MIT Technology Review: Can Photonic Chips "Save" Bitcoin?

Source: MIT Technology Review

Compilation: First Class (First.VIP)

Cryptocurrencies are known for their thirst for energy, especially blockchain (Bitcoin) based on proof of work. The problem is that the price of Bitcoin is difficult to cover the cost of energy consumption, making mining concentrated in certain countries and regions, which will make it easier for regulators to intervene in the Bitcoin blockchain.

So some researchers say that the "outlet" for cryptocurrencies lies in a more energy-efficient way of computing. Is that right?

Photonic Chips is an emerging technology included in arXiv. MIT has reported on this novel "photonic" chip this year, which uses light instead of electricity and consumes relatively little power in the process. When dealing with large-scale neural networks, the chip is millions of times more efficient than today's classic computers.

When the Bitcoin bubble burst in late 2017, the price of this cryptocurrency fell from $ 17,000 to less than $ 7,000 in just a few days. Overwhelming news said that the cryptocurrency boom ended abruptly.

But as Bitcoin's value fell, strange things happened. Bitcoin mining speed has increased dramatically. What is the reason? Although Bitcoin's value has fallen, its mining is still very profitable. In other words, the cost of mining-the price of hardware plus the price of energy to run it-is still lower than the value of the cryptocurrency it produces.

The mining boom has lasted for almost a year. Then in November 2018, the value of Bitcoin dropped again sharply, from about $ 6,500 to less than $ 3,500.

This ruined the work of many miners. Suddenly, Bitcoin no longer had enough value to cover mining costs, and many mines were forced to close. This is the first time in the history of cryptocurrencies that mining hashrate has dropped significantly, from a hash rate of 60 EH / s to 35 EH / s.

Of course, the impact is significant. Prior to this, mining nodes were distributed in every corner of the world, making it impossible for a single country or region to exert improper influence on the Bitcoin blockchain. Now, mining is only possible where energy is cheap enough and profitable, mainly in western China. And China is stepping up its review of cryptocurrencies, closing exchanges and banning various activities.

This poses a threat to the survival of Bitcoin. Since then, cryptocurrency experts have struggled to find solutions.

The fundamental problem is that Bitcoin mining, the computational cost of collision hashing, is expensive, which was set at the beginning of the blockchain design to ensure on-chain security. However, mining calculations are energy-intensive, and as more and more people join mining, energy consumption increases dramatically.

It is estimated that bitcoin mining currently consumes more than 75 terawatt hours per year (first-class warehouse note: terawatt hours is TWH, 1 terawatt hour is equal to 1 billion kWh) , which exceeds the total power consumption of Austria. This is unsustainable. If Bitcoin mining is to survive, a new mining method is urgently needed.

Michael Dubrovsky of the non-profit organization PoWx, Marshall Ball of Columbia University in New York, and Bogdan Penkovsk of Paris Saclay University in France jointly proposed a new method of "protecting" Bitcoin, which is computationally expensive but more energy efficient . They say it is crucial that it is also compatible with the current crypto system, so it should also be compatible in future Bitcoin iterations.

Instead of using traditional computers to crash the hash, they proposed to use optical computers. They believe that optical computers significantly reduce their reliance on energy, which will fundamentally solve the energy bottleneck of Bitcoin mining. Are they right?

The Bitcoin blockchain is a decentralized ledger that records all transactions related to this currency. This ensures that no separate entity controls the currency. The point is that the ledger must be secure so that everyone can trust its content. This security is achieved by periodically encrypting the ledger so that the contents of the ledger cannot be changed.

However, the encryption process must have special attributes. The encryption process of the ledger must be complicated, but once encrypted, it must be easy to check. It turns out that there is a set of mathematical objects called trapdoor functions that have exactly this property. In fact, they have been widely used to encrypt everything from personal information to credit card transactions.

Small science: One-way trapdoor function contains two obvious characteristics: one is unidirectional, and the other is trapdoor. The so-called unidirectionality is also called irreversibility, that is, for a function y = f (x), it is easy to calculate y if x is known, but x = f ^ (-1) (y) if y is known It is difficult. One-way functions are named because they can be calculated in only one direction. The so-called trapdoor is also called the backdoor. For a one-way function, if there is a z such that knowing z, you can easily calculate x = f ^ (-1) (y), without knowing z, you cannot calculate x = f ^ (-1) (y) , The function y = f (x) is called a one-way trapdoor function, and z is called a trapdoor. (Baidu Encyclopedia)

This form of encryption is computationally expensive–it requires expensive, powerful computers to run. Therefore, Bitcoin also has a key feature of success. Anyone who performs the cryptographic process (i.e., a miner who participates in proof-of-work PoW mining) will receive new Bitcoin as a reward.

That's why the process is called "mining." As the value of Bitcoin increases, so does the popularity of mining. But computing is energy intensive. Therefore, miners have been looking for ways to reduce costs : One development is the introduction of application-specific integrated circuits-ASIC chips-optimized for the purpose of bitcoin mining. The other is looking for cheap energy.

The optical computing mentioned by Dubrovksy and his colleagues above will break the deadlock to some extent. The inspiration comes from the rapid development of photonic chips in recent years. The computing efficiency of these photonic chips is much higher than that of silicon chips. They stated:

"This technology promises to provide 2-3 orders of magnitude energy efficiency optimization compared to electronic processors."

To this end, the team proposed an improved encryption protocol called HeavyHash , which is optimized for photonic chips . This means that the best results can only be obtained with digital calculations using a photon processor, and the mining process becomes an “optical proof of work”.

This "optical proof of work" encourages the use of photonic chips, which greatly reduces the original energy cost of Bitcoin. The researchers said:

"The implementation of optical proof-of-work will help accelerate the development of energy-efficient photonic co-processors."

When energy costs are no longer a major consideration, hardware costs will dominate the calculations. This will ensure that miners can make a profit everywhere, not just in areas where energy is cheap.

At least, it works in theory. The problem is that the power efficiency of photonic chips is not yet clear. For example, optical switches work by changing the refractive index, which is currently done with small heaters. The silicon photonic circuit also changes very little and must be compensated with a miniature heater.

But these heaters have greatly increased the chip's power budget in unexpected ways. In fact, Dubrovksy and his colleagues did not explicitly measure the energy saving effect that may (or may not) be realized as the scale of Bitcoin increases, so it is difficult to evaluate the effectiveness of what they call "optical proof of work" .

The researchers also did not show how the “optical proof of work” would address issues related to regional differences in electricity costs. In the future, the hardware cost for all miners will be the same as it is now. Therefore, in the long run, the best way to maximize profits is still to find cheap energy.

This is no different from the problems currently facing Bitcoin. That's why it's hard not to come to the conclusion that this form of energy-saving calculations only delays the inevitable results.

Reprinted please retain copyright information, thanks for reading.