Researchers Argue Radical Redesigns Needed to Scale Decentralized Blockchains

A new research paper has put forth the argument that fundamental changes are needed to the design of decentralized blockchain networks should any iteration of the technology serve as a transaction network for global consumers.

Entitled "On Scaling Decentralized Blockchains", the paper was authored by researchers at Berkeley, Cornell, ETH Zurich University, the National University of Singapore and the University of Maryland, including Christian Decker, co-author of the Duplex Micropayments Channels report; and Ittay Eyal, Adem Efe Gencer and Emin Gün Sirer, authors of the Bitcoin-NG proposal.

Though different in intent, both papers were focused on presenting new ideas for how decentralized blockchain networks could scale at a time when the conversation was dominating industry conversation both among major financial institutions, who are seeking to leverage the technology for existing high-volume transaction systems, and bitcoin developers, who are working on solutions to scale the digital currency's transaction capacity.

The increasing popularity of bitcoin as a digital currency has made scalability a "primary and urgent concern" for the bitcoin network, the authors say, touching upon a topic that has been hotly debated for months in the bitcoin space.

Overall, the position paper analyses how "fundamental and circumstantial bottlenecks" in decentralized, public blockchains like bitcoin limit the ability of its current peer-to-peer network of miners to support substantially higher numbers of transactions at lower latencies.

Likewise, the paper is not concerned with an analysis of permissioned networks, in which a group of entities or entity manages a blockchain-based ledger system in a more centralized manner.

The group's new analysis will be presented at the Financial Cryptography and Data Security 2016 conference, which is being held from 22nd to 26th February in Barbados.

Inherent problems

The statements in the paper echo previous comments from some of the team members, who have been critical about scaling issues they believe are inherent in bitcoin's current design and short-term improvement proposals.

Back in September 2015, some of the researchers spoke to CoinDesk about the "inherent problems" in blockchain design.

Ittay Eyal, for example, said then his work on Bitcoin-NG arose over fundamental issues with the design of blockchains that will make scaling any implementation, public or private, a challenge.

He told CoinDesk at the time:

"For securities markets, for transacting digital assets, if you want to have all of these on a blockchain, you will need significant scaling."

Notably, in the same month, Eyal also made an appearance at the Scaling Bitcoin event in Montreal, where he introduced Bitcoin-NG as a kind of radical attempt to scale decentralized blockchain networks to match the ambitious of those excited about the technology.

Scientific approach

Overall, the paper looks at whether decentralized blockchains can be scaled up to match the performance of a mainstream payment processors, and how they might achieve such functionality.

The team went on to make three contributions that illuminate the problem of scaling bitcoin and blockchains generally, and offer ways to achieve high-performance, decentralized systems:

• 'Measurement study and exploration of reparametrization' – the authors attempt to quantify current scalability limits via experimental measurements covering a range of metrics that characterize the resource costs and performance of the bitcoin network.
• 'Painting a broad design space for scalable blockchains' – concluding that "fundamental protocol redesign is needed for blockchains to scale significantly while staying decentralized", the authors suggest various technical approaches are discussed that could help blockchains scale in the future.
• 'Posing open challenges' – the authors call for better measurement techniques so the health of decentralized systems like bitcoin could be continuously monitored, and further look at the security and robustness of the network following various parameter changes.

Block size is just first step

Supported by their measurement studies, the authors suggest that "more aggressive" scaling will, in the longer term, require fundamental redesign to bitcoin's protocol.

They go on to suggest a variety of "potentially successful" approaches to such scaling, categorizing some recently proposed and new ideas, and posing a number of open technical challenges for the community going forward.

In conclusion, the paper suggests that the process of reworking the parameters of block size and transaction intervals is only a first step toward achieving "next-generation, high-load blockchain protocols", and that for bitcoin to make major advances it will require a "basic rethinking" of the underlying code.

Pete Rizzo contributed reporting.

Image via Shutterstock