How to Build a Decentralized Web

November 21, 2019

The “New” Internet

By: Vishnu J. Seesahai

In season 6 of HBO’s “Silicon Valley” they set the bar high when visionary archetype, Richard Hendricks, sets out to build a new internet. The term decentralized mesh network is bandied about as the gateway to creating this new, more evolved, web. But, what does it really all mean? Why do we even need a new internet? Doesn’t the current one work well enough?

Before we talk about what a new internet looks like, we should first examine why it’s necessary in the first place. To offer a thorough examination of the current state of things, we will look at the stages of development the internet has gone through thus far.

The Next Revolution

The internet as we know it today, is a revolution in technology, but on closer inspection, its development has been more about evolution rather than any single revolutionary technological uprising. Specifically, we can break the internet’s evolution into three distinct chapters Web 1.0, Web 2.0 and Web 3.0.

Web 1.0 — Chapter one was born out of a need to connect computers together over great distances. With the advent of networking protocols such as TCP/IP, exchanging data became cheaper and more rapid than ever before. Prior to this the alternative was to mail data via. snail mail, but Web 1.0 eradicated this. This was the information revolution, this was Web 1.0. But, for all its capability it fell short of making the internet accessible to most. The majority of those who engaged with it were surfers and not content contributors.

Web 2.0 — A decade later, chapter two emerged, when the internet became more programmable, useful and easily authorable. Anyone could contribute content to this internet. In turn, this gave rise to a revolution in human interactions. The so-called Web 2.0 phase had begun. Out of this era came an explosion of online social interactions, bringing producers and consumers of information, goods and services, closer together than ever before. However, these interactions always required intermediaries, or middlemen, which perform as arbiters of these transactions. These middlemen gained significant financial and computing power, leading to the centralization of the internet’s infrastructure vis-à-vis: clouds, data pipes and data centers. Questionable, and often predatory, data practices have lead to abuses of data that range from weaponized marketing, compromised privacy, bandwidth throttling, censorship, and information manipulation, commonly referred to as information warfare. Either through ineptitude or intent, user data has been exposed to hackers, and government agencies alike. Absolute power can corrupt, absolutely. This is Web 2.0’s primary achilles heel, single points of failure, or SPOF.

Web 3.0 —Now the internet is undergoing its latest technological revolution, with the advent of the decentralized web, dubbed Web 3.0. The decentralized web has the power to disintermediate all interactions online, requiring no third party to interact or transact. Anyone can exchange information, goods, or services with anyone else, without the need of a centralized arbiter. This is a data revolution, one which impacts the design and deployment of data storage itself. This revolution places the power squarely in the hands of data owners, and reduces the hegemony of platforms to a fraction of what it is today. Under this new regime, platforms will be unable to weaponize, or compromise, data because they simply do not own it. In this land, the owners of data, get to choose what happens to it. Data storage, computing and bandwidth will be shared across many machines, globally, rather than a few centralized ones inside of data centers. This architecture will serve to support the rapid expansion of Artificial Intelligence, and IoT across the globe.

To achieve this transition, many Web 3.0 web technologies must transition to function as a world wide state machines. As a state machine, many independent machines across the globe work to achieve a consensus on the status of who is who, who owns what, and who can do what, as opposed to just a few machines that live within a walled cable garden. This is good for data, and it’s good for society as well.

Why Mesh?

Bitcoin was revolutionary in creating a disintermediated system of financial transaction. This is very necessary for commerce to occur on the decentralized web. But it’s not enough, because the very infrastructure of the internet is the single biggest weakness of all blockchain technologies. While blockchains are a powerful evolution in software, a similar evolution is required in network infrastructure to make it all harmonize synergistically.

Many see the Web 3.0 evolution to exist solely at the data layer, migrating databases from centralized stores to distributed analogs. But, this is fairly myopic in scope since data must move along physical infrastructure as well. If the physical infrastructure isn’t sufficiently distributed, it too will succumb to the vagaries of Web 2.0. Thus, it is essential that the physical infrastructure of Web 3.0 be rooted in a more robust decentralized topology, e.g. a mesh.

A mesh network is a topology of interconnected nodes in which every node can connect to any other node. These networks are dynamic and non-hierarchical in nature, as well as self-organizing and self-configurable. In other words, it is a network in which connectivity between nodes is maximized. Mesh networks offer solutions to some of the problems plaguing Web 3.0 running over the legacy Web 2.0 infrastructure. This is because the current centralized topology of the internet presents some interesting edge cases, for which Web 3.0 will fail if the infrastructure is not decentralized.

Disasters — During natural disasters, communications networks tend to go down. Resilience in communications systems is critical for rescue response and operations. Mesh networks can offer mission critical backups to support recovery efforts. Providing survivors with the ability to ping mesh nodes which can, in turn, be an important tool for finding and saving them. These would come from independently powered mesh nodes, which either self-power through solar and/or long lasting stores acquired from the energy grid, to be retained for extended periods of time.

Poor Coverage — Internet access is poor in many parts of the world. In parts of the USA there is a dearth of web accessibility, approximately 6% of the population in 2019. Typically, there are only a few internet service providers (ISPs) around to provide backhaul to the internet. In some cases there is only one, or even none. It seems shocking that in a nation that is the spiritual parent of the the internet, that so many would not have access to it. Worldwide, in 2019, 36% of all people lack internet access. Of those who do have it, most of it is mobile, and lacking significant bandwidth. Bandwidth is an issue even in developed nations, as the dependency on ISPs to provide access is high, and the motivation to maintain quality of service (QoS) is low. Most ISPs dominate the areas they cover, with little, or no, competition to compel them to provide better service. This lack of competition coupled with centralized infrastructure, has made it possible for ISPs to offer QoS without restriction.

Building out a wireless mesh network is a more cost effective way of scaling the internet. This is because the cost and complexity of installing fiber, and/or wires, between facilities is eliminated. Furthermore, additional wireless mesh nodes can be added, or removed, to scale up /down network coverage requirements. Mesh is low-powered, using comparable amounts of energy as Bluetooth. Therefore, it is possible to create limited lifespan mesh devices that are disposed of, and replaced in a few years time.

Edge Computing — In the case of edge computing, industries such as IoT, autonomous vehicles and AI are all in need of a new kind of network structure, to support massive computation. This is because computing requires fast data access and processing. Speed in computation is subject to two primary bottlenecks, raw computing power (CPUsGPUsFPGAASICs), and speed of data transfer, ie. the bus, or more broadly speaking: bandwidth. Bandwidth is both a measure of the amount of data that can be sent at a given time and the speed of its movement across a medium, such as fiber or air. In industries such as autonomous vehicles, bandwidth must be maximized at the edge, such that vehicles are able to process data from their environment and make rapid critical adjustments to navigate their terrain, ad hoc. Moving computation, locally, or closer to where it is required, is indeed indispensable to the future of intelligent data processing.

Censorship Resistance — The centralized internet has exposed data to SPOF with respect to hackers, or government agencies who wish to spy on traffic, block it, or even throttle access to certain kinds of data. While virtual private networks, VPNs, can help mitigate, somewhat, against this kind of failure, ISPs can still analyze traffic shapes to determine what a user might be doing, and limit this access regardless of encryption. An alternative method of routing requests / responses over the internet is needed to provide alternatives to the centralized approach. Networks cannot be considered truly censorship resistant until they can offer an alternative route to traditional centralized pathways. This option would not only be less subject to censorship, but it can also serve as a higher speed routing option to the traditional “hot potato” routing strategy offered by ISPs.

It’s clear that web 3.0 cannot merely a change to the data layer of the internet, without changing the highway upon which the data runs.

A Bandwidth Marketplace

Scaling mesh isn’t trivial. Although it is cheaper than building a fiber network, it still requires the financing and maintenance of network gear, such as: gateways, repeaters, and endpoints.

The costs associated have traditionally challenged the growth of mesh networks. However, if these costs could be offset, or if supporting mesh could be profitable, then the technology would take flight.

One path to mesh profitability is bandwidth commoditization. Bandwidth commoditization allows anyone operating a mesh node to profit by supporting the network with bandwidth. This is an effective strategy because of a simple fact, people are greedy.

Let’s face it, virtually everyone would like to make a few dollars more. And, if one could generate a passive income from sharing their internet bandwidth, then they are far more likely to operate a mesh node. Furthermore, the idea of contributing to a project that drives big telco towards a more consumer focused service, is compelling. This rally cry, combined with profitability, is the perfect combination for generating mass mesh adoption. For a mesh network wishing to grow large, externalizing the costs of scaling is strategic sugar.

But, how exactly would bandwidth be commoditized? From the mesh network perspective, any such commodity would need to have a dependent relationship to bandwidth, as well as the ability to be bought and sold, ie. transacted. Bandwidth based transactions naturally give rise to bandwidth marketplaces. A bandwidth marketplace could be considered analogous to the energy market, where energy can be harvested from a source like the sun and sold back to the grid. Similarly, bandwidth can be sold back to a mesh network. At scale, this marketplace will make bandwidth cheaper as supply increases faster than demand.

To service these bandwidth transactions, a ledger is required, to keep track of things. Such a ledger should strive to be truly decentralized, with a proof of work (PoW) based blockchain. Choosing a PoW that is “bandwidth hard” gives the network an incentive to constantly maximize bandwidth, because the PoW requires it. To win blocks, every mining node on this network would compete by maximizing their own bandwidth. This is different from traditional cryptocurrencies whose blockchains only maximize computation. However, creating a PoW rule set that maximizes bandwidth first, is the key to crypto-incentivizing mesh expansion. Mesh nodes, seeking to win a block, must prove that they adhered to the rules bandwidth maximization, and that proof is stored, on the chain, for all other nodes to verify. The reward for creating a block that adheres to the rules is a payout, or cryptocurrency, for playing fairly. Thus, nodes who maintain infrastructure, and help grow the network, are rewarded in crypto. In turn, this crypto could be transacted, or spent on optimizing routes for sending and receiving data, and soon the bandwidth marketplace could give rise to many ISPs all competing to provide the highest QoS for customers.

Striking a Balance

Solar energy is unlikely to replace traditional generation stations entirely, but it does help to remove some of the dependency we have on those stations. Similarly, ISPs are unlikely to vanish; however, they will no longer be the only game in town. The added competition will offer the much needed counter-balance to the internet’s ecosystem, setting us up for an explosion in web 3.0 technologies and services.

But Why Stop at Bandwidth?

The basic concept of a bandwidth marketplace hinges upon the notion of a bandwidth sharing economy. Similar to Uber, or Airbnb, people need only contribute a resource they already own to profit. Widely available excess resources, such as memory and computing power, could be contributed to other kinds of networks. Such networks include edge computing networks or decentralized storage networks. These other networks, can exist in the web 3.0 universe and thrive on top of the underlying decentralized infrastructure. A bandwidth first mesh network, which pays its nodes in crypto, is the ideal environment to foster the growth of other sharing-economy digital networks. This is because it is readily possible to externalize the costs of growing these hyper-networks. Motivated by the desire to increase profits, mesh nodes will also accept compute and storage jobs to increase their holdings of crypto. From this perspective, cryptocurrencies provide the necessary incentive for the digital sharing economy, to transcend merely bandwidth, extending sharing to all manners of technological capacity. With the proliferation of IoT, excess computing and storage may be ubiquitous, and many will attempt to generate passive residual income from this means of gain. But, like most opportunities, the time of greatest profit can be reaped early on in the growth cycle. Early adopters of a crypto-incentivized mesh network, stand to gain the greatest returns from the ensuing bandwidth gold rush. Thus, the time to support such a network is now.