Centralized services like Facebook, Microsoft, and Yahoo keep all your eggs in one basket. This means that bad actors only need to attack one weak link to compromise all your data. And there are many ways to do it, including DDoS, Man-in-the-Middle, and credential stealing to name a few. Decentralized services keep your data safe and secure by distributing it across many redundant servers with cutting-edge cryptography that ensures your data is complete, immutable, and incorruptible. Of course, there are many degrees of decentralization. In this article, we explore the varieties of decentralization, their history and evolution, and the benefits each decentralized consensus algorithm has to offer your business, NGO, or non-profit.
DLTs: Somewhere Between Centralized and Decentralized
Distributed Ledger Technologies (DLTs) exhibit considerable tradeoffs compared to fully decentralized public ledgers like Bitcoin and centralized services like Google. For example, centralized services are designed to process high throughput, low latency computations that scale to meet increased demand. We see this in the success of centralized services like Amazon AWS, which transformed the legacy model of in-house server setups to vastly more efficient, less expensive cloud computation services. Scalability and cost savings are possible with AWS because of their private, trusted setup and the coordinated communications of highly performant centralized servers.
Distributed Ledger Technology, on the other hand, exists on the spectrum between fully centralized and fully decentralized computation. DLTs are shared distributed ledgers that often rely on permissioned setups and servers owned by a centralized authority like IBM, or consortia like Hyperledger (open-sourced) and R3 Corda (closed-source) that follow mutually coordinated operating procedures using shared protocols. The benefits of DLTs are interoperability and scalability; however, they do not maintain the same security guarantees of decentralized public ledgers like Bitcoin or Ethereum.
On the other end of the spectrum, public decentralized blockchains are designed to process computations based on permissionless distributed consensus, independent of a central authority and without relying on trusted intermediaries. As a result, public dAapps like Ethereum and Zcashtake longer to execute trustless transactions and use a lot of resources like compute cycles and electricity to achieve consensus in a secure and private manner. These tradeoffs currently limit fully decentralized use cases because they cannot scale to meet real-world demand.
The History and Evolution of Proof-of-Work (PoW)
There is a wide variety of consensus algorithms enabling blockchain and DLT functionality. The most prominent is Proof-of-Work (PoW), which establishes consensus using mining hardware to agree upon the accuracy of transactions in a decentralized and cryptographically verifiable manner. Bitcoin debuted with PoW and has proven to be a viable consensus algorithm for 10+ years. However, over time, PoW mining hardware has become increasingly cost prohibitive.
As a result of mining hardware competition and human greed, there is a growing trend towards the re-centralization of public blockchains and the decentralized applications that run on top of them. In general, public blockchains exhibit varying degrees of decentralization. Although Bitcoin was originally designed to operate outside of the bounds of centralized authorities, its Proof-of-Work (PoW) mining algorithm, in combination with specialized mining hardware developments, has led to increasingly centralized mining. It is estimated that over 70% of BTC miners operate out of mainland China. This dense geographic concentration is precarious for the resilience of the network because sudden government sanctions or loss of electricity can negatively impact node uptime and, in turn, rapidly reduce network safety and the degree of decentralization.
In 2009, Bitcoin afforded anybody with an internet-connected computer the ability to run a full node for mining, sending and receiving Bitcoin using their computer’s CPU. As Bitcoin gained value, the hardware used for mining became increasingly specialized and costly, displacing basic consumer computer setups. Bitcoin evolved from running on CPUs to more expensive GPUs, FPGAs, and, eventually, to specialized mining hardware called ASICs, Application Specific Integrated Circuits. This evolution of specialized hardware upgrades re-centralized Bitcoin so that the majority of mining rewards went to the people who could afford expensive mining rigs, not the common contributor.
What’s more, miners pooled their resources together in order to concentrate mining authority and the receive more mining rewards. As a result, what started off as a more fully decentralized application slowly devolved so that mining power and rewards were not as evenly distributed as initially intended. Concentrated mining distribution also led to unfair decision-making powers such as the decision to hard fork Bitcoin (seen in the Bitcoin Cash controversy) or the implementation of contentious Bitcoin Improvement Proposals (BIPs) provided majority node consensus.
Overall, Bitcoin and Proof-of-Work (PoW) public ledgers have not end up as decentralized as originally intended. However, private, centralized blockchains like Quorum, by the nature of their design, are vulnerable to centralized attack vectors, so their benefits of speed and throughput do not outweigh the costs of limited security. Ultimately, centralized blockchains undermine the purpose of shared public ledgers, delivering solutions with mixed results.
Performance Trade-offs Between PoS and DPoS
The question remains, how can we achieve the benefits of centralized services such as speed, scalability, and throughput while maintaining the positive attributes of privacy, security, and immutability that decentralized public ledgers provide? As we’ve seen, PoW solutions run into unintended consequences like greedy miners and design tradeoffs like limited scalability. Other mining algorithms such as Proof-of-Stake (PoS) and Delegated-Proof-of-Stake (DPoS) exhibit similar trade-offs but with different outcomes. Here’s how PoS and DPoS consensus algorithms stack up:
- PoS establishes consensus by minting blocks based on staking cryptocurrency claims for the right to validate blocks, instead of mining blocks to establish consensus with a block reward. Staking is another way of having skin in the game. Importantly, the value generation of PoS blockchains cannot be bootstrapped like PoW, where mining organically creates value via block rewards. Instead, value must be pre-established by way of ICO, STO or Airdrop. This means that the security considerations of PoS vary and can be gamed in ways that PoW cannot.
- In PoS, the authority to mint the next block is based on factors like stake — the amount of cryptocurrency collateral a participating member wagers for the right to validate blocks — and additional attributes like the duration of time that a wager is staked or the length of a staker’s hash value. Whereas PoW can run into unintended consequences like greedy miners that affect centralization and governance decisions, PoS blockchains suffer different vulnerabilities such as bad actors who can target the users wagering stake in order to change the outcome of the blocks minted. The tradeoffs for PoS are less security for increased speed and scalability.
- DPoS establishes consensus in a similar way to PoS. What differs is the method of randomization in governance. In DPoS, a select few delegates are in charge of governing blockchain parameters. They get to oversee blockchain governance decisions but do not oversee voting on transaction validation or block production. Users vote to choose witnesses, who vote on validating transactions, block production, and choosing delegates by using stake. Delegates rotate over time to ensure the health of the blockchain.
- In DPoS, security is akin to a liquid democracy, governed by the actions of voters and voting machines. Like democracies, DPoS can be gamed at the social layer and authority can be concentrated to a select few. So, the advantages of speed, latency, and throughput can be offset by bad actors in DPoS. The result is a different type of risky centralization.
Next-Gen Consensus Algorithms
Ultimately, these variations of blockchain consensus — PoW, PoS, and DPoS — all have their own trade-offs. In order to deliver real-world utility, blockchains and DLTs will need to fine-tune consensus algorithms so businesses can be confident that the privacy and security benefits which decentralized applications offer can scale alongside consumer demand. Fortunately, there are many types of consensus algorithms being developed and promising next-gen solutions.
For instance, a new consensus mechanism seen in Algorand’sPure PoS algorithm shows promise for balancing true decentralization with scalability. Algorand was invented by Turing Award-winning cryptographer, Silvio Micali, who developed Pure PoS at MIT. Pure PoS works by generating a new block from a committee “randomly chosen from the set of all users.” In Algorand, player replaceable randomization combined with secret self-selection achieves a decentralized governance model that scales on-chain, solving the trilemma of decentralization, scalability, and security. On paper, Algorand achieves the best of both worlds. Recently, Algorand’s TestNet was released to the public, so we can begin to see how Pure PoS stacks up against the competition. Given it’s wide base of testers and early adopters, I’m excited to see how Algorand pans out.
Another next-gen frontrunner is Elixxir, created by David Chaum, who invented Ecash and DigiCash, early precursors to Bitcoin. Elixxir’s full-stack blockchain enables quantum-resistant security using privacy-preserving mix networks. The primary advantage of Elixxir is its foundational focus on privacy, which hides all transaction metadata by obfuscating transactions using mix networks, homomorphic encryption and the ElGamal signature scheme.
Elixxir’s cMix consensus algorithm is unique compared to other blockchains. With cMix, public key cryptography operations are moved off-chain and calculated in the background by the mix network nodes executing precomputations in real time. In this cryptosystem, senders and receivers do not execute public key operations directly. Public key cryptography demands computationally intensive resources, which typically slow down transaction speeds. To save time and increase privacy, Elixxir nodes process precomputations of transactions across the entire mix network between transaction intervals, minimizing computational overhead for senders and receivers and lowering the latency of real-time transactions. With Elixxir, users can expect privacy, speed, scalability and security all in one high-performing blockchain.
In February, Elixxir started an application process for operating BetaNodeswhich draws to a close tomorrow on April 19th. Due to the nature of node precompuations, Elixxir has built a unique hardware solution that makes use of household GPUs, multi-core CPUs, and performant SSDs to maximize efficiency while limiting the likelihood of a hardware arms race. Because the system “only rewards minimal performance…All nodes receive similar rewards and all nodes are selected at the same frequency.” This means that Elixxir will start out fully decentralized and remain so over time, providing an optimal blockchain to run decentralized applications on for businesses, NGOs and nonprofits alike.