Mobile

2 pages
3 views

OnMove: a protocol for content distribution in wireless delay tolerant networks based on social information

of 2
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
OnMove: a protocol for content distribution in wireless delay tolerant networks based on social information
Transcript
  OnMove: A Protocol for Content Distribution in WirelessDelay Tolerant Networks based on Social Information ∗ RubénCuevas † ,EvaJaho ‡ , CarmenGuerrero † , andIoannisStavrakakis ‡† Dept. of Telematic EngineeringUniversidad Carlos III de Madrid { rcuevas,guerrero } @it.uc3m.es ‡ Dept. of Informatics and TelecommunicationsNational and Kapodistrian University of Athens { ejaho,ioannis } @di.uoa.gr ABSTRACT We present OnMove, a protocol for content distribution inwireless delay tolerant networks for use by handheld de-vices. To improve content distribution, OnMove exploits so-cial characteristics (social similarities and physical encoun-ters) between individuals. We motivate the problem and de-scribe a content sharing protocol based on a ranking algo-rithm that exploits the social and networking characteristicsof individuals. 1. INTRODUCTION In the last years, most of the handheld devices such asPDAs or mobile phones have been equipped with differ-ent wireless interfaces such as Bluetooth or WiFi. Fur-thermore, these devices have tremendously augmentedtheir storage capacity. Both facts make possible a newcommunication paradigm: the  Delay Tolerant Content Distribution  , where individuals carrying these devicescan obtain a given content at a certain time, store itand forward it to other individuals at a later time.In this paper we envision a scenario for content distri-bution in wireless delay tolerant networks (DTNs) froma social perspective. For this we use the concept of   ego-centric networks   [2] which describes the ties that a spe-cific individual has in its immediate locality 1 . In wire-less DTNs these egocentric networks are dynamic and ∗ This work has been supported in part by the EU funded CONTENTNoE (FP6-IST-038423); NKUA’s work has also been supported inpart by IST-FET project SOCIALNETS (FP7-IST-217141); UC3M’swork has also been supported in part by CAM project BIOGRIDNET(S-0505/TIC-0101) and MEC project CONPARTE (TEC2007-67966-C03-03/TCM). 1 In wireless DTNs, a locality is formed by the individual and allother individuals in its range of coverage. Permission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for profit or commercial advantage and that copiesbear this notice and the full citation on the first page. To copy otherwise, torepublish, to post on servers or to redistribute to lists, requires prior specificpermission and/or a fee.ACM CoNEXT 2008 Student Workshop, December 9, 2008, Madrid,SPAINCopyright 2008 ACM 978-1-60558-264-1/08/0012 ... $ 5.00. 1 ACBG HD EFAFGC H DB Et=Tt=T  2 Figure 1: The egocentric network of individual A is formed by ( A , B , C  ) at instant T  1 . Eventually,  C  moves and the egocentric network of   A  becomes( A , B ) at instant  T  2  >T  1 . change as the individuals move. Figure 1 graphicallyillustrates this notion.In our approach, every individual has two types of re-lationships with other individuals: relationships basedon  physical encounters   (i.e., individuals are in the samelocality) and on  social similarities   (i.e., same profession,hobbies, interests, etc). Each individual can collect theinformation about the pattern of physical encounterswith other individuals [3], as well as the social simi-larities. This information will be helpful in the futureto decide how to upload/download contents to/from agiven individual.We present OnMove, a novel protocol that exploitsthe social information (pattern of physical encountersand social similarities) for content distribution in wire-less DTNs. The key point of this system is the selec-tion of the individual to interact with (i.e., upload ordownload content). For this purpose we design a sim-ple ranking algorithm that allows to select the “best”individual based on some parameters described in thenext section. 2. ONMOVEDESCRIPTION OnMove individuals consider the following parame-ters to create a ranking of the known individuals: Social Similarity (SS) : Each of the OnMove individ-uals has a social profile and exchanges it with other  individuals. This social profile contains social details(such as profession, interests, hobbies) about the indi-vidual. It must be highlighted that in order to identifyindividuals across different sessions, OnMove individu-als possess a User-ID that can be a nickname, a largerandom number, etc. Pattern of Meetings (PM) : The pattern of encountersbetween two individuals is defined by the frequency andthe duration of these encounters. Connection Quality (CQ) : Some features can be mea-sured during encounters. We are mainly interested inthe type of used technology (e.g., WiFi, Bluetooth),available bandwidth and the connection stability (ab-sence of connection interruptions). Content Accuracy (CA) : This shows how well alignedare the previously downloaded contents to the indi-vidual’s social profile. Thus, it allows each individu-al to identify those individuals that provide interestingcontents. Egocentric Betweenness  : The egocentric betweenness B i  of an individual  i  is the number of pairs of neigh-bors of   i  that are not directly connected to each other.Individuals with high value of egocentric betweennesshave a lot of influence in the network as a lot of otherindividuals depend on them to make connections withother people [1]. Average Egocentric Betweenness  : Since wireless DTNsare dynamic environments, we use the  average egocen-tric betweenness   B i  to denote the average value of ego-centric betweenness of an individual along the time.This is expressed in (1). In a nutshell,  B i  defines theability of   i  to disseminate contents. B i  = 1 T  T   t =1 B i ( t ) (1)We will use a ranking metric for each individual whichwill be a weighted average of the above parameters. Theweights (assigned on a scale between 0 and 1) for eachparameter will be related to its importance for the spe-cific application scenario (see also Sect. 3). It must benoted that the ranking of the individuals depends onthe dynamic parameters B  and B  (individuals exchangethem in real time). Thus, the rank of an individualshould be also a dynamic parameter.Finally, we explain the content exchange process ina given locality. Each individual sends a content queryto the highest ranked individual in the locality havingan interesting content to the individual. This querycould be accepted or rejected by that individual. If itis accepted the individual starts the download process,otherwise the individual selects another individual andrepeats the process. On the other hand, individuals inthe locality may receive several content queries from dif-ferent individuals. Then, the individual selects amongthe individuals, the one with the highest position in theranking and accepts his query while rejecting the otherqueries. 3. APPLICATIONSCENARIOS OnMove can be applied in different scenarios. In thissection we briefly describe two realistic cases:  advertise-ment platform   and  file sharing on the road  .In the  advertisement platform   the final objective is toachieve the maximum dissemination of the advertisedobject (e.g. photo, video). Thus, the OnMove rankingalgorithm must be configured giving higher weights to B  and B  parameters. However, the advertisement mustbe delivered to those individuals interested in it. Thus,the social similarity is also an important parameter inthis case. On the other hand, the previous pattern of meetings is not very relevant.In the  file sharing on the road scenario , the individ-uals retrieve small contents (news, small videos, etc) tobe consumed on their daily trips from home to work,etc. In this case, individuals are selfish, i.e., they areinterested in downloading contents satisfying their so-cial profile.  B  and  B  become less relevant and thus,the ranking becomes more static. In this situation,the individuals could be interested in downloading con-tents that satisfy the social profile of their best rankedindividuals. 4. CONCLUSIONANDFUTUREWORK This paper introduces OnMove, a novel protocol forcontent distribution in wireless DTNs populated by hand-held devices. OnMove exploits social relationships amongindividuals in order to improve content distribution.The core algorithm running on OnMove individuals is aranking mechanism. The ranking parameters vary de-pending on the application making OnMove an adaptiveprotocol. In the next steps, we will extensively investi-gate the configuration of the ranking algorithm mech-anism in several application scenarios and optimize it.We also plan to analyze the social profiles available onthe current systems such as FaceBook and how to ex-port them to wireless DTNs. Finally, we will investigatethe creation of overlay networks for content distributionin wireless DTNs. 5. REFERENCES [1] L. Hossain, K. S. K. Chung, and S. T. H. Murshed.Exploring temporal communication through socialnetworks. In  INTERACT (1) , volume 4662.Springer, 2007.[2] P. Marsden. Egocentric and sociocentric measuresof network centrality.  Social Networks  , 24(4), 2002.[3] E. Yoneki, P. Hui, S. Chan, and J. Crowcroft. Asocio-aware overlay for publish/subscribecommunication in delay tolerant networks. In MSWiM  . ACM, 2007.
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x