Many-to-Many Matching Games for Proactive Social Caching in Wireless Small Cell Network International Workshop on Wireless Networks: Communication, Cooperation and Competition (WNC3), 2014 Kenza Hamidouche, Mérouane Debbah Alcatel-Lucent Chair on Flexible Radio - SUP´ ELEC, Gif-sur-Yvette, France Walid Saad Department of Electric and Computer Engineering, University of Miami Speaker: Yi-Ting Ch
28
Embed
Many-to-Many Matching Games for Proactive Social Caching in Wireless Small Cell Network International Workshop on Wireless Networks: Communication, Cooperation.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Many-to-Many Matching Games for Proactive Social Caching in Wireless
Small Cell NetworkInternational Workshop on Wireless Networks: Communication,
Cooperation and Competition (WNC3), 2014
Kenza Hamidouche, Mérouane Debbah
Alcatel-Lucent Chair on Flexible Radio - SUP´ ELEC, Gif-sur-Yvette, France
Walid SaadDepartment of Electric and Computer Engineering, University of Miami
Speaker: Yi-Ting Chen
2
Outline
• Introduction• System Model• Proposed Method and Algorithm• Simulation Result and Analysis• Conclusions
3
Introduction
• To ensure acceptable Quality of Experience (QoE) for the end-users – a very dense deployment of low-cost and low-power
small base stations (SBSs)
4
Introduction(cont.)
• However, the prospective performance gains will be limited by capacity limited and possibly heterogeneous backhaul links that connect the SBSs to the core network [3].
5
Introduction(cont.)
• Distributed caching at the network edge is considered as a promising solution to deal with the backhaul bottleneck.
• Basic Idea:– Duplicate and store the data at the SBSs side
6
Prior Works
• The cache placement problem has been mainly addressed for wired networks– Especially for Content Delivery Networks (CDNs) [7]–[10].
7
Prior Works(cont.)
• The placement problem in wireless networks has been studied in [5], [6], [11]– Minimizes the expected delay for data recovery has
been proposed– Defined without considering the limited capacity of
backhaul links– Assignment of data is based on the global popularity of
videos
8
Reference
[5] N. Golrezaei, K. Shanmugam, A. G. Dimakis, A. F. Molisch, and G. Caire, “Femtocaching: Wireless video content delivery through distributed caching helpers,” in Proc. of IEEE International Conference on Computer Communications, Orlando, FL, USA, Mar. 2012, pp. 1107– 1115.
[6] N. Golrezaei, A. G. Dimakis, and A. F. Molisch, “Wireless deviceto-device communications with distributed caching,” in Proc. of IEEE International Symposium on Information Theory, Cambridge, MA, USA, Jul. 2012, pp. 2781–2685.
[7] S. Borst, V. Gupta, and A. Walid, “Distributed caching algorithms for content distribution networks,” in Proc. of IEEE International Conference on Computer Communications, San Diego, CA, USA, Mar. 2010, pp. 1–9.
[8] L. Breslau, P. Cao, L. Fan, G. Phillips, and S. Shenker, “Web caching and zipf-like distributions: Evidence and implications,” in Proc. of IEEE International Conference on Computer Communications, New York, NY, USA, Mar. 1999, pp. 126–134.
9
Reference
[9] M. M. Amble, P. Parag, S. Shakkottai, and Y. Lei, “Content-aware caching and traffic management in content distribution networks,” in Proc. of IEEE International Conference on Computer Communications, Shanghai, China, Apr. 2011, pp. 2858 – 2866.
[10] I. Joe, J. H. Yi, and K.-S. Sohn, “A content-based caching algorithm for streaming media cache servers in cdn,” Multimedia, Computer Graphics and Broadcasting Communications in Computer and Information Science, vol. 262, pp. 28–36, 2012.
[11] N. Golrezaei, K. Shanmugam, A. G. Dimakis, A. F. Molisch, and G. Caire, “Wireless video content delivery through coded distributed caching,” in Proc. of IEEE International Conference on Communications, Pacific Grove, CA, USA, Jun. 2012.
10
Main Contributions
• Developing a novel caching algorithm – Reduce the backhaul load and the experienced delay by
the end-users when accessing shared videos in Online Social Networks (OSNs)
11
System Model(Service Provider Servers)
(Small Base Stations)
(User Equipments)
Video V
Capacities: Q
12
Goal
• To produce a proactive download of video content at the SBSs level
13
Method
• Predicting users’ requests to select and cache videos– By three factors : , ,
• Hence, we formulate this caching problem as a many-to-many matching game
14
Proposed Algorithm – Phase1
• Phase1: Network Discovery• SPSs and SBSs discover their neighbors • Collecting the required parameters to define the preferences
15
Preference of the Small Base Stations
• Social Interactions: – A user is more likely to request a video, shared by one
of his friends• Users’ Interests:
– User can request a video with his interested topic irrespective of the friend who shared it
• Sharing Impact: – If a video is cached in the SBS, sharing with the user’s
friends can have an important impact on the traffic load
16
Preference of the Small Base Stations
• We define the local popularity of a video at the SBS as follows:
17
Preferences of the Service Provider Servers
• SPS would prefer to cache a video at the SBS – Offers the smallest download time for the expected
requesting UEs.• The download time:
18
Proposed Algorithm – Phase2
• SPSs define a preference list for each owned file over the set of SBSs
• SBSs define their preferences over the set of videos proposed by the SPSs.
19
Proposed Algorithm – Phase3
• First Step:• Every SPS proposes an owned video to the most
preferred set of SBSs (shortest download time)• Each SBS rejects all but the most preferred videos• Second Step:• Every SPS proposes an owned video to the most
preferred set of SBSs which have not yet rejected it• Each SBS rejects all the most preferred videos• Repeat the Second Step• Until Convergence to a stable matching
20
𝒗𝟖 ,𝒗𝟗 ,𝒗𝟏𝟎𝒗𝟓 ,𝒗𝟔 ,𝒗𝟕 𝒗𝟏 ,𝒗𝟐
= {s4, s3}
= {}
𝒗𝟑 ,𝒗𝟒
𝒗𝟑 ,𝒗𝟒
= {s3}
𝒗𝟏 ,𝒗𝟐
= {s4, s3} = {s4, s3}
21
Pairwise Stability
• Proposition 1: Offers remain open– For every video , if an SBS is contained in at step k −
1 and did not reject – then is contained in
• Proposition 2: Rejections are final– If a video is rejected by an SBS at step k– Then at any step p ≥ k,
22
Pairwise Stability
• Theorem. The proposed matching algorithm between SPSs and SBSs is guaranteed to converge to a pairwise stable matching.
23
Simulation Result and Analysis
• Setting:
K=80
M=150
N=400
Video V=100
B=80 Mbit/time
R=180 Mbit/time
24
• Compare the two algorithm:– The proposed matching algorithm (MA)– Random caching algorithm (RA)
• We compare with different values of a storage ratio β– represents the number of files that each SBS has the