Conjecture on breadth of applications admitting concurrency under Groundhog

Determine whether the vast majority of important blockchain financial applications can be implemented to admit some degree of concurrent execution under Groundhog’s model of unordered blocks with snapshot reads and commutative typed modifications, thereby avoiding strictly sequential semantics except where actor-style sequencing is required.

Background

Groundhog proposes a smart contract execution engine that processes blocks of transactions concurrently with no intra-block ordering, using snapshot reads and commutative typed modifications plus reserve-commit constraints to ensure deterministic outcomes. This design differs from strictly serializable semantics and requires developers to structure contracts to allow concurrency where possible.

The conjecture concerns the practical applicability of Groundhog’s semantics to real-world financial applications commonly deployed on blockchains. Establishing the breadth of applications that can be adapted to admit concurrency would clarify the limits of Groundhog’s model and guide developers on when actor-style sequencers (as in existing message-passing systems) are necessary.

References

We conjecture that not all but a vast majority of important applications can be built in a manner that admits some concurrency.

Groundhog: Linearly-Scalable Smart Contracting via Commutative Transaction Semantics  (2404.03201 - Ramseyer et al., 2024) in Section 9 Discussion and Limitations, Development Difficulty