Environmentally Responsible Cryptocurrency - Part 3

Thinking Ahead for a Environmentally Sustainable Cryptocurrency

January 1, 2018 (originally posted on LinkedIn Publishing)

If you haven't read the first part of this series of articles, you can go the beginning here: Environmentally Responsible Cryptocurrency.

Coin Magi is designed to be profitable for miners with general purpose CPUs by using a combination of M7 based hashing in addition to a mining reward system utilizing an attraction, repulsion model. However, it is for the most part not designed to discriminate between general purpose CPU types. Since PoW is still a major factor in the reward system, it is not tuned to specifically reward energy efficiency. Granted. PoW implies work, which implies power, so designing a system that rewards power efficient work is certainly not trivial.

Perhaps a new model could incentivize energy efficiency. But this introduces an entirely new problem. How could one incentivize energy efficiency algorithmically? Does that even make sense? What we need is a system that incorporates something like Proof of Efficiency, or PoE. Since PoW is rooted in the limitations of current hardware, perhaps efficiency can be rooted in similar physical limitations.

Additionally, energy efficiency cannot be the only concern since ASIC based miners are very energy efficient per hash for many cryptocurrencies. Various cryptocurrencies optimize different variables to suit their agendas. One popular agenda is the democratic, fair, distributed calculation and integrity of the blockchain/ledger. But how might one reconcile energy efficiency with the other goals?

One approach might be to develop a system that meets the following criteria:

• Carbon neutrality for energy requirements + system purchase costs.
• Designed in an open manner. This ensures that anyone can build or improve the system. With no secrets, there is nothing circumvent or bypass, e.g. DRM, magic hardware.
• Have computational requirements that:
• Require near 100% utilization of the physical hardware resources
• Use a variety of resources that are not easily virtualizable or optimizable, e.g. relatively large amounts of memory, floating point operations, unpredictable logic branches, relatively large amounts of fast-ish persistent storage, medium bandwidth + low-ish latency network connectivity with short term high availability.
• Low power requirements.

One could design an inexpensive hardware system that provides these requirements as economically as possible. The design and mining algorithm would be optimized so that the complete system could be built and sold for $20 including the solar panel, energy storage, etc. The base mining reward would be on a per transaction basis and would simply be the cost of the system ($20) divided by the number of transactions the system could expect to handle in a year. Any further work is pure profit, but naturally limited to 100% per year of initial investment.

By limiting the reward to just above the cost of the system, one can completely deincentivize mass mining, especially the energy expensive types. The idea is, by utilizing 100% of the optimized physical resources, the comparative efficiency of running on VM's, larger platforms with virtualization or simulation becomes uneconomical due to efficiency losses in not running on native hardware.

In the end, the mining is done for the sake of a reasonable profit per device and contributing to a useful system that doesn't have a huge carbon footprint. As Moore's law slows down, individual non-parallel systems do not get much faster or efficient. This means next generation general purpose CPUs have less of an advantage compared to past improvements. As the reward system is controlled algorithimically on a more or less general purpose CPU, there is a fair amount of room for tweaking the reward or algorithm once the hardware is deployed. The implicit goal is to make the optimal way to earn money with the currency be by deploying the solar (or other "free") self powering units. This is as opposed to burning large amounts of carbon to power and cool warehouses of specialized computers profitable only at an industrial scale.

It's even easy to imagine a large swath of units cooperating in a mesh network, requiring little infrastructure and handling distributed tasks efficiently, paying for themselves, and providing a free (possibly even profitable!), reliable and decentralized service for the general public.

You can find the next article in the series here: Environmentally Responsible Cryptocurrency - Part 4