A Decentralised Big Data and Machine Learning Network Powered by a Computing-Centric Blockchain
DxChain Network is a big data and machine learning network which is powered by a computing-centric blockchain with a native protocol token (also called “DX”). Its end users can potentially use this network as a data exchange platform to trade data and as a business intelligence analytics platform to analyse data for supporting business insights. Unlike Bitcoin and Ethereum, where the incentives are driving miners to sustain the huge amount of computation needed to maintain the blockchain consensus, the DxChain Network provides miners incentives based on the usefulness of the work that they passively provide: storage and computation. Since this storage platform is designed through using a principle of decentralisation, the protocol has a mechanism to control the reliability of the file access. In garnering the benefits of P2P network and Hadoop HDFS file systems, the robustness and accessibility is intended to be ensured.
DxChain Network will incorporate Hadoop in its system as it is an industry-proven big data platform. DxChain Network is designed to assign several roles to manage and schedule jobs in the system which achieves the computation goal in a centralised parallel computing system. Built on top of the DxChain Network, a collection of tools would be developed to speed up the computation and analysis process. Additionally, machine learning algorithms can potentially be built upon these to facilitate more computing-driven tasks. At the time of retrieving files, DxChain
Network is also designed to support computation which provides more flexibility of data exchange. With the smart contract, this platform is intended to be especially useful for decentralising data, building and running distributed applications.
Introducing DxChain Network, a decentralised big data and machine learning network, powered by a computing-centric blockchain with four major innovations:
- A new decentralised computing framework to introduce Probable Data Computing and verification game
- A new chains-on-chain design to orchestrate a master chain and two side chains - Data Side Chain and Computing Side Chain
- Incorporate Hadoop to DxChain Network to facilitate big data and machine learning
- Flexible and powerful DxChain Network-based system architecture dedicatedly designed to support most business data exchange and data analytics requirements
Functionally, the entire system is designed as a platform to provide big data and machine learning related computation services with the support of decentralised data storage services. Database query and business intelligence tasks are performed for data that could not be owned by any commercial company but can be traded and stored on the DxChain Network. This platform is designed to connect personal computers, as well as specificallydesigned miner machines, to potentially facilitate easier data computation at a lower cost. DxChain Network provides two fundamental capabilities: computation and storage. Currently, only big companies have the capability to run big data tasks, because not only can these companies afford expensive hardware but these companies also own most of the existing consumer data. For example, Google possesses the navigation data of whoever that uses Google Maps. If Google Maps users could find a way to sell their navigation data, then they will do so for economic reasons. Users will not worry about their privacy since the data has already been seen and used by Google. If third parties can sell navigation data, combined with other commercial data, then advertising campaigns will be more accurate and customised. Most importantly, big companies like Google and Facebook will not monopolise the usage of these data. This is the dream of decentralisation. The DxChain Network is designed to serve as a data trading platform for users who want to sell their data. Big data and machine learning computation is intended to make the use of data easier and more flexible.
PROVIDE SOLUTIONS TO BIG DATA PROBLEMS
- Maximize individual dataset to its capacity
- Empower data subjects to control their own data
- Protect personal data with flexibility
- Support BI and machine learning
INNOVATIONS - CHAINS-ON-CHAIN ARCHITECTURE
The revolutionary Chains-on-chain architecture is designed to solve the data computation, storage, and privacy issues.
- VERIFICATION GAME
Verification game provides a framework which securely perform any computational task and vastly reduces the number of redundant network node computations.
- PoSt & PDP
The Proof of Spacetime(PoSt) is used as the consensus protocol of storage. It provides the foundation for the decentralized storage network.
- PDC (Provable Data Computation)
PDC is a statistical framework to verify the correctness of computation results with few copies of redundant computing.
- PRIVACY PRESERVING
DxChain does not only provide the data storage function, but it also gives the flexibility of retrieving files at a more granular scale.