The rise of specialized hardware (particularly FPGAs) and its impact on the mining community
By Seth David on ALTCOIN MAGAZINE
Proof of Work (PoW) is one of the most commonly used consensus mechanisms entrusted to secure and validate many of today’s most successful cryptocurrencies, Bitcoin being one. Battle-hardened and having weathered the test of time, Bitcoin has demonstrated the undeniable strength and reliability of the PoW consensus model through sheer market saturation, and of course, its persistency.
In addition to the cost of powerful computing hardware, miners prove that they are benefiting the network by expending energy in the form of electricity, by solving and hashing away complex math problems on their computers, utilizing any suitable tools that they have at their disposal. The mathematics involved in securing proof of work revolve around unique algorithms, each with their own benefits and vulnerabilities, and can require different software/hardware to mine depending on the coin.
Because each block has a unique and entirely random hash, or “puzzle” to solve, the “work” has to be performed for each block individually and the difficulty of the problem can be increased as the speed at which blocks are solved increases.
Hashrates and Hardware Types
While proof of work is an effective means of securing a blockchain, it inherently promotes competition amongst miners seeking higher and higher hashrates due to the rewards earned by the node who wins the right to add the next block. In turn, these higher hash rates benefit the blockchain, providing better security when it’s a result of a well distributed/decentralized network of miners.
When Bitcoin first launched its genesis block, it was mined exclusively by CPUs. Over the years, various programmers and developers have devised newer, faster, and more energy efficient ways to generate higher hashrates; some by perfecting the software end of things, and others, when the incentives are great enough, create expensive specialized hardware such as ASICs (application-specific integrated circuit). With the express purpose of extracting every last bit of hashing power, efficiency being paramount, ASICs are stripped down, bare minimum, hardware representations of a specific coin’s algorithm.
This gives ASICS a massive advantage in terms of raw hashing power and also in terms of energy consumption against CPUs/GPUs, but with significant drawbacks of being very expensive to design/manufacture, translating to a high economic barrier for the casual miner. Due to the fact that they are virtual hardware representations of a single targeted algorithm, this means that if a project decides to fork and change algorithms suddenly, your powerful brand-new ASIC becomes a very expensive paperweight. The high costs in developing and manufacturing ASICs and the associated risks involved, make them unfit for mass adoption at this time.
Somewhere on the high end, in the vast hashrate expanse created between GPU and ASIC, sits the FPGA (field programmable gate array). FPGAs are basically ASICs that make some compromises with efficiency in order to have more flexibility, namely they are reprogrammable and often used in the “field” to test an algorithm before implementing it in an ASIC. As a precursor to the ASIC, FPGAs are somewhat similar to GPUs in their flexibility, but require advanced programming skills and, like ASICs, are expensive and still fairly uncommon.
Two Guys One ASIC
One of the issues with proof of work incentivizing the pursuit of higher hashrates is in how the network calculates block reward coinbase payouts and rewards miners based on the work that they have submitted. If a coin generated, say a block a minute, and this is a constant, then what happens if more miners jump on a network and do more work? The network cannot pay out more than 1 block reward per 1 minute, and so a difficulty mechanism is used to maintain balance. The difficulty will scale up and down in response to the overall nethash, so if many miners join the network, or extremely high hashing devices such as ASICs or FPGAs jump on, the network will respond accordingly, using the difficulty mechanism to make the problems harder, effectively giving an edge to hardware that can solve them faster, balancing the network. This not only maintains the block a minute reward but it has the added side-effect of energy requirements that scale up with network adoption.
Imagine, for example, if one miner gets on a network all alone with a CPU doing 50 MH/s and is getting all 100 coins that can possibly be paid out in a day. Then, if another miner jumps on the network with the same CPU, each miner would receive 50 coins in a day instead of 100 since they are splitting the required work evenly, despite the fact that the net electrical output has doubled along with the work. Electricity costs miner’s money and is a factor in driving up coin price along with adoption, and since more people are now mining, the coin is less centralized. Now let’s say a large corporation has found it profitable to manufacture an ASIC for this coin, knowing they will make their money back mining it or selling the units to professionals. They join the network doing 900 MH/s and will be pulling in 90 coins a day, while the two guys with their CPUs each get 5 now. Those two guys aren’t very happy, but the corporation is. Not only does this negatively affect the miners, it compromises the security of the entire network by centralizing the coin supply and hashrate, opening the doors to double spends and 51% attacks from potential malicious actors. Uncertainty of motives and questionable validity in a distributed ledger do not mix.
When technology advances in a field, it is usually applauded and welcomed with open arms, but in the world of crypto things can work quite differently. One of the glaring flaws in the current model and the advent of specialized hardware is that it’s never ending. Suppose the two men from the rather extreme example above took out a loan to get themselves that ASIC they heard about that can get them 90 coins a day? When they join the other ASIC on the network, the difficulty adjusts to keep daily payouts consistent at 100, and they will each receive only 33 coins instead of 90 since the reward is now being split three ways. Now what happens if a better ASIC is released by that corporation? Hopefully, those two guys were able to pay off their loans and sell their old ASICs before they became obsolete.
This system, as it stands now, only perpetuates a never ending hashrate arms race in which the weapons of choice are usually a combination of efficiency, economics, profitability and in some cases control.
Implications of Centralization
This brings us to another big concern with expensive specialized hardware: the risk of centralization. Because they are so expensive and inaccessible to the casual miner, ASICs and FPGAs predominantly remain limited to a select few. Centralization occurs when one small group or a single entity controls the vast majority hash power and, as a result, coin supply and is able to exert its influence to manipulate the market or in some cases, the network itself (usually the case of dishonest nodes or bad actors).
This is entirely antithetical of what cryptocurrency was born of, and since its inception many concerted efforts have been made to avoid centralization at all costs. An entity in control of a centralized coin would have the power to manipulate the price, and having a centralized hashrate would enable them to affect network usability, reliability, and even perform double spends leading to the demise of a coin, among other things.
The world of crypto is a strange new place, with rapidly growing advancements across many fields, economies, and boarders, leaving plenty of room for improvement; while it may feel like a never-ending game of catch up, there are many talented developers and programmers working around the clock to bring us all more sustainable solutions.
The Rise of FPGAs
With the recent implementation of the commonly used coding language C++, and due to their overall flexibility, FPGAs are becoming somewhat more common, especially in larger farms and in industrial setting; but they still remain primarily out of the hands of most mining enthusiasts and almost unheard of to the average hobby miner. Things appear to be changing though, one example of which I’ll discuss below, and it is thought by some, that soon we will see a day when mining with a CPU or GPU just won’t cut it any longer, and the market will be dominated by FPGAs and specialized ASICs, bringing with them efficiency gains for proof of work, while also carelessly leading us all towards the next round of spending.
A real-world example of the effect specialized hardware has had on the crypto-community was recently discovered involving a fairly new project called Verus Coin (https://veruscoin.io/) and a fairly new, relatively more economically accessible FPGA. The FPGA is designed to target specific alt-coins whose algo’s do not require RAM overhead. It was discovered the company had released a new algorithm, kept secret from the public, which could effectively mine Verus at 20x the speed of GPUs, which were the next fastest hardware types mining on the Verus network.
Unfortunately this was done with a deliberately secret approach, calling the Verus algorithm “Algo1” and encouraging owners of the FPGA to never speak of the algorithm in public channels, admonishing a user when they did let the cat out of the bag. The problem with this business model is that it is parasitic in nature. In an ecosystem where advancements can benefit the entire crypto community, this sort of secret mining approach also does not support the philosophies set forth by the Bitcoin or subsequent open source and decentralization movements.
Although this was not done in the spirit of open source, it does hint to an important step in hardware innovation where we could see more efficient specialized systems within reach of the casual miner. The FPGA requires unique sets of data called a bitstream in order to be able to recognize each individual coin’s algorithm and mine them. Because it’s reprogrammable, with the support of a strong development team creating such bitstreams, the miner doesn’t end up with a brick if an algorithm changes.
All is not lost thanks to.. um.. Technology?
Shortly after discovering FPGAs on the network, the Verus developers quickly designed, tested, and implemented a new, much more complex and improved algorithm via a fork that enabled Verus to transition smoothly from VerusHash 1.0 to VerusHash 2.0 at block 310,000. Since the fork, VerusHash 2.0 has demonstrated it's ability to do exactly what it was designed for- equalizing hardware performance relative to the device being used while enabling CPUs (the most widely available “ASICs”) to mine side by side with GPUs, at a profit and it appears this will also apply to other specialized hardware. This is something no other project has been able to do until now. Rather than pursue the folly of so many other projects before it- attempting to be “ASIC proof”, Verus effectively achieved and presents to the world an entirely new model of “hardware homogeny”. As the late, great, Bruce Lee once said- “Don’t get set into one form, adapt it and build your own, and let it grow, be like water.”
In the design of VerusHash 2.0, Verus has shown it doesn’t resist progress like so many other new algorithms try to do, it embraces change and adapts to it in the way that water becomes whatever vessel it inhabits. This new approach- an industry first- could very well become an industry standard and in doing so, would usher in a new age for proof of work based coins. VerusHash 2.0 has the potential to correct the single largest design flaw in the proof of work consensus mechanism- the ever expanding monetary and energy requirements that have plagued PoW based projects since the inception of the consensus mechanism. Verus also solves another major issue of coin and net hash centralization by enabling legitimate CPU mining, offering greater coin and hashrate distribution.
If other projects adopt Verus’ new algorithm- VerusHash 2.0, it could mean an end to all the doom and gloom predictions that CPU and GPU mining are dead, offering a much needed reprieve and an alternative to miners who have been faced with the difficult decision of either pulling the plug and shutting down shop or breaking down their rigs to sell off parts and buy new, more expensive hardware…and in so doing presents an overall unprecedented level of decentralization not seen in cryptocurrency.
Technological advancements led us to the world of secure digital currencies and the progress being made with hardware efficiencies is indisputably beneficial to us all. ASICs and FPGAs aren’t inherently bad, and there are ways in which they could be made more affordable and available for mass distribution. More than anything, it is important that we work together as communities to find solutions that can benefit us all for the long term.
In an ever changing world where it may be easy to lose sight of the real accomplishments that brought us to this point one thing is certain, VerusHash 2.0 is a shining beacon of hope and a lasting testament to the project’s unwavering dedication to it’s vision of a better world- not just for the world of crypto but for each and every one of us. ~S. David
Before moving on, make sure to press follow, leave a clap or 46, share today’s highlight and if you missed the last article, click here.
Read about the Altcoin Magazine Mastermind Event here.
Follow us on Twitter, InvestFeed, Facebook, Instagram, LinkedIn, and join our Discord and Telegram.
The purpose of ALTCOIN MAGAZINE is to educate the world on crypto and to bring it to the hands and the minds of the masses.
