The internet is a magnificently complex, diverse and distributed network which has been the product of decades of research. This has, in a short space of half a century, become one of the dominant means of communication today. Owing to its “agnostic” nature (for the most part), it can be used to carry data from place to place which represents virtually anything – text, images, video, audio, 3D models, binary data or programs, commands, telemetry, etc. It can do so efficiently, almost instantaneously, and relatively affordably. This has made it almost essential to our everyday lives, and something which is coming to be recognized as a human right.
Unless you’ve been “living under a rock”, it’s likely that you have come across the FCC’s move to repeal net neutrality within the USA, spearheaded by the FCC chairman Ajit Pai. While it has spurred a large number of discussions and reactions online, there seems to be a very large number of misinformed people who still don’t understand (or have a partial and distorted understanding) how the internet works, and what net neutrality really means. Worse still, they don’t understand exactly why it matters. But now is no better time than to be concerned about the internet which we take for granted on a daily basis.
The Concept of Neutrality
In a nutshell, the internet as it was conceived, was nothing more than a number of computers connected via a number of links speaking a common protocol. The computers and intervening network equipment handled packets of data, and moved them from the source to the final destination in the most efficient and reliable way possible based on a set of rules. In the early days, there wasn’t the computing power to worry about what was within these packets, and as a result, they were handled as if they were all equal. As a result, the internet was “born equal”, which is an expectation that carries on today (although with some limited exceptions).
You can imagine this as a primitive postal service that handles parcels up to a particular size, with addresses written in a particular format. The postal network looks at the addresses and moves the parcels as efficiently and reliably as possible from source to destination and doesn’t care what your parcel contains – be it a gift, newspaper, baby formula, toys, or junk. To the system, it is just any other parcel.
Is Anything Ever Truly Neutral?
While the concept of net neutrality is great in an ideal sense, the truth is that nothing is quite as ideal in reality due to a number of complexities. While it was initially neutral by default, the internet itself has become increasingly less “neutral” for a number of technical, legal and commercial reasons.
With continually increasing data traffic, and limited investments into network capacity, during peak periods contention is a problem. Because some types of data are more sensitive to delay or loss, some providers have taken steps to implement a Quality of Service (QoS) management system which throttles dataflows which are less time-critical.
With newer protocols that are more aggressive at attempting to “game” the system during periods of congestion (e.g. inelastic flows which do not respond to congestion indications), newer and fairer queue dropping systems have been implemented to prevent those with a larger number of flows or bandwidth from dominating the available bandwidth in a shared link.
For legal (and occasionally business) reasons, deep packet inspection technologies have been implemented on some networks. This technology may be used to help record metadata in regards to data transactions and characterize dataflows for their importance, destination/source or type, thus the type of data is increasingly becoming a concern to the network that carries it.
Where wireless/satellite links and shared-medium mobile carriers are involved, traditional TCP/IP networking protocols may not perform well due to random packet loss and round-trip latency. As a result, some more sophisticated networking equipment will intercept the session and provide “spoofed” acknowledgements, handling the forwarding of the packet to the end user in a manner more suited to the limitations of the underlying transmission medium.
These matters are discussed in more details in the relevant subsections which follow. While some may take this as an argument that neutrality is unimportant, that is not the case. With each of these are relatively benign examples of data-discrimination that have mainly occurred for technical/legal reasons, their introduction is not without their own problems. Instead, this is what has happened more out of need rather than want.
Business-motivated neutrality issues have already come to exist which mirror the frustrations consumers might experience from a telco provider offering “free calls (only) within our network”, and postal services offering “rate discounts for registered businesses”. These are arguably worse, and demonstrably not in the consumer’s best interests.
Why is Neutrality Important?
Net neutrality is important as the internet has become a dominant means of communication, a form of a utility for the public good, and a common carrier. To have net neutrality is to ensure that there is no discrimination against the packets which traverse the network – ensuring everyone has free and equal access to use the network to convey any data they wish. This ensures an open internet where all forms of innovation and technology can thrive and compete on a level playing field where communications can happen truly on a global scale from end to end. This ensures that creators and service providers can be assured of fair access to users, and users can be assured the widest choices as the “internet” means the same thing no-matter where you are or who you’re connected with.
This is a form of anti-discrimination – much as how people expect to be treated fairly without any prejudice no matter their age, background, gender, etc.
There are already a number of cases where neutrality is violated, and its effects can be felt by the end user. This would make the case that neutrality is important, although absolute neutrality may not be achievable or practical.
Without Neutrality: Restricted Access
One possible outcome of a lack of net neutrality is the problem of restricted access. This is where access to certain services are blocked, either by type of service, address or hostname, port or protocol.
One example where you may encounter this is public Wi-Fi access points. Owing to the vast number of potential legal and safety issues, many networks implement blocks of various degrees without much consistency. As a result, users connecting to such networks may find certain instant messaging apps not to work, e-mail access to be impossible, VPN access to be impossible, SIP calling to be impossible, etc. This is an unfortunate outcome – for example, if I’m at an airport and I’m trying to contact a relative and my SIP phone isn’t working, my WhatsApp isn’t connecting and my e-mail isn’t sending … that’s not really real internet. It may be enough to browse, but the internet is more than that.
Another example is on mobile networks where certain destinations/protocols/ports are blocked to avoid unexpected data charges, but also, to keep networks from being saturated by unnecessary P2P or bulk downloads. This could include watching YouTube at higher quality settings (on the premise that you’re on a mobile phone and don’t need them). While this is a convenience for some, it is also a hinderance to anyone that relies on such connections as their primary means of getting online. Ultimately, I believe the user should be the one that decides if they want 1080p or if they want to run P2P or not.
While the first two examples are rather innocuous, the third is the recent ruling that Australian ISPs be required to block access to sites by court order. This interferes with the “transparency” of the internet as a whole, where ISPs and lawmakers become the gatekeepers and sites can be “blocked” for a court order and a mere $50. The fact that this was allowed to pass was a rather disappointing moment for the freedom of the Australian internet, but as a whole has proven vastly ineffective for the simple reason that circumvention is not particularly difficult. Another blight on our freedom was the introduction of mandatory metadata retention beforehand, which we surmised would have resulted in the equipment necessary to perhaps invade privacy and block access being installed and activated. Were we alarmist for being concerned? Arguably not.
Unfortunately, this seems to be a slippery slope which many countries are falling into the trap of. For example, in the UK, there is internet censorship where sites are blocked by default with an “opt-out” list being recorded for those wanting to visit such sites. Numerous incidents of over-blocking have been recorded as a result.
At the extreme end, there is the Great Firewall of China, which is used to control access to overseas websites. This has proven to be a constantly-evolving filter, denying direct access to a large number of western sites and which actively targets VPNs of various types based on protocol. Where access is not directly denied, it is very much limited in speed.
In the future, it may not be unfathomable that access to certain blocked sites or services could be restored with an additional fee, similar to how cable TV subscriptions worked.
Without Neutrality: Throttled Access
While they might not deny you access outright, as that’s an obvious issue that could be bought up, they could take steps to reduce the speed of access. This could be a very subtle technique, which could be effective on easing network congestion for the ISP or denying/degrading access to certain services without being easily caught or with the benefit of potentially shifting the blame elsewhere.
In a world of instant gratification, having the patience to wait for something is becoming rarer and rarer. Users generally expect timely, speedy website loading, and the better sites go to great pains to optimize this to maximise the user experience. By reducing the speed of potential alternatives, carriers can create an advantage for those “in the fast lane” and then charge them for this ability. This increases the cost of providing a service artificially while degrading the user experience. Taken to the extreme, it could make some services completely unusable which makes it almost as good as blocked.
While initially, the web was mainly used to convey text and simple images, the proliferation of multimedia protocols and applications means that there are lots of real-time applications which have stringent requirements on throughput, loss and latency to ensure an adequate user experience. This could involve one-way server-to-client style streaming, as well as two-way video or voice-over-IP calling. By throttling access, ISPs may cause such services to revert to lower-quality lower-bitrate services which reduce user experience and preferentially treat their own services with a higher level of priority, creating an artificially uneven playing field which they have no intention or impetus to resolve. This has happened in the past, due to inter-carrier agreement issues, where Netflix was finding its traffic adversely affected and consumer experience degraded. While part of this comes down to network architecture and limitations, coupled with a lack of investment and arguments as to who should bear the cost, ultimately the consumer loses out by not receiving the quality of service they expect without being able to find a resolution. While neutrality is not the solution to the issue, I’d have to think that if they weren’t so concerned about who or what the data was, they might have sorted it out a lot sooner. That being said, even in Australia, Netflix and YouTube is not immune to such issues.
That being said, a lack of neutrality in terms of throttling access during periods of peak demand may not be such a bad thing if it ensures that real-time services are provided to end users at the expense of non-real-time transactions. This is a case where absolute neutrality may not be as useful, although this comes with the added wrinkle of determining which services require what level of resource and deserve what sort of protection. Traditional forms of QoS marking such as diffserv are prone to being spoofed by end users, making them untrustable, and are not necessarily obeyed by the network carrying the packets.
A truly agnostic network will not have to concern itself with this problem. The same issues exist where protocols designed to “saturate” the network with inelastic flows hoping to gain a throughput advantage may cause less-aggressive protocols to suffer. This is something which is being addressed through new queuing algorithms, although in itself, adds to the complexity and “judgement” of the network.
Without Neutrality: Interference with Access
Another issue of a lack of network neutrality is the potential for interference with access through spoofing packets/websites, man-in-the-middle and man-on-the-side attacks. While it may seem farfetched, some early attempts at controlling P2P traffic involved spoofing TCP/IP connection reset packets causing the connection to “drop” as if an error had been encountered when none had been. The Great Firewall of China a multitude of techniques as well to break VPN connections periodically. American ISPs have been particularly annoying in their pursuit of profit – they have even put in man-in-the-middle injections into HTTP sites to overlay ads and alerts on other sites potentially interfering with tier function and imposing privacy concerns. The US’ NSA has a QUANTUM program which relies on man-on-the-side and deliberate timing exploits to hijack connections over the internet. Finally, such interference could be used to great effect to attempt downgrading connections from secure modes to less secure modes. Had the internet been the “dumb” set of pipes it was, such carrier-level interference would be much less prevalent, at least at the ISP level.
In the present time, there are commercially-motivated interferences to access such as geoblocking, for market segmentation reasons. Unfortunately, this comes about due to a multitude of legal and commercial reasons, and is not something net neutrality can necessarily fix but is a form of interference with access.
There are potentially positive aspects to this – in the case of mobile access, some carriers have employed transparent performance-enhancing proxies which “terminate” the connections at their end and forward them to the mobile end user in a more efficient manner, enhancing the throughput by reducing the impact of high random packet loss and delay on the wireless link from affecting the TCP/IP connection’s flow through buffering and better protocol refinements. While this is conceptually positive, it does come with the potential downside of privacy concerns as well as breaking the true “end-to-end” nature of the communication as acknowledgments received may not be from the end user terminal.
Another potential interference is the provision of “on network” services, which can only be accessed within a particular provider and through no other as a form of vendor lock-in. This was popular in the early days of AOL and Compuserve, but had been “killed off” by the internet. With the allure of profits, such exclusive content may see a return in some way – starting with the next point of “zero-rated” access and package bundle deals.
Without Neutrality: Free “Zero-Rated” Access and Package Deals
Without net neutrality, a future where services are unfairly treated is indeed a possibility. While many “free market” capitalists will continue to scream that neutrality will be maintained as providers would be shooting themselves in the foot otherwise, the truth is far from that simple.
In the first instance, there have been a number of providers which offer “zero-rated” access (i.e. unlimited, quota free) to particular services. This means that those services which are in this “umbrella” are preferentially treated, and any competing services are (essentially) given a financial disincentive. In India, Facebook’s Internet Project was met with negativity precisely for this reason – there’s a difference between a real internet which fairly treats all services, and one where preference is given to a select number of competitors. While some zero-rated access may be beneficial (i.e. the user wants it, or it can reduce network loading), it’s often at the detriment of the provided access arrangements – e.g. free social media + 1Gb per month, or 4Gb a month to spend on whatever you like.
This does extend further, as the other option is to provide “unlimited” access to a particular service (with some restrictions) for an additional monthly fee. I met this even in Singapore when browsing for a SIM to use in my travels, only to be disappointed that the services I wanted to use were not part of any such options. This preferential treatment serves to steer users to use free services. The way the carrier provisions such free services may also be a concern, as they may have a local proxy combined with some DNS rewrites to make it work, increasing the fragility of the service.
This easily extends to us in Australia, where a number of ISPs have made package deals with Pay TV and streaming providers for “bundled” services at one price. This offers the consumer a (potential) saving, but locks them into the bundled provider. Likewise, some ISPs are offering “quota free” access to some streaming services and not others.
Why Should I Care? I’ve got a choice!
One common counter-claim is that the whole issue is a big non-issue as if one provider violates neutrality in a way that was material to the consumer, all they would have to do is change providers. In some cases, this may be true, however, in many cases it isn’t and it also is not practical.
The USA is particularly interesting, as many areas are not well served with options for high speed broadband internet. A large number of users are connected through HFC technology, owned and run by cable providers, some of which have long been recognized as the worst companies around (such as an unnamed company that rhymes with ROM-blast). Owing to the way the coverage and competition within the US cable industry works, and the fact that areas are often covered only by one cable provider, choices of switching to other ISPs are a non-option for many simply owing to the disparity of service between their cable provider and any “other” provider they can use.
Even if you could change providers, there is still the issue of contracts and downtime. There is no guarantee that they would maintain such access arrangements, as they could also change their minds. Short of a solid backing for neutrality, ISPs are essentially free to do what they feel to be in their best interests (including selling your data for money).
Such motivations are not exactly unfathomable, as their bread-and-butter earnings from cable subscriptions have taken a hit due to the number of “cord-cutters” going for streamed “over-the-top” services via the internet. It seems the ISPs probably want to take a cut in some way and may not be above affecting their competitors’ traffic. They’ve already confused consumers with silly fees on their bills and been caught.
Finally, if all the providers in an area are bound to enforce the same rules (e.g. website blocks, nationwide firewall), then there is no recourse for that short of attempting to access the internet through a different gateway (e.g. by VPN).
Why Should I Care? I’m not in the USA!
A common misconception is that the whole net neutrality business is really only something for the people within the USA to worry about. The truth is that it matters to everyone in a number of ways.
The first reason is that the internet knows no geographical boundaries (generally speaking). As a result, while I may comfortably reside in Australia, the sites that I visit may involve traffic passing through several USA-based carriers – this may be through larger backbone providers such as Hurricane Electric or via more local providers such as AT&T, Cogent, Level3, Cox, CloudFlare, Amazon, etc. This may be to visit USA based websites, or even websites in Europe, owing to how ISPs choose to partner with various transit providers to offer connections at the lowest cost. In fact, this very website is being served through USA’s networks – I rely on Tata Communications and Level3 to get there. Not enforcing requirements for neutrality from providers will likely lead to cases where there is a commercial incentive to fast-lane certain types of traffic as opposed to others, which may result in other types of traffic which may not be as “profitable” taking a back-seat. This means slower connections, higher packet loss which may make certain real-time communications unusable.
The second reason is that this may set a precedent for other countries to follow the lead of the USA to abolish any protections they may already have in place, and slowly, perhaps evolve into a more “locked-in” profit-driven model as we fear. Competition (in many cases) just does not truly exist.
Why Should I Care? You can just use a VPN!
A big misconception is that users can just “buy a VPN” or “use TOR”. These people are either in the business of selling VPNs, or they just have no idea how they work. They think the “darknet” is not part of the internet itself, but this is evidently exactly the opposite of what is the case.
A VPN (standing for Virtual Private Network) is basically a way of making a private network connection over a public internet. It does this by creating a network interface, where all the packets to/from this interface are “encapsulated” into encrypted packets which traverse the same public internet. This gives you the illusion that you have a dedicated connection to a private network somewhere else while remaining “secure” through the use of encryption. From there, (as is commonly done) your traffic can “emerge” into the public internet as if you were connected elsewhere.
The obvious issue with this is that your traffic still traverses the public internet, the same internet whose neutrality is compromised. Your VPN traffic could well be discriminated against and interfered with because somebody doesn’t like it, and thus you’re no better off than before.
The same thing applies to “darknet” networks which are an overlay running on the same public internet just with different protocols (some of which are easily identified and discriminated against).
Neither of these solutions are building a new network entirely circumventing the one provided by your ISP (potentially one that is not in favour of neutrality).
Worse still, these solutions make things worse. For one, because your traffic is now travelling through the internet to reach the VPN gateway (which could be in another country), the traffic may have to travel all the way back to access the resources you are after. There is also a penalty in terms of encryption/decryption and packetization. This is likely going to increase latency and decrease performance as ACK packets can be delayed, MTU can be reduced and BDP increases. It will likely result in more traffic to access the same resources, straining the limited resources even worse. Its performance is also highly dependent on the resources of the VPN provider, who themselves may be under resource pressure. It also means that the end user is now paying for an additional service just to get to the things they want to use when it could all be avoided.
While a VPN might get around some aspects of net neutrality issues (e.g. geoblocking, some types of deep packet inspection/QoS throttling), it’s hardly a reliable solution as it could itself be subject to the same tampering that you are trying to avoid.
What is the Upside of a Lack of Net Neutrality?
The repeal of net neutrality doesn’t really present many positive sides by comparison. It’s key to note that most of the claimed benefits are not grounded in facts at all as exemptions already exist within the US framework to handle such cases. The repeal seems more grounded in commercial and government interests than in the interests of the public.
It’s true that absolute neutrality is probably unachievable and counter-productive, especially for real-time services. However, I don’t believe that the solution is to give up on the ideal situation and just “leave it to the operators” to do as they please. That would be throwing out the baby with the bathwater. Arguments that the internet didn’t implode prior to the introduction of net neutrality rules are missing the key point that the internet didn’t implode after their introduction either.
One benefit would be that less regulation would mean less work for the government – essentially wiping the FCC’s hands clean, while also rendering it powerless to intervene should anything happen. Another may be that providers spend less time worrying about whether they’re complying with regulation or not, which saves on costs. It might also allow for more experimentation and flexibility in terms of the way the network is architected and run. But whether that is going to be used towards the good of the end consumer remains to be seen. The ultimate cost is a loss of protection to the end consumers.
I think it’s clear that from a consumer point of view, the “idealist” concept of a neutral internet is important for preserving fairness, equality of access, choice and competition on a network which has become so ingrained in our lives that it should be considered a public utility. The internet was “born” equal, and while it has been pushed further than originally intended resulting in some “inequality” around the edges, the expectations of users and companies that rely on it is that it should remain as close to equal as practical. Having net neutrality codified in law is one way upon which this quality could be protected.
The truth is that equality benefits the users, as well as the providers and creators generally speaking. Those that stand to benefit from the loss of neutrality are likely to be the businesses providing the access, and the government. To see that it was repealed was a sad outcome, leaving consumers less protected. However, the uproar that was generated has been good for the understanding of consumers, and sets a precedent that no-matter what the ruling, most were overwhelmingly in favour, despite the choice of the board members.
It seems surprising that the issue has gotten this far, as I couldn’t think that it would ever be repealed owing to its fundamental nature. Some people believe it will not really mean much as its repeal is only in law and not in the direct operational nature of the ISPs. However, this is an implicit trust we are placing on the ISPs to be in favour of the consumer, something which American ISPs seem to have lost on numerous occasions. For one thing, I couldn’t fathom who would take the name of regular consumers and post in comments in support of net neutrality, if it were not a company with a vested interest in its repeal. The conduct of the FCC is also unconscionable in the regard of how it dealt with submissions and follow-up investigations.
I just hope that nothing bad comes of it, and other countries do not take the lead of the US in repealing protection for the end consumers. Don’t let the internet we’ve taken for granted become further tainted by commercial interests.