SATOS: Storage Agnostic Tokens over Opaque and Substructural Types
Abstract:
Public blockchains in support of Smart Contracts (SC), like Ethereum enable everyone to represent scarce, valuable resources (like cryptocurrencies) as so-called tokens. Token issuing and management was the first blockchain use case. However, programming languages and runtime systems used in the current blockchains for their SCs lack a secure and straightforward way to implement and handle tokens. The unnecessary complexity in doing so can lead to erroneous implementation of tokens and applications built on top of these, including the loss or theft of tokens as it happened. The most known attack was "TheDAO" attack which led to the "loss" of tokens, valued at that time at approximately 60 M US Dollar. A better and secure token representation directly embedded into a SC runtime and SC programming language could prevent loss of tokens. Thus this paper presents an approach including parts of a programming language using it. The core of the model is to use opaque and substructural data types together with an onchain soundness checker to generically represent tokens securely as values similar to integers and booleans. Such opaque data types enforce that only a designated piece of code can create values of that type. The substructural data types allow arbitrary values to express scarcity by preventing the duplication and elimination of values. The on-chain soundness checker ensures that the deployed code does not violate guarantees given by the type system, which includes opaque and substructural data types.