Comparison of Heterogeneous Blockchain Networks

Interchain Economic Security Topology of Cosmos, Polkadot, and Avalanche

With performance, usability, and energy-efficiency issues emerging from first-generation cryptocurrency networks such as Bitcoin and Ethereum, the vision of an open decentralized web has been clouded. In order to solve the current performance problems, Ethereum has launched a new version and corresponding L2 solutions. More importantly, the new generation of blockchain projects Cosmos, Polkadot, and Avalanche have been launched one after another, and excellent infrastructure has been established. These projects aim to achieve horizontal expansion through an asynchronous heterogeneous network model, that is, the dedicated blockchains of each app can coexist and interoperate when needed. In order to ensure the economic security between the chains, these networks have their own advantages in design, and have made their own trade-offs and trade-offs, which also have different impacts, which will be discussed in detail later. The goal of these networks is to build a blockchain internet that can accommodate millions of daily active users instead of today's hundreds of thousands of daily active users, and realize web3's vision of "the Internet is owned and controlled by users". This article hopes to help developers, researchers, entrepreneurs, investors, and everyone looking forward to the advent of a decentralized world understand this paradigm shift in cryptocurrency networks.

Bitcoin has opened Pandora's box and has gradually gained the status of "digital gold", which is the consensus of today's era. Ethereum ushered in the era of programmable internet money and became the home of cryptoeconomic innovation. But bitcoin, ethereum, and their variants have many obstacles to mass adoption. This article will first explore these barriers, and then compare the new generation of blockchain platforms based on the main points.

Energy Efficiency : The proper functioning of an open decentralized computer network requires independent participants to agree on a shared state. At the same time, the network needs to be able to maintain fault tolerance and efficient consensus in the presence of incomplete information or malicious nodes (Byzantine fault tolerance). On the one hand, the network needs to remain open, allowing more nodes to participate in consensus, and on the other hand, the network needs to prevent the same entity from operating multiple identities (Sybil attacks) - these are achieved through a method called Proof of Work (PoW; developed in 1992 Cynthia Dwork invented, originally used to prevent spam) access method implemented. PoW requires nodes to use a lot of computing power, which will exacerbate global warming and lead to high electricity bills. Clearly, there is an economic cost to maintaining the security of a decentralized computing network [1]. The new generation of blockchain projects replaces PoW with Proof of Stake (PoS) as the entry threshold for verification nodes, which requires network participants to deposit and lock tokens. In order to prevent malicious behavior and node offline, this economic threshold needs to be high enough. In fact, PoW and PoS apply the same principle of economies of scale: the cost of running a validator changes from operational expenditure (OPEX) to capital expenditure (CAPEX).

Transaction transparency : Bitcoin, Ethereum, and their variants all use the Nakamoto consensus, and the sent transaction cannot enter an irreversible state until several blocks are created. Therefore, such blockchains are highly available but slow because they use probabilistic finality and need to wait until the blockchain is long enough. In order to speed up the confirmation, many blockchain projects use the classic Practical Byzantine Fault Tolerance (PBFT) consensus, which brings other problems, such as the scale of nodes may reduce the speed of the network, causing the network to prioritize security rather than online time and activity.

Computational throughput : Throughput is the computational workload that a distributed computer network can complete per second, which determines the capacity of the network to expand. However, the commonly used unit of flux, "transactions/second", is misleading, because "transactions" can be simple transfers or complex financial calculations, and their requirements for computing power are different. The throughput is provided by the nodes, and the actual throughput of the network refers to the computing workload that the network can handle per second. There are two ways to improve throughput. One is the vertical expansion strategy, which requires nodes to have high computing performance and node software to be optimized; the second is the horizontal expansion strategy, which divides the network into multiple parts and processes transactions in parallel.

Transaction costs : Blockchains must limit the number of executions, otherwise the nodes running the blockchain are vulnerable to DoS attacks. For this reason, Bitcoin only supports a small number of scripting languages, and Ethereum charges transaction fees based on gas metering executed by smart contracts. The problem is that whether your transaction is a simple transfer or a complex calculation, they are all performed by the same network. Therefore, when network traffic increases, the gas fee for simple transactions will also increase, and only those with deep pockets can afford it. Fees are paid to miners as incentives for priority transactions. On the Bitcoin network, after the Bitcoin circulation reaches the upper limit of 21 million, the service fee will become the only incentive, while in the Ethereum, the service fee is completely used for priority transactions (technical review note: after the Ethereum 1559 protocol upgrade, The handling fees are also recovered and destroyed, and only the extra tip "tip" added by the user is owned by the node). The new generation of blockchain projects more often adopts the mechanism of destroying handling fees. Recently, Ethereum has also begun to burn part of the fees. This way, as network activity increases, token scarcity increases, which benefits all token holders.

Decentralization level : Contrary to most people's imagination, due to the concentration of mining pools (as of November 2021, 90% of Bitcoin's computing power is controlled by 11 mining pools, and 90% of Ethereum's computing power is Controlled by 16 mining pools), the level of decentralization in Bitcoin and Ethereum is actually very low. In the Nakamoto Consensus, as mining costs increase, the difficulty of block generation increases, which will further lead to the concentration of computing power. Faced with this problem, the new generation of blockchain projects have shown their capabilities, which will be discussed in detail below.

Fair distribution : How should blockchain projects distribute ownership shares (tokens) as the network grows? Bitcoin’s token distribution model establishes the interdependence of blockchain security, mining, and exchange rates. It has become a template for many projects: miners join the network, earn token income, and the network becomes more decentralized and secure, which in turn attracts more users. The increase in demand and the rise in currency prices will attract more miners to join the network and maintain network security. However, with the increase of mining costs, the difficulty of producing blocks is also increasing. This leads to a concentration of coins and computing power, creating a situation where miners are run by a few entities. Unlike Bitcoin, Ethereum's strategy is to pre-mine tokens, remove the cap on issuance, sell some tokens through private sales and crowdfunding, and allocate some tokens to foundations for development grants and bugs Offer rewards, and distribute incentives to miners like Bitcoin. Soon, Ethereum’s tokens were also concentrated in a few mining pools, and exchanges became the largest token holders. Ultimately, over time, fair distribution will determine who has power in the network, including the power to produce blocks (initiate, accept, or review transactions), the power to fork the network, the power to participate in protocol upgrade decisions, and the power to The power of the app to invest and pledge.

Governance : Changes to the network protocol can have significant impacts on all current and future users, whether they are aware of the changes or not. In Bitcoin and Ethereum, the core expert community will discuss, make decisions, implement and execute proposals, thereby upgrading the protocol and adjusting parameters. If a certain group of miners pursues something different from the majority, they can fork the protocol and launch a new network, at the expense of not enjoying the previous network effects. In addition, they usually have a central foundation that manages the distribution of R&D funds, as an alternative to a DAO (Distributed Autonomous Organization) that coordinates the funds. Most token holders and users have a very limited say in governance decisions as they may not have the expertise, interest and awareness in the relevant domain. Even if they had this information, they would still have little say compared to those with large token holdings, since votes are often weighted by token holdings. The new generation of blockchain projects will have fairer on-chain governance (including secondary voting, time-locked voting, adaptive voting bias, voting delegation, one-person-one-vote based on decentralized identity authentication) and off-chain governance (Forum signature voting) mechanism combined to allow token holders to participate in governance generally.

These issues will not only constrain the mainstream adoption of decentralized networks, but also cause existing users to continue to rely on centralized exchanges and custodial wallets. It is difficult for non-technical users to regularly use a truly decentralized app. On the other hand, ordinary users did not leave Ethereum and Bitcoin because they did not understand these issues; businesses and investors did not leave these networks because they followed the liquidity; early users and OGs maintained these networks, then It's because of the stakes. However, other possibilities exist for blockchain networks.

Ethereum daily active address. Data source: Etherscan.io

Currently, Ethereum has 500,000 daily active addresses. For reference, Twitter has 200 million daily active users (400 times that of Ethereum), and Facebook has nearly 2 billion daily active users (4000 times that of Ethereum). Even if you add up all the users of the L2 platform and Bitcoin, it is much worse than these mainstream applications. Scaling is a key bottleneck for an open decentralized Internet. This is not a problem we will face in the future, but an urgent problem that needs to be solved here and now.

In order to solve the expansion problem, Ethereum has also launched a new version, trying to cope with the growing demand through L2 solutions. At the same time, the next-generation blockchain platforms Cosmos, Polkadot, and Avalanche, which will launch their mainnets in 2019 and 2020, let us once again see the hope of a truly decentralized Internet. Let's first take a look at how the new version of Ethereum does it.

The new version of Ethereum: EVM ecology

Since its launch, the new version of Ethereum has been adopting various new mechanisms with reference to new research and the practice of a new generation of blockchain platforms. The new version of Ethereum uses PoS, splitting the network into synchronized shards in hopes of increasing aggregate computational throughput. Validation nodes running the same Ethereum Virtual Machine (EVM) will be assigned to different network shards, they will generate blocks, accumulate different user activity data, and then synchronize with each other through Beacon, the beacon chain. However, synchronizing all shards implies full replication, i.e. all nodes store the same data. This is problematic because the purpose of sharding is to scale, not to replicate all data across the network. In the synchronous model or in the heterogeneous network topology model, if the usage of one shard (for example, a very popular DeFi shard) is much higher than other shards, it will produce the same speed as today's Ethereum, Cost and scaling issues. How to efficiently synchronize data between shards is also a problem.

Although Ethereum stated that it will take about a year to transition to the new version, facing the increase in user demand, L2 solutions such as rollup (Optimistic, zkSync), plasma, and state channels have been launched one after another in order to improve efficiency and speed. The problem is that L2’s trust model needs to use a central node as an intermediary, or use multiple incentivized nodes (Polygon is built using Tendermint consensus and runs on multiple validator nodes, Matter Labs hopes to build a network of validator nodes in zkSync), the former will destroy Decentralization and censorship resistance, the latter is equivalent to creating a new decentralized blockchain with its own token (such as MATIC), which will eventually join the competition of the L1 platform. Therefore, sooner or later, these single-chain infrastructures will face the same transaction cost problem as the number of users increases.

Modular Blockchain Design

Recently, Ethereum launched a new strategy of "rollup center roadmap", that is, Ethereum is the data availability layer (L1), and other L2 projects are the computing layer. In other words, Ethereum hopes to serve as the base layer to ensure data availability and shared security for rollup. Therefore, Ethereum is actively integrating EVM chains as computing power. These EVM chains may be dominated by a single rollup, or multiple rollups may coexist (see Vitalik Buterin's Endgame article). In fact, this strategy coincides with emerging modular blockchain designs, where blockchains can outsource data availability or execution to other blockchains. A general model of this strategy was developed by Celestia and EigenLayr. Additionally, Ethereum’s new strategy is similar to the existing shared security models of Polkadot and Avalanche.

On the other hand, since Cosmos, Polkadot, and Avalanche have all deployed Ethereum cross-chain bridges on at least one EVM-compatible chain, they are also sometimes considered L2 platforms. However, these projects often refer to themselves as L0 platforms because they provide the infrastructure to develop interconnected L1 blockchains.

Cosmos, Polkadot, and Avalanche

Cosmos, Polkadot, and Avalanche all aim to scale horizontally through an asynchronous heterogeneous network model. In these three networks, app-specific blockchains have different virtual machines that can interoperate when needed. On these infrastructure platforms, users can build their own personalized blockchain, which provides greater design space for decentralized apps and assets. There are three major advantages to running a project on an autonomous blockchain rather than a set of smart contracts:

1. Performance Isolation : Isolate your blockchain from other blockchains to ensure your user experience is not affected by irrelevant network activities, thereby improving blockchain performance. You can also bridge other blockchains if needed.
2. Predictable and customizable handling fee : On a shared license-free network, you have no control over the handling fee. The high interaction volume of some apps will push up the handling fee of the entire network, and your app can only accept it. A custom rate structure means that fees are more predictable, and it will also eliminate the presence of the underlying platform. Users do not need to hold tokens of the underlying platform to use the app-specific blockchain. Allowing users to pay fees in currencies other than the underlying platform token is critical to mainstream adoption.
3. Verification nodes can be customized : You can set corresponding verification node rules and requirements for your blockchain according to your own app needs. You can require validators to comply with the laws of a specific jurisdiction (such as the European Union's General Data Protection Regulation), have high-performance hardware, or provide specific attestations.

These next-generation blockchain networks have established or are about to establish cross-chain bridges connecting Ethereum and Bitcoin. They are also developing cross-chain bridges to connect each other in order to fully realize the vision of the Internet of Blockchains.

Cosmos, Polkadot, and Avalanche are very different at the protocol level (consensus mechanism, economic security topology, etc.), so their functions (inter-chain communication, token economic model, supported app types, etc.) Node participation, pledge attribution, etc.) are also very different. The following will compare the three to help developers, entrepreneurs, investors, researchers, and those considering building projects on these platforms understand the differences between the three and their respective trade-offs.

Comparison of Cosmos, Polkadot, and Avalanche

consensus mechanism

Consensus mechanisms securely and consistently replicate the state of an application over an open computer network. At the same time, in the case of incomplete information or malicious nodes (Byzantine fault tolerance), the network should maintain the effectiveness of fault tolerance and consensus mechanism. Cosmos and Polkadot use Practical Byzantine Fault Tolerance (PBFT), which requires all nodes participating in consensus to communicate with each other. Therefore, network decisions have absolute finality. PBFT has the characteristics of low latency and fast confirmation speed, but it cannot be extended to a large number of nodes in the global open network, because as the verification work increases, the burden on each verification node will increase exponentially. Bitcoin introduces the longest chain consensus mechanism (Nakamoto Consensus), which allows for probabilistic certainty and extremely low error rates. Over time, it will gradually build a reliable and scalable network, but the process is very slow.

• The Cosmos mainnet was launched in March 2019, adopting the Tendermint PBFT consensus, and the transaction confirmation speed is fast. However, since all nodes must communicate with each other, the complexity of quadratic messaging results in only one block being confirmed at a time.
• The Polkadot mainnet was launched in March 2020, and its consensus mechanism separates block production and transaction confirmation: BABE consensus (a variant of Ouroboros Praos) initiates candidate blocks, and GRANDPA (a variant of PBFT) confirms them in batches . This hybrid consensus mechanism optimizes the complexity of secondary messaging to a certain extent.
• The Avalanche mainnet was launched in March 2020, using the Avalanche consensus protocol. This is a unique consensus mechanism that combines resampling of validators (Snowball) and transitive voting, using a directed acyclic graph (DAG) instead of a linear blockchain. The messaging complexity of Avalanche consensus is constant, so it has the characteristics of low latency and large-scale participation. Like the Nakamoto consensus, the avalanche consensus provides probabilistic finality, but the specific parameters can be adjusted, and the error rate is extremely low.

Verify node access

Blockchains use PoW or PoS mechanisms to prevent the same entity from operating multiple identities (Sybil attacks) while opening nodes to participate. Like other new projects, Cosmos, Polkadot, and Avalanche all use the PoS mechanism because it is more energy efficient and has a larger design space. Some projects on these networks also use lightweight PoW mechanisms or fair coin distribution mechanisms.

transaction delay

• Cosmos takes 6-7 seconds to confirm a transaction.
• It takes Polkadot 12-60 seconds to confirm transactions (technical review note: most of them are around 5 seconds), and block generation and transaction confirmation are separated.
• It takes less than 1 second for Avalanche to confirm transactions. Avalanche, like Bitcoin, adopts probabilistic finality confirmation, and the error rate is extremely low.

Computational throughput

The total amount of calculations processed by the network per second depends on the complexity of the virtual machines used by the network and the functions of the actual operating environment. Cosmos, Polkadot, and Avalanche all support dedicated asynchronous blockchain networks that ultimately have unlimited network throughput. The focus is on how these networks can grow and what their interchain economic security structures are.

transaction cost

Transaction fees rise with increased network activity. Cosmos, Polkadot, and Avalanche all support private networks, where each chain can determine its own rate mechanism based on its state growth.

• In Cosmos, each chain can define its own rate mechanism.
• In Polkadot, each chain can define its own rate mechanism. Fees are pre-calculated through a weighting system. It is up to each chain to decide whether to destroy the handling fee.
• In Avalanche, each chain can define its own rate mechanism. On the main network, some functional fees are fixed, and other functional fees are 0. All fees will be destroyed to maintain the long-term interests of token holders.

level of decentralization

The data below is as of March 17, 2022.

• Cosmos nodes need to carry out secondary information transmission, so the number of nodes is limited. Cosmos has 150 active validators, IRIS has 100 active validators, and Osmosis has 100 active validators. Currently, users need to pledge at least 147,231 ATOM (approximately US$1.3 million) to become an active node of Cosmos Hub, and the delegation threshold is 1 ATOM. The total amount of pledges is about $5 billion.
• Polkadot optimizes the secondary message delivery between nodes, and the number of nodes is relatively limited. Polkadot has 297 active validators and Kusama has 1,000 active validators. Currently, users need to pledge at least 1.75 million DOT (about 33 million US dollars) to become an active node of the Polkadot relay chain, and the nomination threshold is 120 DOT. The total amount of pledges is about $12 billion.
• The message delivery volume of Avalanche nodes is constant, so the number of nodes can be expanded infinitely. The Avalanche mainnet has 1,311 active validators. Currently, users need to pledge at least 2,000 AVAX (approximately $160,000) to become an active node on the Avalanche mainnet, and the delegation threshold is 25 AVAX. The total amount of pledges is about $16 billion.

The level of decentralization also depends on the concentration of stake and rewards of nodes (revenues are weighted according to stake), which usually exhibit a long-tail distribution - a small number of nodes have the majority of stake, and many nodes have a small amount of stake. For blockchain platforms, how to achieve fair equity distribution is an unresolved problem, and each project is trying in its own way. For example, since the core of Polkadot is based on PBFT consensus, its number of active nodes is limited, but these nodes can obtain the same benefits through the Phragmén algorithm. With its novel consensus mechanism, Avalanche can realize unlimited expansion of nodes. At the same time, the average weight of nodes is gradually decreasing, thereby improving the level of decentralization.

Interchain Network Topology

The data below is as of March 17, 2022.

• Cosmos is a distributed blockchain network, and each blockchain can have its own verification node. Inter-chain interoperability is achieved through the Inter-Chain Communication (IBC) bridging protocol. Each chain must implement IBC to connect to other chains. Currently, IBC has been deployed on 28 blockchains, focusing on DeFi, EVM smart contracts, social media, privacy, regenerative agriculture, and gaming. Cosmos is developing a cross-chain bridge between Ethereum and Bitcoin.
• Polkadot allows parachains to inherit security from a central relay chain. Parachains do not have their own validators, but they have collators that collect transactions and generate proofs of state transitions for relay chain validators. Parachains achieve interoperability through the cross-chain message (XCM) format, and the security inheritance mechanism provides the possibility for arbitrary data transfer. Currently, Polkadot has 10 parachains, each focusing on DeFi, EVM smart contracts, social media, privacy, games, etc. Polkadot is developing a cross-chain bridge between Ethereum and Bitcoin.
• Avalanche allows validator nodes to overlap: a subnetwork that runs multiple blockchains while also providing validation for the mainnet. Different blockchains in the same subnet can achieve near-instant asset transfer (export/import). Communication between subnets refers to the communication between one chain in a certain subnet and another chain in another subnet, which is currently implemented through a cross-chain bridge (using the ChainBridge-Solidity contract of the EVM chain). In fact, the more the validating nodes of the two subnets overlap, the higher the security guarantee of the communication between the subnets, because the overlapping nodes have interests in both subnets. If a group of nodes behaves maliciously in a certain subnet, their interests in the main network and other subnets will also be at risk. Although Avalanche has not launched a direct interoperability method between subnets, the Avalanche main network can fully serve as an intermediary between subnets. Currently, the Avalanche main network has 3 blockchains: the X chain is used for transfer, the P chain is used for pledge, and the C chain is used for EVM smart contracts. Other blockchains and subnetworks are also flourishing. Also, like other platforms, Avalanche has the Avalanche-Ethereum cross-chain bridge (AB bridge), which operates through a trusted federation and is one of the most used cross-chain bridges among the 60 Ethereum cross-chain bridges today.

In today's Cosmos, bridging blockchains with different security levels without a security sharing mechanism is no different from ordinary cross-chain operations. Therefore, without a common deterministic guarantee, the risk level of inter-chain communication is not fixed. Polkadot's inherited security model allows for unified deterministic guarantees, upon which parachains can securely pass arbitrary data to each other. Avalanche's verification node coincidence model supports the sharing of security between each chain and the main network, and blockchains in different subnets will soon be able to directly share security without using a cross-chain bridge. Therefore, the more nodes that overlap between subnets (nodes that have interests in both subnets), the higher the security guarantee for communication between subnets. Overall, the more nodes that overlap between different blockchains (similar to merged mining in the PoW mechanism), the stronger the security of inter-chain communication.

governance

• Cosmos adjusts consensus parameters and coordinates fund allocation through an on-chain mechanism.
• Polkadot's operating environment logic is all stored on the chain in the form of WASM binary files, allowing fork-free runtime upgrades, which means that decisions will be automatically executed based on referendum results, without requiring developers or verification nodes to perform any operations. Its governance modules include token-weighted voting, rotating committees, time-locked token voting, and adaptive voting bias mechanisms.
• Avalanche can upgrade some parameters through on-chain voting. Its team is developing a wider governance mechanism based on the characteristics of the Avalanche consensus.

development space

All blockchains have the following core components: database, p2p network, consensus mechanism, transaction processing mechanism and state transition function (running environment or virtual machine). Cosmos, Polkadot, and Avalanche provide the above core components and support developers to build custom state transition functions.

• Cosmos provides the Cosmos SDK and Tendermint middleware to support transactions in any programming language. You can develop your own virtual machines and build your own node sets. If you want to start your own blockchain, you need to build a validator node set from scratch and attract nodes from an existing blockchain. You can also deploy smart contracts on EVM compatible chains (Ethermint or CosmWasm).
• Polkadot provides a Wasm-based meta-protocol and a Substrate development kit in Rust. You can use modules provided by Polkadot (such as accounts, assets, governance, EVM, etc.) and custom modules to develop your own virtual machines. You can also use Substrate's on-chain scheduling, off-chain workers, and execution-free model for fee-free transactions. After bidding for the slot in the parachain auction, you can start your own blockchain, and the new blockchain will inherit the security of the relay chain. Alternatively, you can also increase the size of your own validator. You can also deploy smart contracts on EVM compatible chains (Moonbeam, Acala) or use Ink smart contracts.
• Avalanche provides Avalanche Virtual Machine (AVM) for developers to clone and customize their own instances, or create brand new instances as their own virtual machines (the module SDK for developing virtual machines has not yet been released). To start a blockchain, you need to start a subnet and attract verification nodes. The subnet nodes must be the nodes of the Avalanche main network. (Technical review note: Subnet nodes are currently built or recruited by the subnet creators themselves, not necessarily nodes of the Avalanche main network) There is currently a subnet EVM code that starts a custom EVM chain, and you can run it on an EVM-compatible C Deploy smart contracts on the chain.

Topology of Heterogeneous Blockchain Networks

An asynchronous network of dedicated blockchains has the potential to handle large-scale user activity compared to a network where each blockchain is an instance of the same virtual machine. This section will explore the respective blockchain networks and inter-chain communication mechanisms of Cosmos, Polkadot, and Avalanche in more detail.

Cosmos ecology

The Cosmos ecology adopts a distributed network topology. Different blockchains have different purposes, and each has its own set of verification nodes. When communication is required, these chains will communicate via cross-chain bridges. According to the analysis, this topology is "as secure as the least secure chain" (the most secure chain will be less secure when it receives from the least secure asset). However, this topology also endows the Cosmos network with resilience, because the security issues of no single blockchain will determine the survival of the entire ecosystem. However, what is the difference between such a Cosmos ecology and other blockchains that rely on cross-chain bridges? Cosmos has a “no strings attached” policy, and projects like Binance DEX, Oasis, Terra, Nym, and more can use Tendermint to develop and launch their own app-specific blockchains.

The blockchains in the Cosmos ecosystem are connected to each other through the inter-chain communication (IBC) protocol (see 28 interconnected blockchains on the data platform Map of Zones). The blockchains that implement the IBC protocol will be connected to each other, improving the liquidity of the entire Cosmos ecosystem. The operation mode of IBC is very similar to cross-chain bridge. When transferring assets from one blockchain to another, the user needs to 1) lock the asset on the outgoing chain; 2) a third party (possibly a federation relay node) monitoring each blockchain finds the receipt and sends it Delivered to the destination chain; 3) The destination chain verifies the receipt and feeds back the asset representation to the transfer-out chain. In the Cosmos ecosystem, chains that implement IBC have Tendermint light-client verification tools that can use and verify these receipts in communication. Additionally, IBC is a general protocol that can be implemented in different blockchain architectures (see Substrate's IBC implementation). In addition, the new version of IBC will provide a shared security scheme (see Billy Rennekamps' speech for details).

Polkadot's Inheritance Security Topology

Polkadot adopts a hierarchical inheritance security topology, and arbitrary data communication between parachains is very efficient, but these parachains rely on the security leased from the central relay chain. In Polkadot, parachains do not require their own validators, but instead lease security from the relay chain. Specifically, the parachain needs to win slots in an auction (approximately 100 slots in total) and lock DOT tokens (to raise DOT through crowdfunding). Parachains specialize in their respective domains, and their functionality becomes available immediately after they are connected to and synchronized with the Relay Chain via a checker node. Critics believe that different blockchain chains may not require the same level of security. In addition, the security of a single blockchain should not have the ability to determine the survival of the entire ecosystem. Although Polkadot currently advocates a parachain without verification nodes, users can use Substrate to start the blockchain and build their own verification nodes instead of relying on the central relay chain (see Compound Gateway). In addition, parachains can accumulate their own verification nodes, unlock DOT tokens at the end of the lease period, and use cross-chain bridges when cross-chain communication is required. In addition, Polkadot can set up multiple relay chains, which will benefit the entire Polkadot ecosystem. However, the hierarchical topology of the network is likely to be preserved, because cross-chain communication based on inherited security is more efficient than cross-chain bridges.

Polkadot developed the Cross-Consensus Message Exchange Format (XCM) as a common format for communication between parachains, smart contracts, cross-chain bridges, and Substrate pallets. In addition, there is vertical message passing (VMP), which is used for information exchange between relay chains and parallel chains, and cross-chain message passing (XCMP), which is used for information exchange between parallel chains under the same relay chain. Information in XCM is across programs running on the Consensus Virtual Machine (XCVM) (see Gavin Wood's series). Other heterogeneous blockchain networks are also applicable to this abstract method of writing networks and building apps between composable chains.

As the parachain community expands, parachains may wish to have their own verification nodes (see Acala's ppt), so that they become relay chains that lease security from other chains. Although the nested security sharing mechanism may become complicated, all sub-chains can share the deterministic guarantee, and the amount of state transitions processed per second will also increase, expanding the total computing throughput of the Polkadot network.

Avalanche's Network Overlay Topology

Avalanche has a topology where networks overlap each other. Every node that validates the subnet needs to validate the Avalanche mainnet at the same time. (Technical Review Note: There is no such setting at this stage, and it is not mandatory for nodes to be verification nodes of the main network and subnet.) The subnet is composed of a group of verification nodes. A subnet can verify multiple blockchains, but a blockchain can only be verified by one subnet. That is, a node can participate in multiple subnetworks. When launching a new blockchain, you have to provide incentives to attract validators who also validate the mainnet or other blockchains. (Technical Reviewer's Note: Refer to note above, this is not the case) If your chain attracts new validators, those nodes must validate the mainnet and the subnet that runs your blockchain. Overall, the subnetwork architecture determines the network structure in which validators overlap each other (as shown in the figure above), which is determined by the innovative avalanche consensus. The Avalanche Consensus repeatedly re-samples the verification nodes. It does not require all nodes to communicate with each other, but only a small number of nodes to communicate with each other, which greatly reduces the complexity of information transmission in the network. Therefore, even if the number of verification nodes increases to tens of thousands, the bandwidth and processing power requirements of the nodes are constant. Therefore, from the perspective of node participation, the blockchain of the Avalanche platform is more inclusive than the blockchains of Polkadot and Cosmos, because the verification nodes of each chain of Avalanche can be expanded infinitely. How many blockchains a node can run depends on the runtime/virtual machine design complexity of the blockchain, and there is no definite answer yet.

In Avalanche, cross-chain interoperability is very efficient. This is not only because of the fast transaction confirmation speed of Avalanche, but also because the main network ensures shared certainty guarantees (currently X chain, P chain, and C chain can achieve near-instant asset transfer). Avalanche's secure sharing model is different from Polkadot or Ethereum's latest rollup system. Avalanche's novel subnetting architecture supports higher density networks. This is because security sharing occurs not only between the three chains of the main network, but also between all overlapping subnet blockchains. This endows the Avalanche network with composability and programmability, opens up a new design space, and will support Formed Group Networks (GFN; see Reed's Law) that can scale exponentially to millions of daily active users, helping the vision of Web3 accomplish.

application

Heterogeneous blockchain networks Cosmos, Polkadot, and Avalanche provide a broad design space with the innovation of core infrastructure. As of now, Ethereum has been the home of cryptoeconomic innovation. In fact, the teams that start projects on these heterogeneous networks initially optimize existing projects on Ethereum (DEX, AMM, lending, stablecoins, aggregation tools, insurance, NFT platforms, etc.). However, there are also teams taking advantage of the unique advantages of these heterogeneous networks to explore new application scenarios.

In Cosmos, Osmosis combines transaction privacy (using threshold to decrypt transactions to prevent frontrunning) with cross-chain AMM, and realizes cross-chain through IBC. Celestia encodes block data to improve the security of light clients, which is critical to the interoperability of self-sovereign identity blockchains and their security level differences in distributed blockchain ecosystems. Regen incentivizes regenerative agriculture through a crypto-economic platform and utilizes sensor and satellite data with audited ecology. Nym launches mixnet to prevent attackers from analyzing network traffic, even if the attacker has the ability to monitor the entire network. Nym uses Tendermint and Cosmwasm smart contracts to control directory services, node bindings, and mixnet delegated staking. Penumbra protects the privacy of cross-chain network transactions. Big projects like Binance DEX and Terra also use Tendermint. After interoperability through IBC, these blockchains will release greater value.

On the Polkadot network, the Acala parachain is a one-stop DeFi center, providing a wealth of functions from AMM to stablecoin lending. Moonbeam is an EVM-compatible smart contract chain. Subsocial is developing a decentralized social networking platform. Robonomics is developing autonomous robot services. Bit Country is a platform for launching community-specific virtual worlds/metaverses. Integritee and Phala enable decentralized confidential computing and encrypted data storage using a Trusted Execution Environment (TEE). Polkadot's development framework Substrate can also be used independently (not as a parachain) to run blockchains such as Compound Gateway. Although all parachains are designed to be compatible with Polkadot's cross-chain ecology, they should make better use of the Substrate framework's excellent composability, memory efficiency, and self-upgrading meta-protocol governance capabilities to enable new usage scenarios.

Avalanche's EVM-compatible chain C-chain initially attracted teams who wanted to develop "energy-efficient" Ethereum projects. Pangolin is a high-speed AMM modeled after Uniswap. Sherpa Cash follows the example of Tornado and is responsible for providing private transactions. Trader Joe started out as an AMM, later added a lending feature, and is now making its way to a DeFi hub. Benqi, a Compound-like lending app, recently launched AVAX liquidity staking. Platypus is an optimized version of Curve's stablecoin exchange, adding asset-liability management functions. Leading Ethereum projects such as Aave, Curve, and Sushiswap that adopt multi-chain strategies have also launched on the Avalanche C chain, attracting a large amount of liquidity to cross-chain along the AEB bridge. The Avalanche ecosystem also has some new asset types, such as litigation financing. By combining with DAO, the project may connect the legal system to the cryptocurrency network. In fact, Avalanche's innovative consensus and topology of overlapping subnetworks opens up enormous possibilities for future innovative projects.

in conclusion

Heterogeneous blockchain networks Cosmos, Polkadot, and Avalanche provide excellent infrastructure for the blockchain Internet, prove the efficiency of the asynchronous heterogeneous network model, and also improve the current Bitcoin and Ethereum networks. These networks will eventually accommodate millions of daily active users, realizing web3's vision of "the Internet is owned and controlled by users".

Heterogeneous networks have their own advantages and contribute to the realization of a truly decentralized Internet, because they have their own characteristics in design and have made their own trade-offs and trade-offs. Understanding the similarities and differences of these networks will facilitate the development of new systems for the future. Projects using this infrastructure will go beyond smart contract applications and become scalable production-quality systems with dedicated blockchains and their own communities for previously unimaginable scenarios. But it is too early to say these, and there are still some unresolved issues, such as how to ensure that liquidity flows efficiently between chains, rather than existing in a specific chain in isolation? How will an open organization operating across chains prevent the emergence of multi-chain giant whales and ensure a fair distribution of wealth and power?

[1] The Bitcoin network builds on decades of cryptography research, as detailed in the paper "The Academic Origins of Bitcoin" by Arvind Narayanan and Jeremy Clark.

Special thanks to Sam Hart, İstem D. Akalp, Engin Erdogan, Joe Petrowski for their feedback and review suggestions.

Disclosure: The author of this article may hold assets of the projects described in the article.