Wednesday, June 14, 2006

Tiger Repellers, or why the Net should NOT be 'neutral'

This man is standing at the crossection and waving his arms. The passer-by asks him “what are you doing?” The man replies “I am repelling tigers”. The passer-by said: “But there are no tigers here!”.

“See how good job I am doing?”, the man replied.

'net neutrality' = tiger repellent in a can. Reasons:

  1. So far the internet has worked well without any “net neutrality” legislation even though QoS and bandwidth shaping technologies have long been available. There's no reason now to add legislation to the mix. It's trying to cure the disease nobody has seen yet.

  1. Connectivity is the very thing telcos sell. It is simply bad business to piss your customers off, even if you have local monopoly – pissed customers are potential, the virtual opportunity for the competitors to enter this area. Competition on the market is tough. Degrading connectivity puts telcos in the situation of candy store poisoning the candy they sell. It would be suicidal for them to do it, both in commercial and especially political sense. Arguments that it might be difficult for customers to go to another candy store systematically fall flat on face in practice (one good reason is that it is the candy store that has to make itself conveniently located and appealing to the customers, and the business owners typically are rather good at exploiting business opportunities). Massive service degradation to direct the customers to offers/sites where provider wants creates potential for competitors to attract over the customers; on the other hand, non-massive and selective degradation of service does not create enough of revenue stream to justify frustration and damage to relations with customers. As a business strategy, this is a non-starter for anybody but a proverbial pointy-hairy boss.

  2. The whole hoola is created due to hypocrisy of the bandwidth hogs: they want their privilege and subsidy to be guaranteed forever by the law. Those are mainly big bandwidth users Google and P2P users hogging links with downloading porn and illegal copies of movies that like the concept (P2P users and bandwidth are like Bricktop's pigs and meat). Imagine Google and P2P-ers as transportation companies, huge corporations and small businesses running a few trucks. Would they like to pay in proportion to damage and wear & tear they cause on highways? Or would they prefer everyone to pay taxes that subsidize infrastructure they benefit from to much greater extent than regular drivers? Their subsidies are protected by the technical difficulty of charging highway users individually for their own wear and tear and not that caused by somebody else. 'Net neutrality' is an attempt to achieve the same by introducing legislation that effectively bans bandwidth metering and traffic shaping like quality-of-service (QoS). It is wrapping the most selfish of interests in noble-sounding slogans like 'net neutrality' or 'democratic nature of internet', but it all boils down to plain, seflish, unenlightened interest of hogs using 80% of bandwidth and paying the same price as everybody else using the 20% of overall capacity.

  3. Google wants subsidies and overwhelmingly supports Democrats. The scam works like this: Google wins guaranteed subsidy from all the users of internet done in way they don't have much need and specialist knowledge to understand (how could they detect that instead of paying $20 per month for internet they could be paying, say, $12 – the opportunity destroyed by 'net neutrality' legislation will simply never show up on their bills, it is prevented from coming into existence by legislation putting straightjacket of simplistic and special-interest-ridden lawmakers' ideas onto the network). Democrats get yet another political credit for nothing and more tax-funded agencies doing fictional quasi-work whose main real purpose is providing tax-funded “jobs” for leftist politicians and other thieves feeding at the public trough. P2P users get bandwith. Who pays? Telcos, theoretically – but they would find the way to shift the costs onto the end users. They'd have to – either that or those who don't would simply go out of business. The supply of bandwidth simply has to fall in such case, and thus the prices have nowhere to go but up. Since nowadays there is no option for millions of end-users not to buy connectivity, the remaining telcos would be able to charge end users more for it; they would have to charge users more in order to avoid red ink on the bottom line; and since they are profit-seekers, they would obviously like to do it.

  1. Telcos probably oppose 'net neutrality' to be able to expand the offer: if they can't use QoS, and have to use the same price, there is no way for them to differentiate products they sell and introduce the new, high-premium ones. They probably don't make much money on plain IP service. Rule of a thumb: 20% of customers in most businesses brings in 80% of revenues and vice versa. In order to develop and expand their networks, they need products that would appeal to customers with deeper pockets - and 'net neutrality' means they won't have the right to do it. Little wonder they fight 'net neutrality' tooth and nail. Not having that opportunity means less innovation. Again, it's not surprising that network hardware vendors oppose 'net neutrality' as well: no innovative services means no sales of their hardware. So, no low latency for VoIP packets, and no QoS for digital TV, no premiums for 'blue water strategy' products, no bulk discounts for traffic of peculiar characteristics some business might find use for (say, low-prority, low-reliability massive transfers of CAT scans, media libraries to archives and other data we did not even invent yet - if something gets lost, just retransmit it). Is this really what we want? It should be clear that 'net neutrality' is on the whole bad for end users other than bandwidth hogs, for the economy and last but not least, for the rate of innovation as well. Drab, standardized 'public utilities' for everyone appeal to egalitarian maniacs of 'equitability', but the experience has demonstrated what kind of products this sort of policies tends to create. Equality before quality might sound nice in theory, but trust me, the experience sucks.

  2. All the inane schemes that 'net neutrality' proponents try to scare casual internet users with are based on idiotic conspiracy theories and layman misunderstandings of economics: there is no business for telcos to cut off or degrade the traffic to whatever site, leftist or non-leftist, Amazon or its smaller competitor. From the viewpoint of pure business, they have no reason to care whether traffic they carry is Shakespear plays, leftist ramblings or child porn. Any traffic – that they can charge for! – is good for them, since it means more business. If they cared for what the traffic contains, they would be politicians, not telco businessmen. 'net neutrality' means overusing existing capacity + standard, low prices for using that capacity. This is asking for overgrazing problem, or 'tragedy of the commons'. This is obviously bad business for telcos. It would force 'public utility' model on them, which they obviously don't want – while bandwidth hogs like it, it also means overused, under-invested infrastructure. Yes, telcos oppose 'net neutrality' for selfish reasons, but the strongest proponents of 'net neutrality' are no different in terms of having the vested interest of everyone subsidizing bandwidth for them. In this case, the interests of majority of end users of internet are more in line with interests of telcos than with the interests of maniacal porn downloaders and a few internet companies that grew big only thanks to using capacity they underpay. Neither degrading service depending on how much particular sites could pay to telco is a workable business model - this is yet another reinvention of myth of predatory pricing.

  3. How can we know what will happen in the complex real world & net if 'net neutrality' is enforced too soon? If there is indeed good evidence for massive rigging of Web by telcos, one can always pass the legislation if systematic evidence is actually found – the political force of the evidence and pissed off users in democracy is a force to be reckoned with. It is certainly easier to pass such legislation while having the significant body of evidence available than without it! The politicians' cavalry could ride in and save the day, the saved people would cry “oh thank you, thank you for saving us from the savagery of free market”, Democrats would gloat and the televised orgy of good intentions and huge self-aggrandizement for little work (and not little tax money landing legally in the pockets of messiahs) would follow as usual. I would recommend this scenario to the political left on grounds of greater satisfaction if only.. it were all real. The catch is that the problem most probably isn't real – the 'net neutrality' proponents cry so much and get their knickers in a twist exactly because they know that they won't have enough evidence to show. So they have to pass it before empirical test is ran: no problem to be solved, no taxes collected for the new agency they love, no big political discussions and panels in which they can come off as saviors doing Good Things For The People. Engineers say: one test is worth a thousand opinions. The 'net neutrality' proponents seem to be rather keen not to run the test. Suppose the legislation passes - they will be able to argue forever it is thanks to them that the internet keeps working, even if it were proven by 10,000 intellectual equivalents of Andrew Odlyzko to be untrue, groundless and harmful. Passing the legislation is relatively easy - in comparison to deleting it. All the internet users could be paying through noses and with frustration as well for decades for this crock of an idea.

  4. The Visible Hand of Government is typically worse than whatever imperfections of free market. A distinguished researcher in economics, prof. Mark Blaug, in his Economic Theory in Retrospect, had this to say: “The proponents of government regulation have to show that their policies have less in way of failures than market has of its failures”. Getting politics to tweak in a market is a very serious matter and a slippery slope: unlike naturally decentralized groups of customers, the political process is always centralized and thus always vulnerable to distortion. It is routinely hijacked by special interests, tweaked with by the ignorant dogooders in ways they do not fully understand and with side effects they are way too dim to detect and comprehend, full of miscommunications and errors in thinking, always prone to the law of unintended consequences, and always limits the range of experimental products and policies that businesses can try. If you don't have to really and honestly fix what isn't broken – leave it alone. First, do no harm. As empirical research by Ronald Coase has demonstrated, government politicies in complicated business realities work very poorly. Introducing regulation without good and clear reason is typically result of some sort of complicated vested interests at work, the interests of which the public opinion is mostly unaware of as the interests in question typically deal with complicated, technical realities largely concealed by trade secrets that even the most experienced experts in this area do not know in full, while it is typically portrayed by the proponents as "doing things for The People". From my point of view, 'net neutrality' smacks exactly of such cynical lobbying amalgamated with manipulation by very peculiar sorts of political preferences (like, knack for 'public utility' model of some political orientations, even over public interest or preferences of individual customers).

  5. Finally, net neutrality can cut both ways. The standard rationale for 'net neutrality' is to allow unhindered access to less popular or profitable sites. But imagine that by law the providers can not discriminate against... the video traffic. No, not against the traffic of small, idealistic organizations having their web servers in the discriminated corner of internet that so many of the neutered shed (crocodile) tears over. The popular, 50MB videos of whatever clip or parody. Or pirated DVDs of 4.7GB weight. Thus, the teenagers watching the latest idiotic video on or YouTube or MySpace are going to watch their video all right - but the small website will not be able to get its traffic quickly through the congestion caused by that video. Net neutrality - or not having a specific priority configuration, the dreaded 'special lane' for video could easily come at expense of content that is important, but relatively low-bandwidth, such as pure text of HTML code. It could easily create the very effect its proponents want to protect users and small content providers from: after all, all bits are created equal, right? The 'lane' for small websites could very well be smaller than the 'lane' for the video traffic - but if video cannot fill most of the spectrum available at the moment and thus it is forbidden from clogging the 'Web lane', this 'lane' might very well be faster! It's not like the technology isn't available and wasn't used before for good reasons even by small ISPs and network managers in organizations without very deep pockets, like guaranteeing minimum bandwidth for particular kinds of traffic so that each gets at least some of the spectrum (note for techies: I'm obviously simplyfying things here a bit, I do not want to fill this blog with technical jargon that does not change the essential problem). Letting paranoia go and producing cheap "profitability schemes" by means of abusing the network is easy and emotionally gratifying, but it conceals much more practical and plain reality: most network managers and companies use bandwidth management primarily to provide as comfortable access to the network resources as possible under particular constraints of finance, organization, and people skills (helpdesk, anyone?), because it simply doesn't take much to get the users irate. Continuing analogy from above, if freight companies and high-mileage truck drivers do not pay for infrastructure in proportion they use it, they have the incentive to clog the most popular roads with their eighteen-wheelers, creating traffic jams for everyone else: you lose more than they gain. More bizarre consequences would follow if one wanted to follow the law in consequential and systematic manner: the Internet Cache Protocol would have to be, well, banned. Right now most surfers go through system of transparent Web 'caches' without even knowing that: simplifying things a bit, the network managers often use software such as Squid to cache the requests for pages from popular websites, thus speeding up access to users of particular ISP/telco and economizing use of the wide area links (the requests for many pages go only to the company's Web cache and don't need to go all the way through the internet to the source webserver, thus saving the internet bandwidth - only the first user to get the page needs to get to the original webserver, the following requests are served from the cache). It's not hard to guess what websites get much of their content cached and which don't: strictly speaking, they get preferential treatment by the sheer virtue of their popularity, don't they? And in principle, the law cannot be applied arbitrarily: telcos QoS bad, Web caches good? So either the legislation has to treat some issues simplistically and force it, thus harming the net; or it has to be extraordinarily complex, extensive and thus costly, which means forcing adherence to it will mean increased costs and reduced opportunities for everyone. And that's assuming that the govt officials, companies and end users don't ever try to abuse the net or lobby legislation with their special, vested interests. No sir. We're all good citizens, right? However counter-intuitive it might sound, the requirements of comfortable and quick access to every kind of content might require that the net were exactly not neutral in its treatment of traffic!

In conclusion, 'net neutrality' is inventing imaginary hobgoblin for people to be supposedly 'rescued' by the hodge-podge coalition of hogs, special interests, and leftists politicians always craving to 'rescue' the people from the threat the people do not need rescuing from for complex reasons:

The whole aim of practical politics is to keep the populace alarmed and hence, clamorous to be led to safety - by menacing it with an endless series of hobgoblin, all of them imaginary.

(H. L. Mencken)

Two very insightful articles on the pitfalls of legislative way towards 'net neutrality':

Their conclusion: “Broadband ISPs aren’t likely to succeed in shutting down VoIP, as most do face at least one competitor, and it would be quite easy for VoIP software to evade attempts by ISPs to squash them. What is a threat to the openness of the Internet is FCC regulation in the name of “net neutrality.” Putting the FCC in charge of telling ISPs how they may or may not run their networks is the first step toward politicizing network policies. Although initially the FCC might use that power to do some worthwhile things, in the long run, the FCC is likely to be captured by special interests and do things the Arsians wouldn’t like in the least.”

Update: real world experience shows that so far trying to exclude certain sites from the internet has been practiced not by telcos, but, unexplicably, strangely and amazingly (not really..), by the main proponent of 'net neutrality' itself, Google:

Google Purges The People's Cube Worldwide

Thought Police on the Internet?

More of good comment:

Greg Mankiw's Blog: Network Neutrality