Some Blockchain Challenges

Notes from the Vault
Larry D. Wall
June 2018

A blockchain is a type of distributed database that is being touted as likely to dramatically restructure a wide variety of activities, including accounting, payments, personal identity, property records, and even voting.1 Yet in practice, the use of blockchains tends to be limited to proof of concept and small-scale production.2 One blockchain-based application has had some success in displacing more traditional methods of finance, initial coin offerings (ICOs). However, ICOs' success appears to be less related to any superior feature of the blockchain and more to its ability temporarily to avoid compliance with securities regulation.3

The gap between the blockchain-enabled disruption envisioned by some, and the rather limited usage of blockchain in practice, is due to a variety of factors. One major factor is that many of the proposed blockchain use cases require coordination across different organizations that often have very different objectives and incentives. However, another important factor is that there is no single way of implementing blockchains, and the existing implementations are still immature in many respects. This raises the important question of how to move to implementations that could actually prove transformative. Looking into the future, a similar question arises as to how particular blockchain implementations evolve to meet changes in technology, user needs, regulation, and other developments.

This post is the first of a two-part series related to blockchain evolution. This first part provides a high-level discussion of the workings of blockchain technology and some of the related challenges. The second post will discuss how to change blockchain implementation with a focus on governance.

What is blockchain technology?
As with some other terms used in technology, there is no single agreed upon definition of a blockchain.4 For our purposes, Wikipedia provides a good definition: "A blockchain…is a continuously growing list of records, called blocks, which are linked and secured using cryptography." Although blockchain technology builds on some prior work, the first blockchain is attributed to a paper by Satoshi Nakamoto.5 Nakamoto's intent was to develop an "electronic cash" called bitcoin that could be sent from one party to another without relying on a third party and without creating the risk of "double spending," or the same cash being promised to two different parties.

The critical contribution of the bitcoin blockchain was it provided a way readily to identify any attempt to change any prior block (set of transactions). The blockchain does this by creating a "cryptographic hash," or unique digital summary, for each block that includes the hash from the prior block. If even the smallest change is made to any given block, it will alter that block's hash and the change will carry through to every block subsequently added to the blockchain. (This occurs because the prior block's hash is part of each block in the chain.) Thus, if one wants to rewrite the history of one block (such as to take a payment originally sent to Jane and instead send it to John), then every subsequent block will have to be revised. Otherwise, anyone seeking to verify the blockchain will be able tell that an effort has been made to change one of the blocks.6

Another key part of the bitcoin blockchain is that in order for someone to add a block to the blockchain, he or she must first solve a cryptographic problem that is difficult and expensive to answer, but where it is low cost to verify that answer. The people who solve such problems, called miners, are compensated for their work by receiving new bitcoins from the blockchain program (and sometimes also from payments by people who want timely execution of their transaction). The benefit of such a costly proof-of-work process is that it raises the cost and difficulty of attempting to rewrite history. Anyone attempting to rewrite history without detection must have the capacity to write new blocks faster than they are being created by the rest of the bitcoin mining community.

A potential problem for a proof-of-work blockchain is that two or more miners will sometimes simultaneously solve the cryptographic problem and write out a new block. Bitcoin resolves this problem by designating the longest chain as the definitive one. This solution means, however, that transactions recorded on the shorter chain are discarded as if they never happened. Thus, a widely quoted rule of thumb on bitcoin is wait for six blocks before regarding a transaction as being confirmed.

Given the difficulty of changing prior blocks, blockchain enthusiasts like to promote blockchains as providing tamperproof, immutable records. However, these claims are subject to qualification and are not necessarily a virtue in all use cases. The claim that blockchains are tamperproof and immutable is not based on the argument that changing the records is technically impossible but rather that the costs of rewriting all subsequent blocks would likely exceed the economic benefits. However, as a paper by St. Mary's professor Angela Walch observes, the underlying code in both bitcoin and ethereum (another major blockchain) have been changed in ways that effectively abandoned what had been legitimately entered records.7

Moreover, the argument that it would be uneconomical to attack (try to alter) a blockchain is subject to some substantial qualifications. University of Chicago professor Eric Budish (2018) provides an analysis of the economics of attacking a blockchain that relies on proof of work and concludes that "the model suggests that Bitcoin would be majority attacked if it became sufficiently economically important…which suggests that there are intrinsic economic limits to how economically important it can become in the first place."8

Further, there can be situations in which retaining a record forever is not desirable or legally acceptable. For example, the recent implementation of the General Data Protection Regulation by the European Union provides a "right to be forgotten" that may conflict with the immutability of records on a blockchain in certain circumstances.9

Although the blockchain's features are sometimes overstated, the case for using the technology need not rely on it being completely tamper resistant, perfectly immutable, and never relying on a third party. A sufficient condition is the technology prove superior for some use cases relative to alternatives. For example, a blockchain use case does not need to guarantee it does not rely on trusted third parties for the accuracy of new information added to the blockchain. It is sufficient that the blockchain be no more reliant on trusted third parties than other methods of storing data and the use of a blockchain provides some other advantages—such as greatly reducing the probability someone can tamper with the original records. Whether and for which applications blockchain technology prove superior to the alternatives remains to be seen. However, at this time there are a large number of people working on blockchain applications in the expectation the technology will prove the superior alternative for many uses.

Some of the challenges facing blockchain technology
Bitcoin, the first major and in some respects still the most important blockchain, reflects a set of choices about how to implement a blockchain. There are, however, additional implementation options that have been developed with their own advantages and disadvantages relative to the original bitcoin implementation. For example, the entire bitcoin ledger is public, creating the potential for individual users to be identified personally in certain circumstances, whereas some other blockchains are designed to provide greater anonymity.10

However, probably the most important alternative implementation of a blockchain designed to address a limitation of bitcoin is ethereum. Although there are a number of differences between the bitcoin and ethereum blockchains, probably the most important is the bitcoin blockchain was designed primarily to allow for the transfer of a cryptocurrency, whereas ethereum's blockchain bills itself as "a decentralized platform that runs smart contracts: applications that run exactly as programmed…."11

The choice of implementation options has a number of important implications for the performance and functionality of a blockchain. Arguably, the most important implementation choice is that of how to determine who can add records and how the additions are verified. Proof of work, as currently used by bitcoin and ethereum, (at this time) requires an enormous amount of energy consumption. Indeed, digieconomist.net's page estimates that bitcoin's energy usage exceeds that of many countries, including the Czech Republic, and the average bitcoin transaction uses "several thousands of times more energy than the average non-cash transaction in the regular financial system."12

Another of the more important problems with blockchains that rely on proof of work is that they tend to have low transactions speeds. Whereas Visa reports its network is capable of handling more than 24,000 transactions per second, howmuch.net says bitcoin can handle only seven in the same span.13 The low transaction speed for bitcoin implies that those needing timely completion of their transaction have at times in the recent past needed to pay a median transaction fee to miners of around $30 to include their transaction in a block.14

There are a variety of alternatives to proof of work that have been implemented in blockchains (or are under development).15 Probably the most popular of these is called "proof-of–stake," which assigns the rights to write, or "forge," new blocks based on their ownership of cryptocurrency or tokens on that blockchain.16 Another alternative that is being developed are permissioned blockchains. A permissioned blockchain allows only previously selected persons or groups to write to a blockchain. Permissioned blockchains do not necessarily need costly proof of work given that everyone on a particular permissioned blockchain has agreed to the other participants.

Along with the currently recognized issues with existing blockchain implementations, there are many other issues that could reasonably be expected to arise in the future. One potential issue is that of new user needs that cannot be (efficiently) accommodated within an existing blockchain. If blockchain technology proves as valuable as some observers claim, new uses are likely to arise that can be more efficiently implemented with a modified version of current blockchains. A second potential issue is that of technological changes. For example, Nathana Sharma, a faculty member at Singularity University, discusses the threat that quantum computing (a computer technology that promises far higher speeds) would pose to the cryptography on which blockchains are currently built. A third potential issue is that of changes in government regulation, such as may be necessary to comply with the European Union's General Data Protection Regulation.

Conclusion
A wide variety of uses of blockchain technology have been proposed. Whether the technology ultimately proves the best approach for many of these problems remains to be seen. However, what is clear is that for blockchain technology to reach its potential, it must overcome not only some current challenges but also be sufficiently flexible to allow changes in the implementation to meet future challenges. My next post will discuss some of the issues associated with changing blockchain implementation, including the governance around those changes.

References
Budish, Eric, 2018. "The Economic Limits of Bitcoin and the Blockchain." National Bureau of Economic Research Working Paper No. 24717.

Larry D. Wall is executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Scott Frame and Brian Robertson for helpful comments. The view expressed here are the author’s and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please email atl.nftv.mailbox@atl.frb.org.

_______________________________________

1 Erin Griffith, a senior writer at Wired, identified 187 different problems that blockchain promoters claim can be solved by using a blockchain. However, a paper by ETH Zurich doctoral students Karl Wüst and Arthur Gervais suggests that blockchains are an inefficient database solution for many uses. Some critics take an even more skeptical view of the benefits of using a blockchain relative to other database technologies, see, for example, here, here, and here.

2 The most widely known of the blockchain applications, bitcoin, has a total value that is a small fraction of major government-issued fiat currencies and a transactions volume that is a trivial fraction of that in those currencies. For example, CoinMarketCap reports the total world-wide market value of bitcoin (BTC) as $132 billion on the morning of June 3, 2018, whereas the Federal Reserve's May 24 H.6 release gives the not seasonally adjusted value of the narrow U.S. dollar money supply (M1) at $3,542 billion for the week ending May 14.

3 See my two-part series on ICOs here and here. See also the Securities and Exchange Commission's warnings about ICOs.

4 An article by the Verge reporter Adrianne Jeffries discusses some of the many widely used definitions.

5 Satoshi Nakamoto is the pen name for the person or persons who wrote the bitcoin white paper.

6 A Chicago Fed article by Rebecca Lewis, John W. McPartland, and Rajeev Ranjan provides a relatively nontechnical discussion of the blockchain technology. A somewhat more technical explanation is provided in a paper by Clemson professor Gerald Dwyer.

7 As explained by Walch, bitcoin abandoned some prior records to restore unity after the blockchain unintentionally split (or forked) into two separate ledgers. The ethereum ledger was rolled back to erase the theft of ether.

8 A well-known vulnerability of blockchains that rely exclusively on proof of work is that prior blocks could be rewritten if the attacker has a majority of the computing power applied to solving the cryptographic problem. A recent article by Bloomberg reporter Olga Kharif discusses several recent attacks have hit a number of the small cryptocurrency blockchains. The webpage crypto51.app shows the theoretical cost of attacking a large number of different cryptocurrencies using a "51 percent" attack (or majority of computing power attack) on rental equipment. Finally, Mike Orcutt, an associate editor at MIT Technology Review, discusses several other ways in which a blockchain may be attacked.

9 A Bloomberg article by Olga Kharif provides a general discussion of the General Data Protection Regulation and blockchains, and a paper by the law firm Hogan Lovells reviews the general issue of block chains and data protection in more detail.

10 See an article in MIT Technology Review for a discussion of how bitcoin users can be identified. Bitcoin.com writer Kai Sedgwick discusses two alternative blockchain implementations that seek to provide greater privacy for their cryptocurrency users.

11 My earlier post discusses smart contracts.

12 The amount of energy estimated by digieconomist to be used by bitcoin likely overstates actual usage by a factor of two, according to an article by journalist Paddy Baker that quotes work by CoinShares. However, even if the actual usage is one-half the figures commonly used, that is still a rather large number relative to other electronic payments methods.

13 The cryptocurrency with the highest possible volume in howmuch.net's comparison is ripple at 1,500 transactions per second.

14 A chapter in the Bank for International Settlements' Annual Report highlights another problem with seeking to scale a blockchain cryptocurrency based on proof of work to replace existing payments mechanisms. The size of the required files would soon exceed the capacity of most storage systems and the required communication of updates could overwhelm the internet.

15 A report by Sigrid Seibold and George Samman of KPMG discusses alternative mechanisms for authenticating and validating transactions on a blockchain. A blog post by Kyle Samani, managing partner at Multicoin Capital, provides further discussion of the trade-offs involved in making blockchains more scalable.

16 See a paper by the BitFury Group for a comparison of proof of work with proof of stake. A version of a proof of stake system has been proposed for the blockchain ethereum, and a Github page provides an FAQ discussing proof of stake from an ethereum perspective.