Table of Contents

Why do new blockchain projects prefer PoS over PoW?

Table of Contents

If you have been following the trend of blockchain technology, you will notice that most of the new public blockchain projects are using PoS as their consensus mechanism: Cosmos, Cardano, Dfinity, Algorand, etc. Not to mention that Ethereum is turning from PoW to PoS.

So what is the advantage of using PoS over PoW?

Energy efficiency

The most obvious reason is energy efficiency.

PoW is extremely energy inefficient. In fact, this is built inside the nature of PoW: it provides incentives for miners to increase their computation power to compete with others, therefore the energy consumption of the whole network must rise over time.

Take a look at some statistics. At the time of writing, Bitcoin is consuming 73.04 TWh per year, which is 0.33% of the world’s electricity consumption. Each Bitcoin transaction on average consumes
916 KWh, which could power 30.95 U.S. households for a whole day!

(If we think about this fact, the transaction fee is actually quite cheap compared to the energy consumed. We can claim that the “actual cost per energy consumption” for Bitcoin transactions is much lower than for Visa, though this measure is quite meaningless)

Better security for projects in beginning stage

PoW relies on the assumption that the computation power of a malicious user cannot compete with the computation power of the rest of the network. This assumption makes sense for matured projects like Bitcoin, but not quite true for new projects.

A while ago, Bitcoin Gold, which is a fork of Bitcoin, was being 51%-attacked, which caused exchanges losing at least 18 million USD.

There is a website showing the cost and chance of 51%-attacking different blockchains. For some blockchains, it only costs as cheap as a few hundred (or even dozen) USD for renting computation powers to attack the network for an hour. For a newly started project, which almost nobody is mining on, how vulnerable will it be?

To prevent attacks, some projects relying on PoW protects the chain by extra mechanism. For example, IOTA, the “blockchain without blocks or chains”, is a project which utilizes the transactions (not blocks) confirming other transactions, to make the network scalable when the number of transactions increases. However, since the confirmation (and therefore the security of the whole network) relies on transactions doing PoW, it is possible for attackers with huge computation power to perform double spending. Hence, IOTA is currently operating with a role called Coordinators which are run by the IOTA foundation to set up checkpoints and prevent double spending. See the IOTA FAQ for more details.


As you can see, blockchain projects relying on PoW are vulnerable in their starting stage. What about PoS?

PoS relies on stake instead of computation power for consensus judgement. An important property about stake is that unlike computation power, stake is something “inside” the blockchain. This means it is easier to design mechanisms to lower the risk of double spending attack. For instance, the project team can control the distribution of the coins, making it difficult for attackers to get a large number of stake.

Also, in PoS, the blockchain can introduce more advanced features than in PoW, and utilize these features to disincentivize attacks. For instance, the finality feature makes double spending impossible, since the other nodes on the network knows that once a block is final, it won’t be reverted. Even if the attacker owns a huge amount of stake and takes control of the network, he / she could only attack the network by filtering transactions or shutting down the network, but not making inconsistent state like double spending.

Decentralization

Computation power, which PoW uses to determine which block is valid, has a great degree of economy of scale. That means the more computation power you own, the cheaper each unit of computation power will be.

If your mining farm is large, you can build it near power stations to reduce the cost of electric power transmission. You can choose a power station with lower power cost, say a hydroelectric power plant. You can even design an ASIC chip specifically for the coin you mine. The cost of designing the chip will be spread among the large scale of the mining farm.

The economy of scale encourages people to build or invest into mining farms instead of running mining node themselves. This hurts decentralization.

In PoS, the need for computation power is greatly decreased. The main resource needed is stake, which has much less effects of economy of scale compared to computation power. You may get some advantage for buying a larger amount of stake, but definitely much less than the advantage in the case of electric power and computation power.

New features

PoS allows blockchain projects to have new features.

For example, since the probabilistic PoW is gone, the blockchain can now offer finality, i.e. once the block is confirmed by validators, it will never be reverted. This allows outsiders to validate blockchain data easily, which brings possibilities for features like sharding, cross-chain communication and more light-weight clients.

Another direction is to allow the blockchain itself to determine the next block producers. Unlike computation power, stake is an attribute inside the blockchain, which means that the blockchain knows about it and can make decisions according to it. Algorand is a blockchain project which utilizes this property, together with techniques like verifiable random functions, to sample the next validator set right inside the blockchain logic. These show that PoS is an exciting experiment field which could bring many new features into blockchain technology.

Conclusion

PoS is bringing higher level of security, decentralization and more features to blockchain technology. As PoS projects becoming more mature, blockchain technology will be more usable, finally turns into an essential part in human life.

Related links:Introduction to LikeChain | LikeCoin transfer delegation | How to write a DApp