Blockchain technology has the potential to transform industries by providing a decentralized, secure, and transparent means of recording transactions. From finance to supply chain management, its potential is vast. However, one persistent challenge looms large: scalability. As networks grow and transaction volumes increase, many blockchains struggle to keep up without sacrificing security or decentralization. This article dives into why scaling blockchains is so difficult, introduces Ouroboros Leios—a promising upgrade for the Cardano blockchain—and evaluates its potential to redefine scalability. This article explores the key challenges limiting blockchain scalability and examines current solutions addressing these issues. Through this analysis, you'll gain a clear understanding of what makes Ouroboros Leios unique in tackling scalability. Why Blockchain Scalability is a Hard Problem At its core, a blockchain is a distributed ledger where every node must agree on a single, consistent global state. This consensus requirement is both its strength and its Achilles' heel. Unlike centralized systems that can simply add servers behind a load balancer, decentralized networks demand that every participant validates every transaction. At least that's what first-generation blockchains are like. Transitioning from the current state to a new state requires time, computing power, and network communication. The coordination among nodes creates a bottleneck that goes beyond mere hardware limitations like bandwidth or processing power. Several specific factors compound this challenge: Network Delays: Global synchronization across hundreds (or even thousands) of nodes introduces latency. A new block is usually produced by a single node. The more nodes (hops) a block has to travel through before it reaches the most distant one, the higher the latency will be. Throughput Limitations: Throughput limitations are caused by technical and design constraints that prioritize security and decentralization. Blocks are usually small. Each block can hold only a limited number of transactions, as larger blocks would take longer to propagate through the network. Blocks are produced at fixed intervals. These intervals must respect the network latency. To allow many nodes to participate (even with limited hardware), blockchains avoid pushing too much data too quickly. Higher throughput often means fewer validators can keep up, threatening decentralization. Hardware Constraints: Increasing block size or transaction complexity (e.g., smart contracts) raises the bar for CPU, memory, and storage. Smaller nodes struggle to keep pace, risking centralization as only well-resourced participants can validate. One important consideration regarding decentralization is the affordability of devices enabling participation in consensus. Security Trade-offs: Any scaling solution must guard against double-spends, forks, or adversarial attacks, adding layers of complexity to consensus mechanisms. The majority of honest nodes should agree among themselves on the transition to the new state, including resolving any conflicts, before negotiating the next transition. This interplay of factors—often dubbed the blockchain trilemma—means that improving one aspect (e.g., scalability) often compromises another (e.g., decentralization or security). The question is: how can we scale without breaking the decentralized promise? The Broader Context: Current Scaling Solutions and Their Limits Let’s examine the landscape of existing scaling solutions and their trade-offs: Layer 2 (L2) Solutions The Ethereum ecosystem addressed scalability challenges through L2s. L2s offload transaction processing to secondary layers like rollups (e.g., Optimism, Arbitrum) or state channels. While they can boost throughput to thousands of TPS off-chain, they often rely on centralized sequencers to order transactions. If a sequencer fails or is compromised, the system halts, introducing a single point of failure. This centralization undermines the decentralized ethos, a concern raised by researchers. A low-scalability blockchain (L1) serves as a decentralization and security anchor for L2s. Depending on the implementation details, the level of risk varies from case to case. This solution has many disadvantages. For example, it splits liquidity and users into several separate networks. Users must explicitly transfer liquidity from L1 to L2. Interaction between a user on network A and a user on network B can be a challenge. Scalability is improved within the ecosystem, rather than through changes to blockchain consensus. For layered architecture to succeed, users must be shielded from the complexity of individual layers, just as the Internet abstracts network intricacies, enabling seamless interaction. Sharding Sharding splits the blockchain into a few parallel shards to process transactions concurrently. This can theoretically increase TPS, but cross-shard interactions require complex coordination, potentially weakening security and fragmenting the global state. Transactions that involve multiple shards require coordination, increasing latency and complexity. This can introduce delays or failures in inter-shard transactions. As the number of users and applications increases, the demand for cross-shard communication can increase. This can eventually lead to a bottleneck. However, unlike L2s, communication between shards is more secure because it is secured by the blockchain itself. Validators must be assigned, rotated, or shuffled between shards, which requires coordination and synchronization mechanisms. Poor design can lead to inefficiencies or vulnerabilities. Sharding is implemented in projects like MultiverseX, Near, and others. High-Performance Blockchains Chains like Solana and Sui achieve high throughput with optimized consensus mechanisms—Solana’s Proof of History and Sui’s parallel execution. In both cases, the hardware requirements are enormous. This hardware barrier excludes smaller operators, centralizing control to well-funded entities. The advantage of highly scalable monolithic blockchains like Solana is that all execution, consensus, and data availability happen within one tightly integrated system. Users and liquidity are all together in one place. Although very user-friendly, these blockchains optimize for high scalability at the expense of decentralization. The key question is whether this approach truly solves the blockchain trilemma. An ideal solution should enhance scalability, security, and decentralization simultaneously, without sacrificing one for the others. Ouroboros Leios: A New Approach to Scaling Currently, node computational resources and bandwidth are not efficiently utilized. Cardano mints a new block every 20 seconds, broadcasting it across the network. Within moments, all nodes validate the block, but after this brief peak in activity, network resources remain idle until the next block is minted. This results in spikes in resource utilization rather than steady, optimized usage. Enter Ouroboros Leios, the latest evolution in Cardano’s Ouroboros family of consensus protocols. Designed to address these scalability hurdles, Leios introduces a sophisticated three-tiered block structure that enables parallel transaction processing while preserving a unified global state. The valid global state is maintained through Ranking Blocks (RK), which reference higher-frequency minted data blocks. This approach preserves blockchain linearity while enabling parallel processing. The image shows the structure of blocks and the resource usage of nodes in slots. Note that in each slot, nodes can perform some work on different blocks. Let's take a closer look at the blocks: Input Blocks (IBs): These are the workhorses of the system, containing raw transactions. Minted every 0.2 to 2 seconds by a distributed set of nodes, IBs allow for rapid transaction minting. This parallelism is where Leios aims to boost throughput significantly. Endorsement Blocks (EBs): EBs ensure that only non-conflicting, valid transactions move forward, resolving potential double-spends and other conflicts. EBs can be minted approximately every 5 seconds. Ranking Blocks (RBs): These serve as checkpoints, finalizing the order of endorsed transactions and cementing the global state. Minted every 20 seconds, RBs provide stability and a canonical reference point. This architecture decouples the stages of transaction handling—submission, validation, and finalization—allowing the network to process transactions concurrently. Several Input Blocks can be minted and validated at the same time a new Ranking Block is minted. The last valid RB acts as an anchor, ensuring that new transactions reference an agreed-upon state, while EBs filter out conflicts. This pipelined approach leverages underutilized computational resources. At any given moment, there is a known single global state that is referenced by newly minted blocks (let's ignore blockchain reorganization). Simulations conducted by the Leios team, using both Haskell and Rust environments, suggest a potential throughput of up to 11,000 TPS on a mainnet-scale network of 3,000 nodes. This figure, while impressive, is based on controlled conditions and awaits real-world validation. The system is able to scale linearly with network resources. Deploying Leios will result in higher hardware requirements. However, not necessarily as high as Solana. We will have to wait for the details. Is Ouroboros Leios a True Breakthrough? Ouroboros Leios remains in the research and development phase, with no mainnet deployment yet. Still, its theoretical advantages are compelling: Unified Global State: Unlike sharding or L2s, Leios maintains a single ledger, simplifying consensus and enhancing security. This brings Leios closest to a highly scalable monolithic blockchain like Solana, with the difference that Cardano remains decentralized. Unlike sharding and L2s, users and liquidity will remain on one network without cross-chain or cross-shard overhead. Decentralization-Friendly: With lower hardware requirements, it avoids the high barriers of Solana or Sui, broadening participation. Security will remain under the control of L1. Scalability Potential: The 11,000 TPS target, if achieved, would position Cardano as a leader among L1 blockchains, rivaling high-performance chains without their centralization risks. Some L1s can reportedly achieve higher TPS. It is possible, but at the cost of compromises that should not be accepted in a decentralized world. If successful, Ouroboros Leios could pave the way for mass adoption, supporting advanced decentralized finance (DeFi) and global dApp infrastructure on Cardano. It addresses not just technical scalability but also the accessibility needed to bring blockchain to the mainstream, beyond crypto enthusiasts. The upcoming months will be critical as the Cardano community deliberates funding, with potential deployment targeted for 2026 based on current timelines. If Cardano can achieve high scalability while preserving decentralization, it could effectively solve the blockchain trilemma. However, this process won’t be simple—the community may need to vote on potential trade-offs to balance scalability with security and decentralization.
