A Dag Based Decentralized Oracle Model: Implementation and Evaluation

Blockchain comes up to solve centralization problems and to be an alternative solution to traditional finance. This technology seems to be very promising and has proven a big efficiency in ensuring trust among the network. It innovates and proposes new forms of governance and capital allocation, it promises decentralized financial services of credit, exchange, insurance, etc. It is the interweaving of computer programs, called smart contracts in a decentralized network, which replace traditional intermediaries and automate business processes. To make smart contracts in the blockchain context consistent and worthy, they must rely on solid data from the outside world, such as market prices observed on all exchange places for example. Blockchains are autonomous systems that are not natively designed to integrate data from the "outside", thus, oracles are external systems that solve the problem of feeding an isolated network with data from the outside. On the other hand, these systems are centralized, consequently, they grant a high risk of data corruption and unreliability issues. In this paper, we present our decentralized oracle approach, based on a Dag (directed acyclic graph) model, its implementation as well as an evaluation of its accuracy in the context of live dynamic financial market data.