This week, Solayer, a Solana ecosystem restaking project, unveiled its 2025 roadmap, with its centerpiece being the forthcoming hardware-accelerated SVM blockchain, Solayer InfiniSVM. For Solayer, “Solayer InfiniSVM” represents a critical step toward achieving its long-term vision.
Chaofan Shou, Solayer’s Chief Engineer, joined the team after leaving his PhD program at UC Berkeley. Shou revealed that the team drew significant inspiration from Solana’s validator client, Firedancer, and decided to offload most of Solana’s components onto SmartNICs and programmable switches. This approach significantly enhances network processing capabilities, making transaction execution far more efficient.
In Solayer Chain, every transaction follows a structured workflow. Transactions are first routed through a scalable ingress cluster of hundreds of thousands to millions of nodes, where they undergo cleansing and pre-execution based on probabilistic future state predictions. Next, execution snapshots are sent to a sequencer built with Intel Tofino switches and additional FPGA components. Notably, most transactions are already validated during the pre-execution stage and do not require re-execution on the sequencer. For conflicting transactions, Solayer Chain’s sequencer leverages a state-of-the-art (SOTA) scheduling algorithm based on granular account access patterns collected during pre-execution to ensure fairness and efficiency.
According to Chaofan Shou, Solayer Chain achieves over 16 billion transactions per second (TPS) for simple workloads, while handling up to 890,000 TPS for workloads with conflicts. This allows the platform to process billions of USDC transfers per second and support millions of users trading the same memecoin on Raydium simultaneously.
But how does Solayer InfiniSVM make this possible?
As outlined in the Solayer Chain Lightpaper, the platform achieves infinite scalability for single-state blockchains by distributing workloads across specialized hardware and clusters while maintaining a global atomic state.
Solayer leverages SDN (Software-Defined Networking) and RDMA (Remote Direct Memory Access) to achieve 100 Gbps throughput while preserving atomic state. By offloading tasks to hardware circuits and kernels, Solayer InfiniSVM achieves 1ms transaction finality.
Here’s a brief overview of the Solayer Chain workflow:
Transaction Ingress
Each transaction enters an initial ingress point (illustrated in the top-left corner of the diagram) for signature verification (sigverify) and local deduplication (to avoid duplicate transactions).
Pre-Execution Phase
Verified transactions are then sent to a pre-execution cluster for probabilistic execution.
Sequencing and Scheduling
Transaction results and intermediate snapshots are transmitted to the sequencer via InfiniBand, a high-speed, low-latency network architecture designed specifically for high-performance computing and data center environments. The sequencer leverages SDN switches and FPGA to determine whether a transaction follows the simple path or the complex path:
Simple Path: If all accounts involved in a transaction are at their latest versions during the pre-execution phase, state changes are directly applied via RDMA (Remote Direct Memory Access) using local caches on the SDN, bypassing further processing by the sequencer.
Complex Path: If at least one account has an updated version, the transaction is routed to a local memory pool. The sequencer schedules transactions in the memory pool to ensure fairness and optimal parallel execution.
State Updates
After execution, transaction state changes are updated in a sharded database. The sharded database uses the RDMA protocol for efficient cross-node data access.
Transaction Broadcast
Once transactions are executed and state changes are written, they are broadcast globally via Points of Presence (PoPs).
Solayer Chain adopts a Proof-of-Authority-and-Stake (PoAS) hybrid consensus protocol, where transactions are batched into fragments called shreds. Each shred includes a slot number, a vector of transactions, metadata about account version access, and a link hash. Trusted entities act as sequencers to publish shreds, while stakers vote to decide whether the shreds can be accepted.
What sets Solayer Chain apart is its focus not only on performance but also on significant user experience improvements through chain-level support features, including:
Hooks: Allow developers to embed post-transaction logic, such as arbitrage, liquidation, and accounting, directly into the chain.
Large Transactions: Support for larger transaction sizes, enabling cross-program calls.
Cross-Chain Calls: Built-in system programs to facilitate atomic cross-chain operations.
Integrated OAuth Support: Enables users to leverage OAuth services like Google, X, or Reddit as wallets.
Each time a hook is executed, the bidding amount is distributed as follows:
40% to the Transaction Initiator: Encourages active network usage.
40% to the Program Owner: Incentivizes the development and maintenance of high-quality programs.
20% to the Network: Offsets the additional on-chain computational costs.
This fee distribution model encourages greater participation from both developers and users, enhancing the platform’s functionality and increasing network activity. Moreover, by requiring bids, Solayer Chain mitigates spam transactions and prevents off-chain malicious MEV exploits, providing an additional layer of security to the network.
The vision of Solayer Chain extends beyond enhancing performance; it aims to integrate more user-centric and developer-friendly features into blockchain technology.
This week, Solayer, a Solana ecosystem restaking project, unveiled its 2025 roadmap, with its centerpiece being the forthcoming hardware-accelerated SVM blockchain, Solayer InfiniSVM. For Solayer, “Solayer InfiniSVM” represents a critical step toward achieving its long-term vision.
Chaofan Shou, Solayer’s Chief Engineer, joined the team after leaving his PhD program at UC Berkeley. Shou revealed that the team drew significant inspiration from Solana’s validator client, Firedancer, and decided to offload most of Solana’s components onto SmartNICs and programmable switches. This approach significantly enhances network processing capabilities, making transaction execution far more efficient.
In Solayer Chain, every transaction follows a structured workflow. Transactions are first routed through a scalable ingress cluster of hundreds of thousands to millions of nodes, where they undergo cleansing and pre-execution based on probabilistic future state predictions. Next, execution snapshots are sent to a sequencer built with Intel Tofino switches and additional FPGA components. Notably, most transactions are already validated during the pre-execution stage and do not require re-execution on the sequencer. For conflicting transactions, Solayer Chain’s sequencer leverages a state-of-the-art (SOTA) scheduling algorithm based on granular account access patterns collected during pre-execution to ensure fairness and efficiency.
According to Chaofan Shou, Solayer Chain achieves over 16 billion transactions per second (TPS) for simple workloads, while handling up to 890,000 TPS for workloads with conflicts. This allows the platform to process billions of USDC transfers per second and support millions of users trading the same memecoin on Raydium simultaneously.
But how does Solayer InfiniSVM make this possible?
As outlined in the Solayer Chain Lightpaper, the platform achieves infinite scalability for single-state blockchains by distributing workloads across specialized hardware and clusters while maintaining a global atomic state.
Solayer leverages SDN (Software-Defined Networking) and RDMA (Remote Direct Memory Access) to achieve 100 Gbps throughput while preserving atomic state. By offloading tasks to hardware circuits and kernels, Solayer InfiniSVM achieves 1ms transaction finality.
Here’s a brief overview of the Solayer Chain workflow:
Transaction Ingress
Each transaction enters an initial ingress point (illustrated in the top-left corner of the diagram) for signature verification (sigverify) and local deduplication (to avoid duplicate transactions).
Pre-Execution Phase
Verified transactions are then sent to a pre-execution cluster for probabilistic execution.
Sequencing and Scheduling
Transaction results and intermediate snapshots are transmitted to the sequencer via InfiniBand, a high-speed, low-latency network architecture designed specifically for high-performance computing and data center environments. The sequencer leverages SDN switches and FPGA to determine whether a transaction follows the simple path or the complex path:
Simple Path: If all accounts involved in a transaction are at their latest versions during the pre-execution phase, state changes are directly applied via RDMA (Remote Direct Memory Access) using local caches on the SDN, bypassing further processing by the sequencer.
Complex Path: If at least one account has an updated version, the transaction is routed to a local memory pool. The sequencer schedules transactions in the memory pool to ensure fairness and optimal parallel execution.
State Updates
After execution, transaction state changes are updated in a sharded database. The sharded database uses the RDMA protocol for efficient cross-node data access.
Transaction Broadcast
Once transactions are executed and state changes are written, they are broadcast globally via Points of Presence (PoPs).
Solayer Chain adopts a Proof-of-Authority-and-Stake (PoAS) hybrid consensus protocol, where transactions are batched into fragments called shreds. Each shred includes a slot number, a vector of transactions, metadata about account version access, and a link hash. Trusted entities act as sequencers to publish shreds, while stakers vote to decide whether the shreds can be accepted.
What sets Solayer Chain apart is its focus not only on performance but also on significant user experience improvements through chain-level support features, including:
Hooks: Allow developers to embed post-transaction logic, such as arbitrage, liquidation, and accounting, directly into the chain.
Large Transactions: Support for larger transaction sizes, enabling cross-program calls.
Cross-Chain Calls: Built-in system programs to facilitate atomic cross-chain operations.
Integrated OAuth Support: Enables users to leverage OAuth services like Google, X, or Reddit as wallets.
Each time a hook is executed, the bidding amount is distributed as follows:
40% to the Transaction Initiator: Encourages active network usage.
40% to the Program Owner: Incentivizes the development and maintenance of high-quality programs.
20% to the Network: Offsets the additional on-chain computational costs.
This fee distribution model encourages greater participation from both developers and users, enhancing the platform’s functionality and increasing network activity. Moreover, by requiring bids, Solayer Chain mitigates spam transactions and prevents off-chain malicious MEV exploits, providing an additional layer of security to the network.
The vision of Solayer Chain extends beyond enhancing performance; it aims to integrate more user-centric and developer-friendly features into blockchain technology.