Blockchains are not the only database systems that require verifiable computing. From media rights and royalty management to confidential financial and medical data, there is a plethora of applications that need Verifiable computing. Applications that need complex data structures and  which data (Big Data!) are stored in large datasets. Application that needs to combine privacy and confidentiality requirements (storing data in private tabular databases) with requirements of inferring insights. Think doing SQL queries on a tabular private data repository and needing to know that the query was done correctly, completely and on the right subset of data chosen by the query creator.

Since the computation’s input data is part of the Prover’s input, general-purpose ZK proof systems are not suited for verifying the correctness of computations over large datasets, as typically happens when performing analytics-based queries. In fact, with a moderate-size table containing 1 million numerical values, the corresponding circuit for performing even the simplest operations (like summing together all the table’s values) contains more than 1 million gates that approach the current frontier for which proof systems have reasonable performances.