Do you know how a typical blockchain consensus works?
In most cases, the blockchain consensus requires every node to process all transactions and store the entire network state to achieve consensus.
If this is the case, the blockchain infrastructure is incapable of scaling beyond what a single node can manage!
That’s where blockchain sharding helps.
Source: PBS.Twimg
Applying sharding in blockchain infrastructure helps in data and node partitioning to divide workload and process transactions in parallel for better throughput, scalability, and efficiency.
This article discusses sharding in crypto, the pros and cons of sharding blockchain, and popular blockchains that use to to achieve scalability.
Understanding Blockchain Scalability Issues
The Scalability Trilemma
Visa and Mastercard alone handled 64% of the 791 billion global card transactions in 2024. FYI, Visa has a transaction throughput of 24,000 transactions per second (TPS) while Mastercard scores a decent 5,000 TPS.
Mastercard and Visa are centralised payment networks. Blockchains, on the other hand, are decentralised networks that require all nodes to participate in consensus, computation, and full storage. This decentralisation gives these networks immaculate security and immutability.
However, blockchains have to compromise on scalability as node consensus increases finality time and storage requirements decrease network efficiency. Note how in the table below, three of the most popular blockchains stand at 7, 15, and 2,000 TPS.
Source: pbs.twimg
This is known as the blockchain trilemma. Vitalik Buterin, one of the founders of the Ethereum ecosystem, introduced the concept in one of his articles, in which he mentioned it as a possible solution to the blockchain trilemma.
The blockchain trilemma proposes three main issues—decentralisation, security, and scalability—that developers encounter when building blockchains. It forces them to ultimately sacrifice one aspect as a trade-off to accommodate the other two.
Current Limitations in Transaction Throughput
Blockchains are essentially decentralised networks of nodes that verify and store transactions using different consensus protocols. Bitcoin uses Proof of Work, Ethereum uses Proof of Stake, etc.
Source: EuroMoney
The transactions are grouped as blocks. Suppose a blockchain has 100 validators or nodes. Theoretically, each transaction block would have to be passed by at least 66 validators (2/3rd majority) to become a part of the blockchain.
On an average day, a blockchain processes thousands of transactions. Imagine the load each validator would need to bear to validate each transaction and keep up-to-date with the network state.
Blockchain nodes have limited computing power, bandwidth, and storage. Although decentralisation makes blockchain highly autonomous and secure, it hampers blockchain performance and network efficiency.
Sharding is the process of dividing a blockchain into parts or shards, distinct and with their own set of data and validators, and capable of validating transactions independently of other shards.
Sharding in blockchain is a scalability technique that reduces transaction latency by improving throughput (transactions per second or TPS) and building network scalability. It also improves blockchain performance, reduces gas fees, and improves energy efficiency.
Explore different kinds of blockchain sharding techniques in this paper by researchers at Moscow State University.
Source: X
Some successful implementations include Near Protocol, MultiversX, TON, Polkadot Parachains, Harmony, Ziliqa, etc. Though Ethereum had it as a part of its Ethereum 2.0 Merge roadmap, there has been no substantial development concerning Ethereum sharding, except the Beacon chain going live in 2022, and Layer-2s and rollups assisting Ethereum with scaling.
Currently, Vitalik has proposed to shift Ethereum from the EVM machine to the entirely new RISC-V. Other developers have proposed increasing the gas limit which would push TPS to 2,000 under EIP-9698.
How Sharding Works in Blockchain
Let's take a real-world example to understand the ‘sharding blockchain’ meaning. Suppose a class has 60 students and 1 class representative. The teacher asks the class representative to check whether each student has done their homework. If the representative takes 1o seconds to check each student’s homework, it would take a total of 10 minutes to confirm the same for 60 students.
Now, suppose the teacher allots one monitor for each row of 15 students—four monitors in all and one class representative. The four monitors check the homework in parallel, and within 2.5 minutes, all 60 students are checked. The monitors then inform the class representative of the number of students who did their homework.
That’s basically what it does.
image
It splits a blockchain into smaller chains or ‘shards.’ Each shard acts as a mini-blockchain with its own set of validators who process its transactions independently. Each shard is specifically assigned to some validators who verify and add transactions.
Instead of hundreds of validators working to verify a single transaction, multiple subsets of validators now verify several transactions at once, saving time and enhancing speed and scalability for the blockchain.
Here’s how it works:
Data partitioning: The blockchain is essentially a shared database. Under sharding, the blockchain is divided into shards, each with only a piece of network data. The shard is responsible for processing transactions involving that data.
Assigning Validators: A blockchain may have hundreds or even thousands of validators in bigger chains. The next step involves distributing or assigning validators or nodes to each shard. This reduces each validator’s workload, which leads to faster transaction finality, shorter congestion times, and greater efficiency.
Cross-shard communication: Sometimes, transactions may involve data from two different shards. In that case, shards must be able to communicate with each other to validate the cross-shard transaction and ensure the network state remains the same universally.
Benefits of Sharding in Blockchain
Here’s how it helps to scale blockchain infrastructure, improve storage efficiency, and lower costs:
Enhanced Transaction Throughput
Sharding uses ledger partitioning to distribute the workload among the nodes. This kind of node distribution helps in parallel transaction processing, making the network faster and more scalable.
It can improve transaction throughput by several degrees. Let’s compare Ethereum and Near Protocol to better understand. After the Nightshade upgrade, Near Protocol has a 1-2 second finality, while Ethereum takes anywhere between 15 seconds and a few minutes. Similarly, Ethereum has a TPS of 14-20, while NEAR has a one-shard TPS of 800-1000, and once quadratic sharding is in full force, a 4-shard will have a minimum TPS of 2500-3000.
Improved Storage Efficiency
When database sharding happens, nodes are no longer required to store the entire copy of the blockchain database. Instead, they can make do with the share of data allotted to them and verify transactions based on that. This helps to reduce storage requirements and computational load, making the blockchain network leaner and congestion-free.
Lower Costs
With fewer computation and storage needs, nodes can function efficiently at lower costs, and validators don’t need expensive hardware to join the network. When the blockchain network works at full efficiency, fewer gas fees ensure better user adoption and enterprise-level use cases.
Challenges
Security Concerns
Since shards break down the decentralisation, which gives blockchain its inherent security and perpetuity, protocols need to ensure the network stays secure and efficient to pull off data partitioning successfully. Cross-shard communication can bring in vulnerabilities, and any bugs in smart contract codes may be an easy target for bad actors looking to manipulate the network.
Also, shards need to secure themselves from a new kind of attack, called the 1% attack. Under this kind of attack, bad actors can target any one shard to exploit any smart contract vulnerability to compromise shard security, and take over the entire network.
Complexity in Implementation
Sharding isn’t an easy technical feat to achieve when it comes to blockchain. It involves blockchain partitioning, which leads to data fragmentation. Decentralised ledger technology is built on the idea of a shared ledger and a single source of truth. If there are multiple shards, each of them needs to concur with the remaining shard chains for the ledger to be in a unified network state.
Shard management also involves synchronising the shards and ensuring that transaction processing across shards happens accurately. Different shards are built differently and independently of each other. When transactions involve multiple shards and smart contracts, it often leads to transaction conflicts and aborts.
To avoid such case scenarios, blockchains need to adopt multi-round cross-shard consensus mechanisms to ensure shards talk well with each other and smart contracts can do their job smoothly during
This paper details the challenges faced when it comes to practical deployment.
Blockchain Projects Utilising Sharding
While sharding is still in early stages, there are a couple of projects experimenting with the idea of network partitioning to meet scalability challenges:
Ethereum 2.0
Earlier, there were plans to switch the Ethereum network to sharding to improve throughput and scalability. However, the complexity in implementation and security concerns around cross-shard communication caused developers to drop it from the Ethereum Merge roadmap. Danksharding is in the roadmap, but it may be years before there’s some actual implementation.
image
Instead, the focus shifted to Layer-2s, Layer-3s, and rollups to enhance Ethereum’s scalability. Ethereum adopted proto-danksharding as a short-term solution, a pseudo-sharding technique where L2 fees are reduced, but there would be no impact on TPS. Read in detail about it here.
Near Protocol: After its Nightshade upgrade, Near Protocol enables fast blockchain production and high scalability while lowering barriers to entry.
Zilliqa 2.0: Zilliqa can handle 15,000 TPS with a transaction finality of 2 seconds while ensuring high decentralisation. Ziliqa employs linear scaling via sharding.
Shido Network: Shido uses a sharding module for horizontal scaling, where each shard has its own independent governing and validating system. These shards can talk by sending packets of data and help scale TPS exponentially.
Alephium: This blockchain uses the BlockFlow algorithm and Proof-less-work mechanism to bring scalability and security for its dApps.
TON: TON’s dynamic sharding and ‘infinite blockchain’ design ensures speed and scalability to help it onboard millions of users and ensure real-world adoption.
MultiversX: This blockchain uses sharding to partition network, execution, and state, and builds decentralised, open, and scalable ledgers.
Future of Sharding in Blockchain
Since 2008, when Bitcoin emerged as the first blockchain network, scalability has been a major cause of concern. New advancements in blockchain technology are trying to find a solution to avoid trade-offs between the three essential qualities a good blockchain network must possess.
Ethereum is taking an ecosystem approach to keep the network moving and thriving, while newer blockchains are working on sharding and other scalability solutions to increase throughput and scalability.
To approach an epoch where real-world adoption becomes a reality for distributed systems, sharding might emerge as a true, cost-effective scaling solution. However, it is yet to be seen how developers navigate the turbulent waters of complexity and security while implementing shards.
If NFTs and blockchain intrigue you, check out other interesting pieces on NFT.EU.
