A Trustless Internet

Vishnu Seesahai
Mar 14, 2019 · 12 min read

Cult leaders are some of the most dangerous people in society. Perhaps it is because they wield great influence over their followers, controlling their thoughts and actions through mechanisms like “blind faith.” Blind faith is a tool utilized to get followers to accept a message as truthful rather than requiring any evidence of that truth. In modern times, the cult of personality has soared to new heights. Fractionalization of beliefs and identities have lead to the rise of a new type of cult leader. They command loyal followings and use platforms to proselytize to their believers. The mind-manipulating force they wield is nothing short of a superpower. Sadly, that power is often used to wreak havoc upon the cults’ own constituents, in one form or another. In thinking about the origin of their power, we have to consider the source of their control. It has been proposed that trust is one such fountain of their power. Cult leaders have the uncanny ability to build trust through their charisma and influence. While we might have a tendency to focus on their personal attributes, ie. charisma et. al., this would miss the bigger point, trust.

The Hierarchy Of Trust

Philosophically speaking, trust has been eroding in society for a while. Humanity’s migration from worshiping god-kings to post postmodernity is evidence of the deprecation of trust. However, prior to the rise of large social constructs, tribes cultivated trust in a more meretricious manner. Tribe members had very specific functions, and if those functions were not precisely executed, the entire tribe would suffer. In this structure, it is difficult for an adversarial tribe member to sustainably misrepresent their output without eventually being unveiled.

As tribes evolved into large societies, hierarchical social orders arose. Typically, this ordering of people involved the tiering of access, in one form or another, and trust was established to be greater primarily at the top of the pyramid.

In this way, the leaders of these large societies could employ trust as a pillar of mass control. This totemic structure posits the necessity of an intermediary to validate all exchanges between members of the society. This structure is most evident when considering money.

Thus far, in society, the validity of a currency has been dependent on its central issuance, however bitcoin has offered an alternative to this paradigm. Bitcoin, and its notions of censorship resistance and “trustless” value store / exchange, is novel. And though it is still a very nascent experiment, its underlying philosophy, of in-math-we-trust, subverts the notion of requiring a centralized intermediary for transacting.

Non-blockchain people often remark that they don’t understand why the world needs bitcoin because it’s not “real money”, to which I often retort, “what is real money.” The de facto response is typically, “something I can buy things with.” While this is a highly pragmatic answer, it doesn’t get to the heart of the question. Though there are competing philosophies on what money is, they would all agree on one thing, that there must be trust in the medium of exchange for it to work.

Trust is earned, or at least that’s the way it should work. Centralization is a form of trust that is based on the idea of being too big to fail.

But, all parties must trust the issuer of the currency and that trust isn’t necessarily built on warranted trust. Indeed, over time, currencies tend to fail. Typically fail when the issuers’ trust has been eroded, either through corruption, currency devaluation, or general poor macroeconomic management. In today’s world of fiat currencies this can happen very easily. Fiat currencies aren’t backed by real world assets, which makes creating more currency as easy as pushing a button. As a consequence, these currencies tend to be inflationary, and require a centralized monetary policy to control this inflation. Centralized trust is flawed because of human malfeasance such as corruption, cronyism, censorship and human error. Any one of these flaws can lead to currency inflation which can give way to economic crashes cycles. Over time, this erodes centralized trust. Decentralized trust, through cryptography, has been proposed as a viable solution to the centralized trust problem. While decentralized trust isn’t entirely trustless, the necessity for centralized trust is greatly minimized by spreading trust across many independent parties, or nodes in a network. If this solution is in fact sustainable it can be migrated to other arenas.

Cryptographic Trust

Broadly speaking, cryptographic solutions to trust can be applied to any network which requires trust to function optimally. Cryptography can ensure, tokenize and commoditize trust as an asset within these networks. One such network which desperately requires decentralized trust is the internet. A “trustless” internet would allow its netizens to transact freely and openly without fear of being centrally compromised. Currently, we have an internet where any nation state, or bad actor, can censor, or otherwise invade the privacy of others. Internet Service Providers (ISPs) can meter or deny internet access at will destroying any hope of net neutrality.

The topology of the internet is highly centralized and with its dependency on ISP data centers and fat telecom pipes. This centralized dependency has introduced egregious bottlenecks which are currently affecting the internet’s scalability.

The internet, centrally hosted at data centers

The Centralized Web

The internet was built upon an antediluvian, 100 year old, telephone system infrastructure. As such, it has many choke points and bottlenecks. With the demand for wireless connections going exponential, the telecoms are racing for solutions, such as 5G and expansion of the cloud. Cloud expansion requires the development of new data centers and additional fiber-optic cables carrying trillions of bits per second across their lines. However, this all comes at cost of overloading and already overtaxed infrastructure.

Topology of the Internet

The predecessor of the internet was built by the Defense Advanced Research Projects Agency (DARPA) to be a decentralized network that could resist attacks by foreign powers. It was called ARPANET and it started in 1969 when a message was successfully sent from a computer at UCLA to another computer at Stanford Research Institute (SRI). Eventually this network would evolve into Usenet, which was the first really large online community. On this network, anyone could operate a server and connect to other machines allowing these computers to exchange messages. No one controlled this network. And, as long as you could dial into at least one server, you could gain access to the rest of the network. You could run an IRC server if you liked, join chat rooms or connect to public servers.

The World Wide Web came a decade after Usenet and it allowed anyone to host a web site on their own server. They could buy a DNS, a human-readable hostname, and point back to their server. Web sites could also link to each other which created a decentralized internet of information. But, even though this made life easier, the internet was still clunky to use and difficult to navigate. Hosting and maintaining a fast connection to your own server was a challenging task, even for the tech savvy. Thus, in an effort to onboard the masses, the internet became centralized. ISPs, like AOL, arose to offer internet access and email addresses all on a free trial basis. CD-roms were directly mailed to anyone with a mailbox and so the centralized web was born.

The modern iteration of centralization is more concentrated than it’s ever been. Our dependency on platforms such as google, youtube, facebook, instagram, twitter, and others has created a dependency on a highly centralized cloud.

The Centralized Cloud

The cloud is a global collection of high capacity fiber connected data centers. At these data centers are servers which host much of the content we engage. As such, most of the web’s information moves through the cloud. Hence, as the web grows, the cloud will need to be scaled. But this scaling effort is overly dependent upon telecoms and ISPs to achieve the required growth. This is problematic from a net neutrality standpoint as well as a routing perspective. Since the web is a connection of global servers which host the internet’s content, fast and open routing content is of primary concern. Metaphorically speaking, we can think of the web as a complex network of interconnected highways which move data across the planet. To navigate this highway system it is necessary to maintain routes to destinations. The web uses border gateway patrol (BGP) routing tables to route data. The growth of these tables has been enormous and there is significant debate as to whether or not it will be able to handle the internet’s future growth. BGP is poorly protected against manipulation, so existing ISPs are reluctant to connect with ,or trust, inexperienced ISPs. Consequently, participation in this market requires registration and approval from centralized registries to join this “web of trust.” This web of trust dependency has formed a veritable oligarchy of ISPs which dominate the internet today.

High hardware maintenance costs and large economies of scale have also contributed to ISP centralization, leading to high consumer costs and inefficiencies in routing. These inefficiencies cause bottlenecks when multiple people in the same area use the web simultaneously. To make matters worse, BGP peering is configured manually and subject to “hot potato” routing, which fails to route in the most efficient manner.

In practice, web traffic is routed from service provider to service provider until it reaches its final destination. As the name hot potato implies, ISPs route traffic by passing it off to another provider’s network. The routing is not necessarily optimized for speed, ie. the shortest route to the destination, instead it is optimized for lowering costs. Even though ISPs take the cheapest path to provide service, they continue to charge their customers as though they were not. For you and I, they consistently fail to deliver on the speeds and bandwidths they advertise. And since there is far too little competition in the ISP space, we have no assurance that they will not abuse their power. ISPs can throttle connectivity or sell browser histories in an unchecked manner, since their customers have nowhere else to go. It is clear that if the internet is to grow in a healthy way, with respect to openness and speed, it will need to do so in a more distributed fashion or the cycle of centralization will continue, perhaps to detriment of the internet itself.

To create the “trustless web”, the infrastructure of the internet must be open and resistant to censorship. Creating an internet for the people, by the people, is a fair way to approach this problem. Topologically speaking, this network’s nodes would form a mesh. In a graph theory, a full mesh is a graph in which every node is connected to every other node. A less constrained version of the full mesh is called the mesh graph in which every node has at least two or more connections to other nodes, and that there is a path from every node to every other node in the network.

Traditional Web Versus Mesh Web

Networks which have mesh topologies offer greater resistance to node censorship because there are no Single Points of Failure (SPOF). Similarly, bitcoin is resistant to SPOFs through its consensus method.

The Blockchain Internet

Technically, bitcoin can be thought of as a Byzantine Fault Tolerant (BFT) deterministic state machine. A state machine of this nature is a kind of distributed software which is fault tolerant, in other words, it has no SPOF. State machines are distributed over multiple processors, or server nodes on a network, in which these nodes are replicas of each other, thereby achieving a deterministic state. Bitcoin uses a Proof of Work (PoW) consensus mechanism to derive the deterministic state of the network by mining blocks of an immutable distributed ledger. If we apply bitcoin’s system of block mining to achieve consensus to the internet centralization problem, we would derive a structure by which the blocks mined would need contain verifiable proof that the internet’s nodes were behaving in a fair manner. Any nodes behaving unfairly, ie. not abiding by the rules, would be discovered and their blocks forked from the main chain. In this sense, the network disincentivizes adversarial nodes by preventing them from winning blocks, and thus restricting them from financially profiting from deception. Moreover, if the blockchain’s consensus algorithm could prove a minimum level of bandwidth you would have the ingredients for faster internet. Such an algorithm could assure bandwidth requirements are being met through a bandwidth hard function. Additionally, latency could be minimized through faster block resolution times. A high speed network could certainly support faster blocks since it was message propagation latency that lead Satoshi to safely settle on 10 minute block times in bitcoin. Indeed, the centralized web, and its inherent latencies, has conspired to limit bitcoin’s block speed maximization potential. However, on a meshed network which rewards nodes for low latencies, and requires high bandwidth to compute its PoW, choke points could be minimized and block speed maximized. With point-to-point encrypted traffic, and encrypted node addresses, censorship through Man in the Middle (MITM) attacks are significantly limited. If nodes cannot be targeted by their IP addresses they cannot succumb to malicious adversaries. Distributed Denial of Service (DDoS) attacks which plague the internet today could be reduced by enforcing routing timeouts on packets, preventing them from entering infinite loops or for lasting more than a few hops.

The Internet Of Money

As an added benefit, the trustless internet would allow its constituents to profit from maintaining the integrity of the network. This would give rise to a marketplace in which bandwidth could be both bought and sold. Most mesh networks today have failed to scale massively due to their dependency on volunteers and limited capital. However, a marketplace could incentivize the growth of the network through market forces. Since, all nodes can compete fairly to offer the best bandwidth experience to consumers, the marketplace would influence the expansion of the network’s efficiency and capacity.

While a blockchain could offer a provable mechanism to assure fidelity in a trustless internet, it is unrealistic to wait for block confirmations in order to access the web. However, faster routing over lighting payment channels could negotiate paths at near instantaneous speeds. The marketplace would incentivize the growth of these fast payment channels as it would be lucrative to do so. And since anyone can set up a lightning node, the network would not become centrally concentrated over lightning nodes as it is a free and open marketplace. Thus far, the internet has failed to establish a middle class which could support its infrastructural growth through proper economic incentives. If everyone could have profited from enhancing the internet’s expansion then a middle class would surely have arisen through mutual contribution to the web.

Though technological decentralization as an ideology is relatively new, it’s roots are old. Decentralization probably has its roots in democracy. For a democratic web to exist, it would need to maintain a high degree of freedom between its nodes, such that any node can reach any other node on the network. This equates into an internet which affords everyone the freedom to express themselves. A decentralized web could still censor nodes, but it could only be accomplished in a democratic manner, ie. the majority of nodes would have to agree to censor a node on the network. This would be challenging to coordinate in any centralized manner since nodes on this network are all private by design. And even if a node was booted off the network, it would have an opportunity to rejoin at any point in the future with unrestricted access to the web. This is a sort of native redemptive quality is a feature which doesn’t really exist in our physical world today.

Bitcoin’s BFT solution to the centralization problem seems very applicable to other network graphs. In the case of the internet, it is wholly possible that this sort of solution can help the web scale properly into the future. And if this is true, then maybe the future of trust itself relies upon its decentralization. Perhaps, if religion was decentralized in similar fashion, then cult leaders would be asked to proof their beliefs rather than relying upon the blind faith of their followers.

Vishnu Seesahai

Written by

Mathematician, Blockchain Architect, and Artist.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade