Bytemaster's Blog

Delegated Proof of Stake (DPOS) vs Proof of Work (POW)

Delegated Proof of Stake (DPOS) vs Proof of Work (POW)

Posted by Max Wright on .
Featured

Delegated Proof of Stake (DPOS) vs Proof of Work (POW)

Posted by Max Wright on .

Today’s post is an excerpt from “BitShares 101” talking about the benefits of Delegated Proof of Stake vs Proof of Work.


In this chapter, I am going to explain the technological leap that occurred in August of 2014 that made DACs far more viable. I should warn you that this bonus chapter may get a bit technical, so if you are not interested in the technicalities of it all, then you may want to skip over it.

However, there is good reason to become familiar with these concepts – depending on your personal opinion and risk-aversion, they just may sway your investment decisions in the future.

The first question that may pop into your mind is this: “If Bitcoin was the world’s first DAC, then why can’t other DACs just copy it?”

This is a great question, but in order to answer it, we first need to understand just how Bitcoin works.

An anonymous inventor named Satoshi Nakamoto created Bitcoin. He hashed out his ideas in a small forum with a handful of other people including Dan Larimer.

Five years later, after millions of people have studied the Bitcoin code, several have identified areas in which the Bitcoin algorithm is not as efficient as it could be.

We will explore some of these arguments and the proposed solutions.

Let’s review how Bitcoin works under the hood. The consensus algorithm used by Bitcoin is called Proof Of Work (POW).

At present, every 10 minutes, the Bitcoin protocol rewards miners with 25 newly created Bitcoin.

Let’s do some math here. We’ll use the price of $500 per Bitcoin. Although this may not be the current price of Bitcoin, it pretty close to Bitcoin’s average price in 2014 and it will keep our example nice and simple.

Every 10 minutes, $12,500 worth of Bitcoin is created and paid to the miners for their services. That translates to 1.8 million dollars per day, or $650 million dollars per year. That number is astounding. Nearly two-thirds of a billion dollars worth of Bitcoin will be created and paid to the miners in 2014 alone.

So where does that money come from? The answer is that the owners of Bitcoin, you and I, pay for it through the mechanism of inflation. When those new Bitcoins are created, Bitcoin holders either pay for it through a drop in the price of Bitcoin (miners have to liquidate their new Bitcoins to pay for mining rigs and electricity), or through the fact that the price of our Bitcoin doesn’t increase as fast as it would have if those miners keeping Bitcoin had to buy them on the open market.

Inflation is always a tricky concept to understand, so let me explain it in a slightly different way.

In 2014, $650 million worth of new Bitcoin will be created.

If less than that amount of new money buys into Bitcoin, then the price will fall. If more that than amount of money buys into Bitcoin, then the price will rise.

To demonstrate this point, let’s play a hypothetical game. Let’s say the price of Bitcoin went up one order of magnitude; that is to say, the price was not $500, but rather $5,000 per Bitcoin.

So instead of the yearly $650 million of new money that would be required to keep the price stable, $6.5 billion of new money would be required, each year, to keep the price from falling. If any less than this amount entered the Bitcoin ecosystem, the price would decline.

As you can see, as more and more people buy into Bitcoin, the price of each Bitcoin goes up. However, as this happens, the value of those 25 Bitcoins being created every 10 minutes increases simultaneously. Thus, it is as if there are reigns pulling back on the price and stopping it from continuously increasing.

If a new Bitcoin consensus algorithm could reduce the power of those reigns considerably, then the Bitcoin price would skyrocket and in my opinion would probably already be over $100,000.

If there were a way to reduce the cost of securing the network without making it less secure, this would be a tremendous victory for Bitcoin and would make future DACs a lot more viable.

So let’s take a look at the security Bitcoin receives for the $650 million per year spend.

Running the Bitcoin network at its current size actually costs very little, in the order of thousands of dollars per year. What Bitcoin holders are really trying to purchase with their $650 million per year is decentralization.

The Bitcoin network is only as secure as it is decentralized.

The more miners securing the network, the less dependent on each miner the network is. Each new miner makes the system more robust.

But what is the value for money that Bitcoin is getting for its $650 million spent yearly on decentralization?

First let’s look at the trend. As Bitcoin scales, there is a clear trend toward centralization of mining power.

Once upon a time, anybody could mine Bitcoins from home using their PC or laptop.

Soon, mining became a little bit more specialized, firstly by GPU’s, and later by specialized computers called ASICs (Application Specific Integrated Circuit), whose sole purpose is to mine Bitcoin.

These first two iterations mean that it takes thousands of dollars of hardware just to begin mining Bitcoin. This has caused some centralization of mining power already, and the nice little retail ASIC units in pretty casings you can use at your home are already fast becoming obsolete. Huge mining farms are cropping up around the world with just the bare-bones essentials.

In the very near future, most people agree that giant purpose-built, water-cooled ASIC farms will be required to make mining profitable.

Already, Bitcoin is centralized to the point that with just three mining pools, you can control 51% of the network. With just four ASIC chip manufacturers, you can control 90%+ of production of future hashing power. And as we rapidly approach giant purpose built mining centers, it is certainly conceivable that by controlling just a dozen or two of these mining centers, one could control 51% of the hashing power.

That is not a lot of decentralization for $650 million dollars.

Don’t get me wrong here. I am not saying Bitcoin that is doomed or that it is not decentralized enough. I am just stating the facts about how centralized Bitcoin hashing power is becoming.

If an algorithm were created and adopted by Bitcoin that could either deliver more decentralization or the same level of decentralization for less cost, then that could send the price of Bitcoin soaring.

Remember, too, that the Byzantine General’s problem that Satoshi solved was thought to be an impossible nut to crack. But it was done.

Getting more decentralization than four chip manufacturers, two mining pools, and a dozen mining farms for less than $650 million per year seems like a walk in the park compared to solving the Byzantine General’s problem.

Whether we have to wait one year, two years, five years, or even 10 years, I am completely certain that eventually, a vastly more efficient consensus algorithm than Proof of Work will be created.

As a side note, it is interesting to think about how Bitcoin will adapt when this new algorithm is found. One great feature of Bitcoin is that it is software and therefore can be updated at any time with community consensus.

Bitcoin is the 400-pound gorilla of crypto-currencies, with hundreds of millions of dollars of VC money pouring into it, and mind-boggling network effects. When consensus is reached that a more efficient algorithm has come along, then Bitcoin can update its code and still remain the number-one crypto. However, if Bitcoin fails to adapt, then I think it will lose market share in the long term.

So how will consensus be reached? In my opinion, getting consensus on an unproven theory will be impossible. I think this is where the world of alt-coins is so valuable.

I see alt-coins as a little laboratory with lots of little experiments running. Without risking anything to Bitcoin, we can watch plenty of experiments fail and find the odd one that succeeds.

Only after years of an alt-coin proving itself as secure and more efficient, and of course exploding in price, would the Bitcoin community entertain the idea of updating to its consensus algorithm.

That is, at least, how I imagine how this transition will take place. Here is the exciting part:

Although we are in the very early days of this technology, I don’t think we have to wait five or 10 years for that efficient algorithm to be invented. I think it was invented and released into the wild in July of 2014.

It will be very interesting to watch over the next few years.

This new consensus algorithm is called Delegated Proof of Stake (DPOS). Dan Larimer invented it.

By my calculations, if this algorithm were the backbone of the current Bitcoin network, it would produce more decentralization for less than 5% of the cost. The result would be a reduction in inflation and an increase in the price of Bitcoin.

DPOS also allows for reliably confirmed transactions every 10 seconds. With Bitcoin, this takes six confirmations at average time of 10 minutes each. 10 seconds versus one hour is a pretty significant difference.

DPOS has already been implemented, and is the backbone of Bitshares.

Bitshares launched in July of 2014.

So how does Delegated Proof of Stake work? It works by using reputation systems and frictionless, real-time voting to create a panel of limited trusted parties. These parties then have the right to create blocks to add to the Blockchain and prohibit un- trusted parties from participating. The panel of trusted parties take turns creating blocks in a randomly assigned order that changes with each iteration.

It is important to note that not much trust is required. Block creators (called delegates) can create blocks or not create blocks, include transactions or not include transactions. That is it. They cannot change transaction details like senders, recipients, or balances, so they have little power to do anything harmful. If delegates fail to create a block or include a transaction, the next delegate’s block will be twice the size or will include the missing transaction, and the confirmation time will be 20

seconds instead of 10 seconds. No serious harm done. The malicious or tardy delegates behavior is publically available and the community can vote them out quickly and easily. This would result in the delegate losing their income as delegates with no potential upside.

Built into the client is a voting system. Every owner of Bitshares votes to create a panel of 101 delegates who take turns creating blocks.

The number 101 is arbitrary. The community could allow more delegates for more decentralization, but their costs would go up. They could also reduce costs at the expense of decentralization by having less than 101 delegates. 101 is the first attempt at that sweet spot and can be changed by the community at a later time.

So the 30,000-foot view is that because the number of delegate positions is limited, delegates actually compete against one another for the job. By voluntarily lowering their salaries, delegates can attract votes, and thus, the cost of securing the network is kept in check by competition among delegates.

The degree to which it is decentralized is firmly in control of the holders of Bitshares, as they decide how many delegates can be block creators.

Meanwhile, nefarious delegates get no value from bad behavior and are voted out quickly.

Where as Bitcoin pays miners by inflating the currency supply, or said another way, diluting the Bitcoin shareholders, Bitshares can afford to pay its 101 delegates out of fees alone, and no dilution need take place. In fact, because delegates can offer to lower their pay to get more votes, the remaining fees are paid to Bitshares holders as dividends.

Something else very cool happens with this model.

Rather than simply taking a pay cut to be a delegate, delegates may get voted in because they use the delegate pay on other tasks like marketing, legal work, or lobbying. They can thus perform some of the duties that employees of other types of companies would do.

Thus, there are strong incentives for all delegates to not just secure the network, but also to provide value to shareholders in other ways.

Also, rather than the ad-hoc fashion in which Bitcoin comes to consensus, DPOS has a built in, real-time shareholder voting system. This allows the system to act like a continuous shareholder meeting where shareholders vote in or out various changes to the company’s charter. As compared with Bitcoin, voting rights in BitShares are firmly with the shareholders and not the employees.

While the Bitcoin Proof Of Work algorithm may work well for a currency, I believe the features and efficiencies of DPOS make it a lot more suitable for building DACs.

I believe the development of DPOS will unleash a number of DACs into the world in a way that would have been impossible before. Exploring just some of these opportunities is what this book has been all about.

Recommended by Bytemaster

      (see more...)

 
user

Welcome!

My name is Daniel Larimer and I am the founder of BitShares. My mission in life is to find free market solutions to secure life, liberty, and property for all.

The purpose of this blog is to help create a free society by encouraging people to join our community which is centered around Bitshares, a next generation fully decentralized crypto-currency exchange. I use Austrian Economics to engineer the economic incentives which make freedom and non-violence profitable.


Help spread the word.


Featured posts

Recent Articles