Traffic Offloading Algorithm Considering Social Context in Mobile Network
- 주제(키워드) Traffic offloading , small cell backhaul networks , social context , social networking services , LIPA/SIPTO , LTE , LTE-A , 5G
- 발행기관 아주대학교 일반대학원
- 지도교수 김재현
- 발행년도 2019
- 학위수여년월 2019. 8
- 학위명 박사
- 학과 및 전공 일반대학원 전자공학과
- 실제URI http://www.dcollection.net/handler/ajou/000000029131
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
Traffic offloading is one of the promising solutions to deal with the traffic overload problem in the mobile network operator (MNO)’s core network. Many works for traffic offloading have been studied based on the various radio access technologies (RATs), WiFi, device-to-device (D2D), backhaul networks for small cell, and so on, but the traffic offloading via WiFi and D2D have the limitations in terms of coverage, battery, storage, and etc. In addition, social networking service (SNS) traffic is growing dramatically as a result of the high popularity of SNS services, such as Facebook and Instagram, and so on. For the design of the effective traffic offloading scheme, it is necessary to consider the social context. Thus, this dissertation proposes the traffic offloading algorithm considering the social context through the small cell backhaul network. First, the traffic offloading algorithm exploiting the selection frequency for each application is proposed to achieve the effective data rate to the maximum. To achieve the objective of the proposed traffic offloading algorithm, the estimation model of the application selection probability is proposed based on that the application selection probability is proportional to the selection frequency of application. The results of the performance evaluation show that the proposed traffic offloading algorithm enables to effectively satisfy the effective data rate objective compared with the QoS-based offloading algorithm with no considering social context despite the less traffic offloading in the proposed algorithm. Second, the traffic offloading algorithm considering the user's social impact and application's popularity is proposed to maximize the user's QoS. For this traffic offloading algorithm, the estimation model of application selection probability is proposed from the perspective of the user's social impact and the application's popularity. The results of performance evaluation show that the proposed traffic offloading algorithm has a good performance compared to the algorithms that do not take into consideration the user's social impact and the application's popularity, especially from the aspects of data rate and transmission delay. Third, the traffic offloading algorithm involved in the direct and indirect user’s impact is proposed to maximize the user's QoS versus the QoS requirement, and alleviate the core network load of MNO to the maximum. To achieve this goal, the estimation model of application selection probability is proposed from the aspects of direct impact by each user and indirect impact by other users. For this, the social relationships and social impact for each user are considered in terms of users, and the popularity and selection frequency for each application are considered in terms of applications. The performance evaluation results show that the proposed algorithm can improve the QoS of users and reduce the core network load of MNO better than the other algorithm that does not exploit the social context.
more목차
1 Introduction 1
1.1 Background and motivation . . . . . . . . . . . . . . . . 1
1.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Related Works 9
2.1 Traffic offloading in mobile network . . . . . . . . . . . . 9
2.2 Social context in mobile communications . . . . . . . . . 11
2.3 Social context in mobile traffic offloading . . . . . . . . . 13
3 System Model 15
3.1 Network architecture . . . . . . . . . . . . . . . . . . . . 15
3.2 Social context model . . . . . . . . . . . . . . . . . . . . 17
4 Selection Frequency Based Traffic Offloading Algorithm 21
4.1 Estimation model of the application selection probability
by the selection frequency of application . . . . . . . . . 22
4.2 Traffic offloading algorithm by the selection frequency of
application . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.3 Performance evaluation . . . . . . . . . . . . . . . . . . . 26
4.3.1 Evaluation environment . . . . . . . . . . . . . . 28
4.3.2 Evaluation results . . . . . . . . . . . . . . . . . . 31
4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5 Social Impact and Popularity Based Traffic Offloading Algorithm
36
5.1 Estimation model of the application selection probability
by user’s social impact and application popularity . . . . 37
5.2 Traffic offloading algorithm considering user’s social impact
and application’s popularity . . . . . . . . . . . . . 39
5.3 Performance evaluation . . . . . . . . . . . . . . . . . . . 46
5.3.1 Evaluation environment . . . . . . . . . . . . . . 46
5.3.2 Application selection probability . . . . . . . . . . 47
5.3.3 Social context influence . . . . . . . . . . . . . . . 50
5.3.4 QoS weighting factor influence . . . . . . . . . . . 56
5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6 Direct and Indirect User’s Impact Involved Traffic Offloading
Algorithm 59
6.1 Estimation model of application selection probability by
direct and indirect user’s impact . . . . . . . . . . . . . . 60
6.2 Traffic offloading algorithm involved in the direct and indirect
user’s impact . . . . . . . . . . . . . . . . . . . . . 63
6.3 Performance evaluation . . . . . . . . . . . . . . . . . . . 75
6.3.1 Evaluation environment . . . . . . . . . . . . . . 76
6.3.2 Application selection probability . . . . . . . . . . 80
6.3.3 QoS values according to the user’s social impact . 81
6.3.4 Influence of social context weighting factor . . . . 84
6.3.5 Influence of objective function weighting factor . . 87
6.3.6 Influence of QoS weighting factor . . . . . . . . . 90
6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7 Conclusion 93
References 96
