Design and Analysis of Decentralized Incentive and Reputation Mechanisms for Device to Device Ecosystems

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering

PhD Thesis Defence

Title: "Design and Analysis of Decentralized Incentive and Reputation 
Mechanisms for Device to Device Ecosystems"




The proliferation of computationally capable mobile devices gave birth to 
device-to-device (D2D) ecosystems where devices communicate, exchange 
resources and assist each other with the execution of tasks to extend 
their battery lifetime and improve users' quality of experience. The 
effectiveness of such collaborations depends on the characteristics of the 
selected devices and the willingness of their owners to help. Thus, mobile 
applications need to adopt mechanisms that motivate collaboration and 
compensate those who share their resources. Research has proposed 
incentive schemes and reputation-based mechanisms designed to operate via 
centralized authorities.

In this thesis, we design and analyze decentralized mechanisms to 
facilitate the exchange of resources, reputation, and credit, and to 
maintain a ledger in D2D ecosystems. For resource exchange, we introduce 
FlopCoin, a cryptocurrency operating as an incentive scheme for 
computation offloading and maintained by Internet-connected users called 
miners. We show how FlopCoin rewards collaborating devices and sidelines 
selfish ones. For reputation exchange between mobile users, we present 
OPENRP, a lightweight and scalable middleware that provides a unified 
interface to applications. OPENRP updates the reputation of participating 
peers based on their interactions and chooses the best peers with which to 
collaborate. The traffic generated by applications using OPENRP is lower 
compared to benchmark strategies. For credit exchange in D2D ecosystems, 
we propose LocalCoin, a cryptocurrency that requires minimal computational 
resources and works with off-the-shelf mobile devices. We prove that under 
the assumption of a sufficient number of mobile users and adequately 
selected tuning parameters the probability of double spending converges to 
zero. Finally, to maintain a decentralized ledger in D2D ecosystems, we 
design Mneme and propose a consensus protocol that is based on users' 
context. We prove that the proposed protocol guarantees persistence and 
liveness, which are essential properties of a distributed ledger.

Date:			Tuesday, 17 July 2018

Time:			11:00am - 1:00pm

Venue:			Room 5560
 			Lifts 25/26

Chairman:		Prof. Volkan Kursun (ECE)

Committee Members:	Prof. Pan Hui (Supervisor)
 			Prof. Gary Chan
 			Prof. Dimitris Papadopoulos
 			Prof. Yi Yang (ISOM)
 			Prof. Dah-Ming Chiu (Inf. Engg., CUHK)

**** ALL are Welcome ****