Subnets

Hierarchical subnets

Most blockchain scaling solutions either inherit L1 security without their own consensus algorithms (e.g., rollups) or have their own consensus but lack L1 security (e.g., sidechains). In contrast, the hierarchical subnet strategy employed by Recall operates as a hybrid of these approaches, allowing subnets to have their own consensus algorithms while also inheriting security from parent subnets. This hierarchical structure also means subnets themselves can optimize for reducing latency, while the overall network can scale horizontally to accommodate greater throughput.

State replication

Subnets provide semantic isolation, allowing for the creation of specialized subnets tailored to specific use cases. Examples include subnets defined by geographic region(s), service-level agreements (SLA), feature sets, public/private, encryption, or even targeted data types and use cases (e.g., AI, ML, or private app data).

Recall's core is a deterministic state machine that interfaces with the consensus engine, which in turn ensures the blockchain's state is consistently replicated across each subnet. At a high level, Recall combines onchain contracts, offchain services, and async computation through functions. Its EVM compatibility ensures that it can integrate smoothly with existing Ethereum tooling, while its native support for asynchronous compute makes it uniquely suited to data-intensive applications.