decentralize, black and white

Ariel Deschapell is content manager for blockchain real estate startup Ubitquity, and a recent Henry Hazlitt fellow at the Foundation for Economic Education.

In this opinion piece, Deschapell argues that in order to solve issues around decentralized governance, the blockchain community must ask difficult questions about what kind of decentralization solutions need.

Screen Shot 2016-08-12 at 11.31.23 AM

Since the earliest days of bitcoin, decentralization has been key to its value proposition. But it's also been its greatest obstacle.

Whether it's the block size debate, or the ethereum classic debacle more broadly, decentralization in public blockchain networks presents significant hurdles to what can seem like straightforward objectives. After seven years of open-source study, decentralized governance remains a little explored and unsolved enigma.

But solving it will likely mean going back to the beginning, asking what exactly we mean by 'decentralization'. Do we mean the distribution of hash power? The number of nodes? The inherent ability to fork and secede as recently demonstrated by ethereum?

The exact definition of decentralization in cryptocurrency debates depends on the context. Yet, semantics and technical terms tend to obscure the fact that, at its heart, "decentralization" refers to a system of voluntary cooperation between peers.

Efficiency trade-offs

Decentralization is a means to an end.

As a tool, it is not optimal for every possible use case. In fact, decentralization tends to be incredibly inefficient compared to centralized solutions. When it comes to transaction throughput, for example, bitcoin lags far behind centralized payment networks like MasterCard or Visa.

Yet, network distribution is a necessary prerequisite for preserving the consensus rules that give bitcoin its value, such as immutability and progressive reward halvings. Distribution of the network promotes extreme redundancy to safeguard against censorship and attacks that would threaten these well-defined consensus rules.

This comes at the inherent expense of efficiency in transaction throughput, power consumption and the overall pace of development.

Yet, the value of a digital network that can preserve these rules is so great that these costs are proving to be justified. As the foundation of the entire decentralized ecosystem, it becomes essential we understand and are able to maintain network distribution before we can start on the problem of governance.

A relative measure

Like hash power, network distribution is a pivotal factor when it comes to blockchain security.
 However, we have no definitive metric for measuring hash power distribution.

Fortunately, that doesn’t mean we’re blind when it comes to determining preferences. Network distribution itself consists of identifiable contributing factors.

These include the number of nodes that propagate transactions, the amount of mining machines, the number of operators behind both miners and nodes, the geographic distribution of it all, and the number and size of mining pools.

If we isolate any of these factors, it's trivial to determine what more and less distributed looks like.

For example, 100 independent miners spread across the world are clearly more decentralized and less vulnerable than 100 clustered in the same region. But it's when we start to consider trade-offs that it begins to get murky. When we do this, the weight of these individual factors becomes largely determined by subjective preference.

Is it more decentralized to have 20 globally spaced miners, or 200 in close geographic proximity?

More decentralized

As we can see, when all else is equal, identifying what is more 'decentralized' is easy despite the lack of a standard unit of measurement. It's when it comes to making potential development trade-offs on the protocol level that it gets difficult.

This is compounded because, while decentralization is a means to an end, there is an important fact we are missing. How much and what kind of decentralization do we actually need? What minimum amount of network distribution is required to ensure bitcoin or any cryptocurrency continues to maintain its security? The answer: nobody has any idea.

The reason for this is no one can predict the scale and method of future attacks that will be carried out on the bitcoin network. Given the right circumstances, a single Black Swan event in the form of a formidable attack could have tremendous negative impact.

If bitcoin is to become as successful as many hope, such attacks should not be out of the question.

Whether they are carried out by extremely well-positioned private parties shorting the currency, or an organized collection of state institutions determined to stamp out its popularity, entertaining wildly successful bitcoin adoption demands we take serious the possibility of coordinated attacks on the network.

For this reason, if any public blockchain hopes to become the core of a truly pervasive and global financial web, it must be prepared for the worst. This means developers should be inclined to encourage network decentralization.

Delicate balance

But this alone is no easy task. If we could determine the minimum amount of distribution needed to ensure the worst possible attacks would fail, and could ensure network changes do not push it below that threshold, then vetting development decisions would be trivial.

But we can't do either of those things. Bitcoin is a system of voluntary peers, and therein lies the difficulty.

We can't force stakeholders to run full nodes, or prevent miners from joining pools that are already a certain size. All we know is that greater decentralization is generally more secure, and the only way to encourage greater decentralization in a voluntary network is by incentivizing it.

For node distribution, this means lowering the cost of running one or increasing the value of doing so. For mining, this includes improving block propagation to negate an advantage of larger pools. Such developments would see the network become more distributed than it otherwise would be by making it easier, less costly or more advantageous to become a peer on the network.

This isn’t to say greater distribution must come at the cost of everything else.

There are in fact trade-offs that are likely worth making for less decentralization. Bitcoin's hashrate dwarfs the combined power of all the world’s supercomputers. It represents the raw computing power securing each new block of transactions on the blockchain, and it wouldn't be possible without specialized mining centers.

Most trade-offs are not so clear however.

By allowing for larger blocks, the proposal would increase throughput, but like all economic actions this comes at a cost.

Larger blocks demand more computational resources from nodes and are more difficult to propagate amongst miners. However, because non-mining full nodes lack monetary incentive, the benefits derived from doing so will remain the same after the limit increase. Because the cost of running a node rises with the size of a block, and the benefits do not, then all else being equal there must be more nodes dropping off the network than there otherwise would be.

Next challenges

This alone doesn’t tell us if such a change is worth implementing. But because the cost to the network in terms of distribution is not zero, the burden of proof rests with showing there is a beneficial and pressing need to do so.

In this particular case, this means showing that the block size limit is the limiting factor when it comes to bitcoin adoption. If it is not, then there is no pressing need to increase throughput and we can wait for solutions that do not risk impacting distribution at the base network layer.

Given the importance of network distribution and the inherent constraints in measuring and controlling it, this should be the standard criteria for vetting decisions which can sway it one way or another. Ultimately, the more distributed the network is, the more secure and certain its future.

If the prevailing ethos of modern law is "innocent until proven guilty", then the guiding ethos of blockchain development should be "decentralized until proven otherwise".

So long as network distribution can be adequately maintained, it can serve as the basis for a large and decentralized ecosystem of stakeholders and contributors who depend on a secure and reliable blockchain. The next challenge is understanding the relationships and incentives of these disparate stakeholders and determining how they can best work together to progressively improve the ecosystem without a centralized decision maker.

Early objections against bitcoin focused on the question of whether or not a deflationary currency with no sovereign backing could possibly become proper money. Yet the skepticism was misplaced. This isn't the most immediate or even the greatest challenge facing bitcoin and other cryptocurrencies.

The much more important question pundits should be asking is can bitcoin or its successors pull off decentralized governance?

As a new phenomenon, the decentralized ecosystem of cryptocurrencies presents novel problems for stakeholders and independent thinkers alike. These challenges are undoubtedly great, but so are the potential rewards.

Black and white distress image via Shutterstock

The leader in blockchain news, CoinDesk strives to offer an open platform for dialogue and discussion on all things blockchain by encouraging contributed articles. As such, the opinions expressed in this article are the author's own and do not necessarily reflect the view of CoinDesk.

For more details on how you can submit an opinion or analysis article, view our Editorial Collaboration Guide or email [email protected].

Block SizeDecentralization

Load Comments