Home / Blogs

Would the Real Network Neutrality Please Stand Up?

I’m sure this is something that’s been raked over before, but I don’t see a common understanding of what ‘Net Neutrality’ actually is. Despite many of the Internetorati demanding it by law. Whatever.

There appear to be several different camps, which you could paint as “bottom of IP”, “middle” and “top”.

The bottomistas would see enforced Internet Protocol itself as a premature optimisation and violation of the end-to-end principle. Unhappy that you only get IPv4 or IPv6? Still grumpy that you only have IPv4 and not even IPv6? Really miserable that your VoIP packets are staggering under the poisonous load of IPv6 headers? You’re a bottomista.

I suspect there are some fundamentalist bottomistas who would object to your service providers not giving you a choice of Ethernet, ATM or roll-you-own-L2-protocol. We’ll pretend to be out and not answer the door when they knock.

The middlemen draw a distinction between “raw IP” (before the ISP gets a hold of it), and “retail IP”, which is what you and I get to experience. This kind of suggests that the OSI 7-layer model got it horribly wrong, because there’s a fundamental cleave right in the middle of layer 3, where IP sits. Fair comment, but sounds pretty radical to me. Although I’ve never really got layer 6, so maybe they’re onto something.

Then you might be a “top of IP” kind of girl. You can cope with the discrimination creeping higher up the stack to the next layer, where particular TCP and UDP ports and flags are screened off. But you only get queasy if particular commercial service providers or applications are targeted. Blocking off port 25 is OK to you, since it doesn’t discriminate against any particular email service provider.

Sadly, these are all hogwash and bunkum.

Net Neutrality is a dead end, because as Searls and Weinberger correctly noted, the Net isn’t a thing, it’s an interconnected set of agreements. These are bilateral and freely entered into. And since those agreements weren’t modeled off a viral template such as the GNU General Public License, they are all unique. There’s no contagious clause that insists the Internet becomes a “thing” by virtue of everyone having to agree to freely and neutrally pass packets in an ever growing pool of Neutraldom. So to impose neutrality you’re going to have to interpose yourself into a lot of contracts. (Another reason why “Internet Governance” is an oxymoron when referring to anything beyond IP address allocation and routing, which do require some central agreement and co-ordination.)

There’s no grand “first principle” from which you can derive network neutrality as an economic argument. No public choice, competition, game theory or otherwise construct that leads us there. Indeed, saying that the public would benefit if there was a transfer of wealth from providers to users isn’t good enough. You’re playing with matches in the oil refinery when you start messing with property rights. Yes, those networks are mostly funded by risk capital. The local loop copper of a fixed operator may still be hangovers from monopoly days, but generally those assets were brought into the private sector on clear rules, the stockholders took a punt, and some of the better informed ones who saw the long-term potential of DSL etc. got to reap a windfall. Of course in parallel the telcos have done a superlative job of lobbying for rules that keep competition out, but that’s a different issue.

But wait a moment, it gets worse.

What if I wanted to allow people in the street to access my WiFi? But I only want to offer web and email, so as to make P2P file sharing tricky. As a good public-spirited citizen I put up a splash page so they know exactly what’s going on. Am I allowed to? Or is Net Neutrality only for the mythical mystical “them”?

When in deploying my network do I need to “design-in” neutrality? Concept, build or operation? Should we be outlawing the deployment of PSTN-specific GSM networks because they’re “unfair” to non-PSTN voice applications like Skype? Am I allowed to deploy non-technological measures for neutrality, such as contract terms? Am I allowed to read the packets, but not block them, in order to enforce my contract (repeat - freely entered into by both partners)?

What level of jitter and congestion is perceived as “neutral”? What if I deploy technology like Qualcomm’s 1xRTT, which separately supports voice and data, with PSTN-only voice, but the data is a bit lousy for VoIP? Is that being unfair, or merely a realistic response to the limitations of technology?

Is neutrality a wholesale or a retail problem? What if the access infrastructure owner offers “neutral” IP connectivity, but no retail provider chooses to pass that on directly to the public without layering on some filtering and price discrimination?

Oh, and what’s so special about the Internet? Do other IP-based networks need neutrality principles? Do any networks? Should more network industries be forced to forego “winner takes all” rewards? Google looks awfully dominant at adverts, doesn’t it… I wonder if that ad network needs a bit of “neutrality”?

Incidentally, although I’m against blanket rules enforcing neutrality, although I would reserve it as a tool for post hoc competition and antitrust law enforcement. And I think you can make a stand on Network Neutrality on political and free speech grounds, but that requires a very different policy approach (i.e. not one that confiscates the proceeds of private capital investment).

And if the users value a neutral connection so much, perhaps it’s time for them to self-organize a bit, build their own networks, or tender for connectivity together—rather than rolling over and accepting whatever the local telco can cableco provide by default. But that would burst the illusion that government is here to save us from ourselves and we’ve no need to take personal responsibility for our connectivity freedom.

The moment you try to define Network Neutrality, you have to choose a layer, a time, a market, the participants. You have to make non-neutral choices in order to define the boundary of your Neutralsphere. There is no ‘neural’ space devoid of favouring the interests of particular market players. The contradiction is inherent. There is no way to finesse it away.

Everything’s bass-ackwards. Neutrality is a sign of healthy supply competition and sophisticated ways of demand expression. It’s an output, not an input. Beware demanding net neutrality as a blanket principle, rather than a scalpel to excise particular local anti-competitive acts. Kruschev declared the corn harvest was great, too—but it didn’t create the incentives for more corn to be sown and for the system to succeed on future iterations. And net neutrality rules are likely to have the exact opposite effect of that intended too.

Net neutrality messes up freedom of contract, freedom of association, and property rights.

I don’t buy it.

By Martin Geddes, Founder, Martin Geddes Consulting Ltd

He provides consulting, training and innovation services to telcos, equipment vendors, cloud services providers and industry bodies. For the latest fresh thinking on telecommunications, sign up for the free Geddes newsletter.

Visit Page

Filed Under


The Famous Brett Watson  –  Nov 16, 2005 2:27 AM

There are technical issues here, and there are political issues.

The OSI seven-layer model of networking isn’t God’s Own Truth; rather it’s a useful pedagogic tool. It illustrates the usefulness of dividing the networking problem into subtasks which interact in a layered manner. It identifies lines of demarcation between the layers which are practical and not to be lightly ignored. Layering is an important concept, not only in networking, but in computing generally—and in many other kinds of analysis as well. Certain kinds of structure, when used appropriately, are a definite aid to clear thinking.

We could do with some of that clear thinking here—not necessarily structuring in layers, but at least separating orthogonal issues.

The general principle of “neutrality”—from a technical perspective—means respecting the network layers and not cutting across them without good cause. The layering used in the TCP/IP stack means that only the endpoints need to implement TCP: intermediate routers can operate purely at the IP level. Neutrality is maximised when the intermediate network devices operate at the lowest level possible to do their job.

In general, neutrality is a good thing because it respects layering, and layering exists for good reason. But we need not be slaves to the layers. Arguably the layering in TCP/IP isn’t perfect, and there may be good reason to consider data at higher layers than is strictly necessary for basic operation. Firewalls are an obvious example of such intervention, and there are some mixed views about the benefits and ills of firewalls as a result of this, but they’re a widely used tool and not going away any time soon.

From a technical perspective, I conclude that “neutrality” equates to respecting the layers. For an Internet Service Provider, the most “neutral” service is one that operates at the Internet Protocol layer (the lowest Internet-specific layer) and does not discriminate on the basis of payload, but only on the IP layer header data.

Whether or not such extremely neutral service is the best kind of service is a different question. If fact, it’s a different kind of question, because it depends on the wants and needs of the parties involved. We’ve entered the social realm: an orthogonal issue.


The Famous Brett Watson  –  Nov 16, 2005 2:30 AM

[continued from previous post]

There are circumstances in which I want my ISP to become rather “non-neutral” on my behalf. If some host elsewhere on the Internet has gone haywire and is flooding me, I’d like my ISP to firewall it off as far upstream as possible so that my physical link isn’t saturated with unwanted data. Where we can agree to move away from neutrality, or where I at least acquiesce to such a move (say I’m sympathetic to their blocking of TCP port 25 as an abuse control mechanism), there is no problem.

A fresh problem arises where I’m not happy with my provider’s non-neutrality for one reason or another; where they are discriminating on the basis of packet content in a way that is contrary to my wants and needs. Maybe they are being entirely self-serving about it—attempting to sabotage Internet services which compete with their other products, or similar. Maybe they’ve interposed an irritating web content filter and made it hard to circumvent. Whatever the case, we now have a dispute. Disputes result in politics, which typically means the application of law.

So now we have yet another orthogonal issue: the political angle. To what extent should behaviour be regulated? At this point the philosophical implications overwhelm me, and my analysis fails. I can’t give a good rational argument for how people ought to be governed which isn’t based on some other set of values equally in need of support.

On the other hand, I’d like to point out that other arguments—such as Martin’s—fall afoul of the same difficulty. I agree with Martin that “neutrality is a sign of healthy supply competition”, although I would say that the important factor is not “neutrality” so much as the absence of disputes. Where disputes can be resolved by finding a different source of supply, there seems little need for political intervention: it’s the free market working the magic it’s supposed to do.

But where disputes exist and won’t go away of their own accord, an argument exists for intervention. The question of what form this intervention should take—if any at all—is a political quagmire, and I won’t go there. Contrary to Martin, however, I believe that some kind of legislative requirement for neutrality is a reasonable approach towards bringing disputes under control. Not necessarily the best approach, and not necessarily free of unintended consequences, but worth considering.

In particular, I disagree with Martin’s concluding remarks: that “net neutrality messes up freedom of contract, freedom of association, and property rights.” If genuine freedom of contract and association exist, then disputes resolve through those avenues. Conversely, where disputes persist, it is generally because there is no genuine freedom of contract and association. If such freedoms don’t exist, then it’s just blowing smoke to protest that they might be damaged.

And as regards “property rights”, there is a question of using one’s property as a means of leverage in areas outside that property. If it turns out that you can use your property to put the squeeze on other people, one way or another, and you contend that it is your “property right” to act in this manner, then there is a direct tension between “property rights” and “freedom of contract and association”. You can’t be naively in favour of both things where such tension exists. Personally I think the situation is resolved by denying that your rights of property ever had the extent you imagined, but there’s room for argument.

Comment Title:

  Notify me of follow-up comments

We encourage you to post comments and engage in discussions that advance this post through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can report it using the link at the end of each comment. Views expressed in the comments do not represent those of CircleID. For more information on our comment policy, see Codes of Conduct.

CircleID Newsletter The Weekly Wrap

More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Co-designer of the TCP/IP Protocols & the Architecture of the Internet




Sponsored byDNIB.com


Sponsored byVerisign

Domain Names

Sponsored byVerisign

Brand Protection

Sponsored byCSC

Threat Intelligence

Sponsored byWhoisXML API

New TLDs

Sponsored byRadix

IPv4 Markets

Sponsored byIPv4.Global