The Ethereum ecosystem continues to evolve at a rapid pace, with a new paradigm emerging for Layer 1 (L1) scaling. This approach is not about abrupt overhauls but a systematic, repeatable framework designed to safely and predictably increase the L1 gas limit. Unlike previous reactive strategies, this model emphasizes proactive optimization—identifying performance constraints before they become critical and applying targeted solutions in a structured manner.
This new methodology reflects a maturing network, where upgrades are guided by real-world data, rigorous benchmarking, and a clear multi-phase roadmap. The goal is simple yet ambitious: enable Ethereum to scale sustainably without compromising security or decentralization.
A Three-Step Framework for Sustainable Scaling
At the heart of this new paradigm is a closed-loop process composed of three key stages:
- Bottleneck Identification
Using advanced performance monitoring tools, developers can detect system limitations before the network reaches capacity. This includes tracking execution speed, state access patterns, and propagation delays under stress conditions. Classification of Constraints
Once identified, bottlenecks are categorized into three types:- Client Implementation Limits: Issues rooted in specific execution clients like Geth or Erigon.
- Encapsulation-Level Inefficiencies: Problems related to how transactions are packaged or priced.
- Protocol-Level Barriers: Fundamental architectural constraints requiring deeper changes.
- Preemptive Optimization
Instead of waiting for congestion or failures, upgrades are planned and deployed ahead of demand spikes. This enables smoother transitions during major network upgrades and reduces the risk of chain instability.
This proactive cycle allows for predictable growth in throughput while maintaining robustness across diverse network conditions.
Three-Tiered Mitigation Strategy
To address different classes of bottlenecks, the strategy proposes three distinct response paths:
1. Pure Engineering Optimization
These are client-side improvements that require no network fork. Examples include memory management enhancements, faster trie traversal algorithms, or optimized signature verification. Since these changes don’t alter consensus rules, they can be rolled out independently by client teams.
2. Encapsulation Adjustments
Lightweight protocol tweaks that involve minor forks. One example is gas price re-pricing, which adjusts how certain operations are charged without changing their behavior. These adjustments help rebalance transaction economics and reduce spam vectors without deep architectural shifts.
3. Protocol Restructuring
Major upgrades that redefine core components of Ethereum’s architecture. Key examples include:
- Slot restructuring: Modifying the timing and structure of block production intervals.
- ePBS (execution Proposer-Builder Separation): Decentralizing the builder market to prevent centralization risks and improve censorship resistance.
Each path is chosen based on the severity and nature of the bottleneck, allowing flexibility in execution across development cycles.
Scaling Roadmap: Phased Evolution Through 2027+
The proposed timeline divides Ethereum’s L1 scaling journey into three distinct phases:
Phase 1: Now – Q4 2025 (Glamsterdam Upgrade)
Focus: Client tuning and minor protocol patches.
Efforts center on optimizing existing clients (e.g., Geth, Erigon) through performance profiling and incremental fixes. Benchmarking on test environments like Devnet-1 helps validate gains under high load.
Phase 2: 2025 – 2027
Focus: Data-driven protocol upgrades.
With more empirical data from real usage patterns and stress tests, larger consensus-layer changes will be introduced. This phase will likely see the deployment of ePBS and other structural improvements informed by long-term performance trends.
Phase 3: Post-2027
Focus: zkEVM-driven re-architecture.
Zero-knowledge technology will play a transformative role, enabling a full rewrite of Ethereum’s state, data availability, and execution models. This could lead to native zk-proof validation on L1, drastically improving scalability and finality guarantees.
👉 Explore platforms integrating zero-knowledge proofs and high-performance settlement layers.
The Missing Piece: Parallel Execution
One of the most critical unsolved challenges in Ethereum’s current design is parallelization. Today, transaction execution remains largely sequential, meaning each transaction must be processed one after another—even if they operate on unrelated parts of the state.
True scalability requires deterministic parallel execution, where non-conflicting transactions can be processed simultaneously. Several research directions are being explored:
- Block Access List: Pre-declaring which state elements a block will access, enabling parallel processors to schedule work efficiently.
- Long Access List: Extending access lists to cover deeper storage paths for complex contracts.
- Optimistic Concurrency: Speculatively executing transactions in parallel and rolling back only when conflicts occur.
Implementing any of these will require tight coordination between protocol specifications and client implementations—an effort already underway through dedicated performance branches and cross-client testing initiatives.
Performance Benchmarking: The Foundation of Safe Scaling
A major enabler of this new paradigm is the launch of a real-time performance dashboard, where each execution client runs stress tests on isolated perf branches. The primary metric? Sustained execution speed under worst-case scenarios.
The target is clear: ≥20 Mgas/s per client. Why this number?
Because to safely support a 60 Mgas block, clients must validate it within 3 seconds—requiring ~20 Mgas/s throughput. For 100 Mgas blocks, the bar rises to ~33 Mgas/s.
Recent tests on Devnet-1 and mainnet shadow forks have pushed the network to 100 Mgas, revealing real CPU bottlenecks and state access inefficiencies. These findings not only validate the testing framework but also highlight which clients lag behind—enabling focused optimization efforts.
Future benchmarks will expand to include:
- Re-org resilience testing
- Combined state + compute pressure scenarios
- RPC tip delay measurement under high load
- Standardized block processing timeout (currently ~2 seconds in Geth/Erigon)
Establishing a unified timeout threshold will improve node synchronization and reduce uncle rates during peak congestion.
Dynamic Gas Limit: A Vision for Adaptive Throughput
An intriguing proposal gaining traction is the idea of a dynamic gas limit, where the consensus layer continuously broadcasts a "safe" gas cap based on recent network health metrics.
Rather than a fixed number set by miners or validators, this adaptive limit would adjust in real time—lowering during propagation stress or high verification latency, and increasing when conditions are stable.
Such a system would enhance network resilience, prevent propagation failures during traffic surges, and allow organic throughput growth aligned with actual client capabilities.
Frequently Asked Questions (FAQ)
Q: Why can't we just raise the gas limit now?
A: Raising the gas limit without ensuring client performance is risky. If nodes can't process large blocks fast enough, it leads to longer propagation times, increased orphan rates, and potential centralization as only powerful nodes keep up.
Q: What is the difference between ePBS and current PBS?
A: Today’s Proposer-Builder Separation (PBS) relies on centralized relay systems. ePBS (execution PBS) aims to decentralize the builder market, allowing anyone to participate and reducing censorship risks while improving liveness.
Q: How does parallel execution improve scalability?
A: By allowing non-overlapping transactions to run simultaneously, parallel execution can multiply effective throughput without increasing block size—making better use of available hardware resources.
Q: Is zkEVM integration part of Ethereum’s official roadmap?
A: While not yet finalized, zkEVM research is actively funded and tested. Full L1 integration post-2027 is considered a plausible path for achieving quantum-leap scalability.
Q: What role do clients play in this scaling model?
A: Clients are central—they must implement optimizations, support new features like access lists, and meet performance targets. Without client-level readiness, even the best protocol upgrades fail in practice.
Q: How does the performance dashboard help developers?
A: It provides transparent, comparable metrics across clients under identical stress conditions. This fosters healthy competition, accelerates debugging, and ensures upgrades are data-backed rather than speculative.