Most conversations on the topic of blockchain and computing gravitate towards the astronomical costs of mining bitcoin and other major cryptocurrencies. Indeed, the recent doubling of the bitcoin hash rate means that any price below $7,300 will result in a net-loss for miners who “mint” new coins. This is a serious issue though it is larger in scope than it may appear at first sight.
While bitcoin is still a niche application, the absurd computing cost involved in mining it highlights a very real problem for hardware. Although software continues to take quantum leaps, hardware design is bounded by very real limitations that make it harder to keep the pace. As a result, the only way to increase computing power is to physically add more components to do the work—an expensive and cost-ineffective strategy. This is true across several fields, from standard computational capacity to rendering and even neural networks and AI.
The blockchain industry, fully cognizant of these problems, has come up with a clever solution. The distributed ledger’s architecture is based on the power of collective computing—multiple nodes maintaining the network in collaboration with one another. Several companies have used this model to create their own decentralized super computers, a development that could unleash serious computing power and change the way several industries operate.
The Shackles that Hold Computing Back.
Currently, the biggest cost of mining bitcoin is tied into the hardware and physical costs involved. As hashing becomes progressively more difficult, hardware can only process data so fast, and the only viable solution is to add more of it.
These components are not cheap, and the result is a highly stratified mining sector restricted to those who can afford the expensive machinery. Similarly, the resource costs of mining bitcoin are becoming astronomical. The bitcoin network consumes roughly 2.25 gigawatts a year, almost as much as Ireland and several other small countries.
The problem is similar in other areas of computing as well. Graphics rendering, which is also completely reliant on hardware, sees costs increase exponentially when the tasks become major projects such as movies. Even neural networks are limited by the hardware that houses them. As these networks become more complex and thus larger, hardware has struggled to keep up, effectively limiting how much they can develop.
Even cloud computing, which claims to alleviate many hardware problems users face, is not immune from the restrictions that hardware places on it. The infrastructure to maintain towering racks of servers is expensive and has vulnerabilities that make it less than optimal.
Decentralizing Processing Power
With few avenues left to continue closing the hardware-software gap, the solution may come not from hardware, but from new software and network architectures. Blockchain’s decentralized networks are built on the concept of shared processing power. The technology uses a peer-to-peer network that reduces the need for costly infrastructure and expansive needs for power.
To simulate and improve on hardware-based computing centers (such as rendering farms, cloud servers, and more) blockchains mobilize the surplus processing, storage, and network capacity of their many nodes to create a massive pool of power and utility. Unlike the hardware-based model which requires adding more physical components, blockchains can simply add more nodes, which have significantly lower costs.
This model can be applied to several sectors of computing outside of simply creating a faster PC. In terms of raw computation—such as a “supercomputer”—blockchains provide a source of processing capacity that takes small increments from nodes on their networks.
Tatau, a B2B marketplace for computing power, built a system that matches suppliers with businesses looking to unlock further processing capacity. The company’s platform uses idle GPU computing resources available from its network’s nodes and puts them on an open market where companies can rent as much computing power as they need on demand, with a strong focus on AI-related processing. Tatau also handles parallel tasks, so computations can occur simultaneously.
Simplifying the Computational Equation
Companies like Tatau seem to be asking a poignant question. why increase hardware costs and force users to build highly isolated processing centers, when they can leverage collective, idle power to improve computing? The result is a model that provides better computing capacity while removing many of the barriers to advancement that hardware-based systems impose.
Blockchain-based computing tools are built on the strength of their networks, which is both a negative and a positive. On one hand, they are entirely reliant on the number of users connecting and collaborating, so they could theoretically fail before they reach critical mass. On the other, they provide a truly viable model to unlock the real potential of computing power to continue fueling technological innovation. The first wave of these platforms is only now becoming viable, but progress achieved thus far bodes well for the future of this blockchain sector.