DIAS

designed for deployment over crowdsourced computational resources to open up the democratization of data analytics

DIAS, the Dynamic Intelligent Aggregation Service is a decentralized, privacy-preserving, networked information system that performs lightweight real-time data analytics such as computation of aggregation functions, for instance, summation, average, maximum, minimum, top-k and other. A DIAS node is actually a server, a cloud computing node or even a personal computer running a piece of software that shares data to other interconnected DIAS nodes and receives back collective information about all shared data in the network. DIAS is designed for deployment over crowdsourced computational resources to open up the democratization of data analytics and their transformation into a public good.

Examples of DIAS applications include the following:

 

Existing data analytics systems usually perform centralized computations or they are centrally managed, in contrast to DIAS that is designed to operated in a fully decentralized fashion. Common implications of centralized data analytics methodologies are the low privacy-preservation, the enabling of surveillance or discriminatory actions and the undermining of citizens’ autonomy. On the other hand, DIAS allows self-determination of information sharing, exchange and storage of hashed or encrypted data, decentralized computations and collective bottom-up self-management of the computational resources in which data analytics are performed.

collective bottom-up self-management of the computational resources in which data analytics are performed

Some of the novel and innovative features of DIAS include its fully decentralized design, its privacy preservation and the computations of several aggregation functions without changes in the main system operation. Moreover, DIAS is intelligent and efficient as it can adjust its computational and communication load based on application requirements and available resources. DIAS is capable of ensuring highly accurate estimations of aggregation function even under rapid changes in the input data.

Related Publications

Comments are closed.