsight.

      Modelling tended to be done after the event. I remember one sage had a saying - "if you need Queueing Theory calculations - then the system is under-sized".

      Modelling after the event did confirm surprising answers. IIRC telephone networks smoothed random bursts of traffic - but bursty data packet behaviour just gets amplified.

  31. Cynic_999

    It's the reason that's important

    It's not the fact that some packets are given greater priority than others that is of concern, it is the *reason* for the difference that needs to be examined in order to decide whether the reason is acceptable. Slowing down a download from NetFlix so that a robo-surgeon has all the bandwidth he needs may be considered acceptable, whilst slowing down a NetFlix download so that users get frustrated and switch to a different movie service may not be considered acceptable, especially if it involves deliberate throttling rather than packet prioritisation.

    1. Anonymous Coward
      Joke

      Re: It's the reason that's important

      A robo-surgeon linked to the unclean interwebs!!!!

      I shudder to think what the result would be:

      "Anonymous" tattooed to your intestines, twitter "Uh-oh! Lol!! #RoboS" updates, Google ads post op recommending second-hand hand markets, etc.

    2. Destroy All Monsters Silver badge
      Holmes

      Re: It's the reason that's important

      It's the reason that's important

      "Not basing traffic management on packet attributes that are irrelevant for said management."

      On a highway toll booth, cars with the "already paid elsewhere" sticker, the VIP marker or the blue light may pass quickly. The rest should be an amalgamated mass. Once vans of delivery company X are suspiciously often held up compared to vans of delivery company Z, something is going on. But then the contract should stipulate a max delay time for 99% of the time or penalties.

      If there are daily jams, you may want to consider using our VIP#2 lane, usable for a low, low fee... (and then you have to hope that the next tool booth is either owned by the same operator or is contractually bound to you, too)

      In the software world, there is talk about moving away from the "Not responsible, LOL" EULA model for paid-for software. Is it time to move away from the "best effort, LOL" EULA model for packet delivery?

  32. Anonymous Coward
    Anonymous Coward

    Parcel pixies

    doesnt help when companies using advertising like Shaw in Canada showing little anthropomorphised packets wandering around big tubes for their internet service adverts.

  33. noominy.noom

    Orlowski as usual did a great troll

    I typically comment on net neutrality BS, even when I know I am being trolled. But there have been some good posts here already and I don't want to duplicate the ideas. See above:

    Aedile

    Joe Gurman

    Confused Vorlon

    Zippy's Sausage Factory

    auburnman

    Naselus (Well said, the robo-surgery rebuttal)

    Cynic_999

    and sometimes: TheOtherHobbes (not meaning some good/bad, but meaning some on topic some not. I'm just referring to those on topic.)

  34. relmasian

    The first thing I did after readings Orlowski's rather twisted article was to check to see if he was shill for cable networks; he is not. It seems he is a longtime Register columnist who has a history of trying to be an iconoclast. This is not the first time he has written something "outrageous."

    Orlowski's arguments about packets and ISP intents is cuckoo. One does not have not have to know much about network management and packet traffic to know whether an ISP is violating new neutrality. One just needs to find contracts/offers that promise special treatment after paying special prices, especially if the contracting company chews bandwidth when delivering product to an ISPs retail customers. Hulu and Netflix are examples of such companies. One could also check whether an ISP actually delivered special treatment to some companies. However, ISPs do not have to actually deliver special treatment in order to violate net neutrality. They just have to make and to collect on the offer of special treatment.

    1. ipghod

      yes, exactly! the ongoing problem is attempting to run this backwards: presuming there is intent and that common network resources have been 'sold twice' when performance is bad.

      if, for example, carrier x has poor performance to customer Y -because- Ys traffic demands back to carrier X's customers overwhelmed a common public connection, the lack of performance is not 'proof' of an evil intent, nor is proposing customer Y buy a larger, private connection to carrier X as a solution to the problem.

      it also begs the question if carrier X can hold customer Y responsible for spamming it's public interface with traffic that starves out other legitimate traffic (an unintentional denial of service), and would therefore be justified in taking active measures to protect their network (rate limiting).

      the point of the article is well taken, but should not be turned into the old 'engineering vs philosophy' argument again. the philosophy of the open network, we all get. just because some ethusiasticly vigilant philosophers have misunderstood technical jargon, and attempted to hijack engineering's means to deliver their end, doesn't make their points irrelevant, it's just their attempt to force a method based on a metaphor that needs to be stopped cold.

      when you shout 'technical foul', and then spout metaphors that fall apart with the most casual of technical inspection for justification, it tends to reinforce the idea that you should keep your hands to yourself.

      your wishes have been made clear, and will be respected. now keep your hands out of the config files, please.

  35. Diarmuid Pigott

    Packet is a metaphor

    Spot the computer science majors with a poor understanding of logic as well as argumentation.

    All reasoning is in part metaphoric, because we need common terms for debate. So picking holes in other people's metaphors is fine, but objecting them to using them in the first place is daft, and show a lack of understanding of argumentation in the first place. (And as for an appeal to Boolean logic? in 2015?)

    Packet, Exchange, Medium are all naturalised metaphors - they're metaphors that were so useful they became terms in the jargon set. But if the grounds of the metaphor (the reason the model was considered applicable in the first place) fails, then the metaphor does.

    As Stamper pointed out in the 70s, all accounts of information are ridden with metaphors and ultimately incoherent. (Stamper, R.K. (1985) Towards a Theory of Information: Information: Mystical Fluid or a Subject for Scientific Enquiry? The Computer Journal 28, 195.)

    Carnot explained the heat cycle in terms of caloric, a metaphoric-derived term for the transmission of heat. Likewise Elan Vital (Bergson) and the Ether (everyone). A better understanding of the nature of heat (Clausewitz) gave us entropy as a non-metaphoric term. This may yet happen with information.

    1. Destroy All Monsters Silver badge
  36. Anonymous Coward
    Anonymous Coward

    Packet discrimination based on the packet type or protocol (ie SIP/streaming video> http > ftp/smtp etc.) is fine and necessary, and while, strictly speaking, not neutral, it's also not really what net neutrality is about at it's core.

    Net neutrality is mainly about source/destination neutrality, ie treating all traffic by the same rules, regardless of where it comes from or is going.

    Analogies only serve to muddy the waters because they're imperfect in themselves. Instead of dumbing the subject down with analogies, we should be teaching the really rather simple broad concepts to the people involved.

  37. Anonymous Coward
    Anonymous Coward

    It's about censorship not being neutral.

    Having paid good money to an ISP that restricted the sites and addresses I could access, and even called me on the phone (not email) to say they did not like the sites I was visiting (including any .ru sites) or the software I was using to do so I know first hand what "net neutrality" can look like.

    When your most important connection to information is being controlled, for profit or information content or both, you end up being controlled. Even if you know the information is being selected for you "your" ideas and thoughts on topics will fall into line. Sure there are work arounds but they take effort and on topics of little concern you, me, we do not make that effort, we just accept the information that is being fed to us.

    Maybe it would be better to just call it what it is, censorship.

    If you make it harder, or require people to pay, to find uncensored or unthrottled content, they won't, they will

    1. Destroy All Monsters Silver badge

      Re: It's about censorship not being neutral.

      Censorship is never neutral dog. That's, like, a totally different mass of state-issued shit.

  38. laurence brothers

    Is it any better for a liberal arts grad to comment on a tech issue than for an engineering grad to comment on a business issue?

  39. Horsie

    Net neutrality? Doesn't really exist...

    Traffic management is needed. That is a given until we find a way to simultaneously transmit infinite amounts of data at infinite speeds between all the different points in the world that should need them and for free.

    As all those conditions are unlikely to be happening anytime soon, this means that there is a cost involved in putting in network infrastructure, and although the more bandwidth you put in makes things cheaper on a per-bit basis, there are limits to how much infrastructure companies are willing to put in.

    Let's not forget that transporting data is a business. Service providers want to recover their investments in a timely manner and see a worthwhile return on that investment (ie that the cash put in makes a profit when factoring risks, amortization, maintenance, etc....).

    So... limited infrastructure based on cost and/or business model, which will be (presumably) less than the potential demand for bandwidth.

    If service provider "A" has a flat fee schedule where all of its customers pay the same amount and bandwidth gets shared around on some sort of equal footing (erm.... should we put in traffic management to make sure that the customers get the same amount, or should the customer that is hammering the network more get more? hmmm.... already the notion of traffic management cropping up.....), and along comes someone (a rich individual or a company) who says "look, that amount of bandwidth at that service level is not good enough for me, can you give me more if I pay more?", I really don't see the issue about it.

    Also, either due to a business opportunity, or due to excess stress on the network, a provider may choose to limit a certain type of traffic (either throttle it or block it) unless there is a payment involved.

    Note that we are no longer in a telco monopoly situation (in most developed countries at least... the issue of monopolistic situations is a completely different one), and thus if a certain service provider is doing things that some people consider unacceptable ("OMG, I can't get to site www.joeblogs unless I pay them $$$$!!!" or "Crap, as www.joeblogs doesn't pay the provider, I can't see them"), then move over to another provider that acts in a more acceptable way.

    The argument of "I want unlimited access to everywhere for free (or at a price that *I* decide) is just not going to happen.

    Alternatively, if you (and enough like you) are convinced that there is a valid business model in giving out unlimited access at unlimited speeds to everyone (and stating in a contract that you ARE giving unlimited speed/volume), then write up the business case and seek funding for your wonderful new service provider... I'm sure you will make loads of cash...

    John.

  40. Daz555

    Traffic management fine.

    Net neutrality - traffic management without corporate $ bias. Also fine.

  41. PeteCress

    To me, it is all about who pays and who knows who is paying.

    If the end-user has to pay extra for, for instance, video downloads that's all well and good....'You pays your money and you takes your choice."

    But if the end user, when browsing web sites gets terrible response for site A and great response for site B - and has no idea that site B is paying somebody off under the table for that preferential response time, that is a huge problem to me - and everybody else.

    Free economies are founded in part on a knowledgeable consumers - and hiding things like who is paying for faster speeds defeats a free economy.

    Imagine Mom-and-Pop's site offering productr"XYZ" for $25.00 while Big-Corp's site offers the same product for $30.00...... but Mom-and-Pop can't sell anything because the response time of their site has been throttled - while Big-Corp's site has not.

    Consumer loses, economy loses, Big-Corp wins.... one more step on the path to oligarchy.

  42. Anonymous Coward
    Alert

    Lessons from telco?

    It pains me to admit it, but there may be lessons from the telco regulatory regime that could be applied.

    1. Local loop maintained by a neutral third party. Where "local loop" means the fiber/DSL/whatever link between you and the edge of the Internet (operator's network).

    2. Non-discriminatory access from the provider edge to an ISP. This is analogous to the local PSTN loop terminating in a central office but being wired to equipment installed in the CO by a competitive carrier (CLEC). US law in particular is very rigorous here and infractions by the CO owner are extremely rare. Net effect: you can connect to any ISP that has gear in the data center you connect to.

    3. There is a universal floor for standards. In telephony this is enforced by rules around uptime, latency, call quality, call connection completion rates and suchlike. In broadband you would have rules like latency/jitter, packet throughput and suchlike. An ISP would have to deliver at least this standard.

    4. Customers sign up for "at least" service. You buy 10Mbit fiber to the home, you get at least X, Y and Z. Beyond that, the carrier makes no promises and can manage traffic how they see fit. People like Skype and the next Whatsapp tell their users that they need certain values for X, Y and Z and that's that. It's actually in their interest to write tighter code now. In telco, the FCC mandate certain minimum standards, which is why wacko stuff like 1960s fax machines work alongside the medical bracelet you bought Grandma last Christmas.

    If I'm a small startup, the prevalence of "at least" connections means I am guaranteed to have my service used at acceptable speeds. If I am Google, I don;t lose - I can still peer with carriers. If I am Netflix, I can continue to cut deals with ISPs to put my CDN servers close to users. If I am an ISP, I can traffic manage to my heart's content so long as I maintain the floor standards. Everyone's happy.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like