Managing social contents in Decentralized Online Social Networks: A survey
Introduction
Online Social Networks (OSNs) are nowadays one of the most popular applications in the Internet. They have attracted a huge amount of users during the last years by changing the way people communicate and interact. Facebook can be considered the most representative OSN with, at the beginning of 2018, more than 2 billion active users, and the highest number of daily users connections. OSNs provide several services [1] offering to their users the opportunity of building a public profile, looking up new friends among the registered users, establishing relationships, and sharing content. Furthermore, these platforms also allow sharing of information within groups of users and the possibility of building communities of users characterized by common interests.
One of the major problems of current OSNs, which are mainly developed on centralized platforms, concerns the privacy of the users’ data. Indeed, social data are stored in centralized servers, and the companies running the OSNs, use these data for commercial goals. In May 2015, a report1 commissioned by the Belgian Data Protection Authority based on the analysis of OSN policies and terms-of-use, concluded that Facebook gives users a false sense of control over their data privacy. More recently, it has been evident to the general public that Facebook data might have been sold without any consent of the legitimate owners. These are only few examples of several legal issues that involve not only Facebook, but also further OSNs, like Twitter, or Google+. Furthermore, centralized OSNs may suffer of other problems, like limited scalability and high maintenance costs to manage data of large number of users [2].
All these issues have led researchers to propose alternative solutions based on the decentralization of OSN services. A Decentralized Online Social Network (DOSN) [3] is an online social network implemented on a distributed platform. In a DOSN, there is no single service provider but a set of nodes that cooperate to guarantee all the functionalities offered by a centralized OSN. Decentralization gives several benefits in terms of privacy. Indeed, there is no central entity that has the control on all users’ data or changes the existing terms of service, and this gives to the users more control over their data.
Moving from a centralized to a distributed architecture gives the opportunity to develop the social network platform by exploiting different distributed models. Following the way of how people interact with each other in a social network makes P2P architectures [4] a natural way to implement DOSNs. However, several alternative solutions are possible, like exploiting network of trusted servers, or mobile and opportunistic networks. Also hybrid solutions where the user exploits both the storage of its own device and a cloud storage service are possible [3].
Researchers have presented several solutions for DOSNs in the last decade. Furthermore, recent years have seen several initiatives to implement and deploy real DOSNs. Just to provide a few examples, the precursor of current DOSNs can be considered Diaspora [5], a solution based on a federation of trusted servers, which has gained a lot of popularity, despite not fully decentralized. Other DOSNs, like Tent and Friendica, are based on similar concepts. On the other hand, Retroshare is a fully decentralized DOSN, designed to provide maximum security and anonymity to its users.
While decentralization gives interesting possibilities for increasing the privacy level of users’ data, it introduces many challenges, still to be solved. These mainly concern the management of social data in a distributed environment. With the term “social data”, we identify all data exchanged in the Social Networks concerning both information related to the users (contact details, describing the userâs identity, relationships, community memberships, etc..) and generated contents (comments, posts, etc..). One of the main problems is to guarantee the availability of social data, in an environment characterized by a high level of dynamism. Another main problem is related to the development of techniques for propagating social updates in an efficient way. Finally, even if data is no more stored on centralized servers, new privacy issues have to be solved, for instance detecting trusted nodes that may host the profile of off-line users.
This paper reviews the main solutions proposed in the literature to solve these challenges and presents a classification of the existing solutions. We describe in detail the DOSN research challenges, focusing, in particular, on the data management problem. First we introduce a classification of the main architectural solutions for DOSNs. Then, we introduce existing approaches for guaranteeing data availability, and the main techniques used to spread social content among the users of the social network.
Existing surveys or taxonomies [3], [6], [7] provide a detailed description of DOSNs, mainly from a privacy and security point of view. They address only partially other important issues, such as managing social data in a distributed environment, and do not provide a classification of existing proposals based on this characteristic. On the other hand, we focus on the data management problem in DOSNs and present current techniques for data availability, information diffusion, and data privacy in a distributed environment.
The rest of the paper is organized as follows. In Section 2, after a general overview of DOSNs, we propose a coarse classification of existing systems and we introduce the main research open issues. Section 3 presents the structures used to represent social data. Section 4 introduces current techniques to guarantee data availability and classify them, while Section 5 discusses and classifies the techniques used to spread social information. In Section 6 we describe data privacy in DOSNs. In Section 7, we present relevant examples of DOSNs. Section 8 discusses the limitations of current solutions and open problems and concludes the paper.
Section snippets
Decentralized Online Social Networks: overview and key research challenges
A general model for the decentralization of the services in an OSN is given in [8], where a DOSN is defined as a distributed system including a Social Network (SN), a Social Networking Service (SNS) and a Communication and Transport (CT) level. Using a network stack layering representation, SN is located on the top layer, SNS is the intermediate layer, and CT is the bottom one. The SN level provides the common social network functionalities, such as chat, mails, wall posts and/or tweets,
Social data representation: an overview
In this section, we briefly review how social data is generally represented. Like in OSNs, also in DOSNs a user is represented by a profile that can be considered his/her digital proxy, stored on one or more nodes. In detail, a profile represents the virtual identity of the user, and usually contains personal information, like private information (name, date of birth, town, etc.), friends list, posts (or tweets), comments, videos and photos [12]. DOSNs adopt several solutions to represent
Data availability
One of the main challenges for distributed systems in general is guaranteeing data availability when the application of the data owner is offline. Several techniques have been proposed to solve the data availability problem in the general context of distributed systems; for instance, distributed caching [17], [18], dynamic data replication [19] and erasure coding [20]. However, traditional approaches can not be easily adapted to DOSNs because of the different usage patterns [21]. Indeed, OSN
Information diffusion
Nowadays, OSNs produce a huge amount of contents and interactions. The strategy adopted to propagate social data plays a key role to deliver data to targeted users with a limited latency, and avoid message duplication. While in legacy OSNs information diffusion can be carried out by the central provider by exploiting the whole social graph, it is still an open research challenge for DOSNs.
The problem of data dissemination in distributed systems like sensor networks, P2P platforms, and mobile
Privacy and Security
DOSNs effectively address the main privacy concern of centralized OSNs, because personal data are no more under the control of a centralized provider, such as in Facebook, but are managed by the users themselves. Users do not depend on an external OSN service provider to maintain their data, and the absence of a single control point cuts out several privacy breaches. However, the decentralization of OSN services raises new privacy and security challenges that were previously addressed by the
DOSNs: current proposals
In this section we introduce a set of relevant DOSN proposals by classifying them according to the three classes introduced in Section 2: DHT-based, SO-based, or External resources-based.
Conclusions and open problems
DOSNs represent a valid alternative to centralized OSNs. However, the decentralization requires specific approaches for the social data management. This paper presents an analysis of the existing DOSNs, focusing on the techniques to manage social contents in a distributed framework. In detail, the paper highlights three major problems for data management: data availability, information diffusion, and privacy. We describe in detail the problem of data availability by introducing a classification
Conflict of Interests
None
Barbara Guidi is currently a postdoctoral researcher at the Department of Computer Science of the University of Pisa. She received her B.Sc. and M.Sc. in Computer Science from the University of Pisa, Italy, in 2007 and 2011, respectively. She received her Ph.D. degree in Computer Science from the University of Pisa, in 2015. In 2014, during her Ph.D., she was a visitor at the Heinrich Heine University of Dusseldorf. She was a Co-Chair for the conference EAI GoodTechs 2017, and she is a workshop
References (87)
- et al.
A survey on decentralized online social networks
Comput. Netw.
(2014) Egocentric and sociocentric measures of network centrality
Soc. Netw.
(2002)- et al.
Privacy and temporal aware allocation of data in decentralized online social networks
Proceedings of the International Conference on Green, Pervasive, and Cloud Computing
(2017) - et al.
Privacy, cost, and availability tradeoffs in decentralized OSNS
Proceedings of the 2nd ACM Workshop on Online Social Networks
(2009) - et al.
A comparative study of pub/sub methods in structured p2p networks
Databases, Information Systems, and Peer-to-Peer Computing
(2007) - et al.
Selective propagation of social data in decentralized online social network
Advances in User Modeling
(2012) - et al.
Social network sites: definition, history, and scholarship
J. Comput.-Med. Commun.
(2007) - et al.
A first step towards user assisted online social networks
Proceedings of the 3rd Workshop on Social Network Systems, SNS ’10
(2010) - et al.
Decentralized online social networks
Handbook of Social Network Technologies and Applications
(2010) - et al.
P2P Networking and Applications
(2008)
A taxonomy of decentralized online social networks
Peer-to-Peer Netw. Appl.
Safebook: a privacy-preserving online social network leveraging on real-life trust
Comm. Mag.
Distributed Systems: Principles and Paradigms
Efficient dissemination in decentralized social networks.
Peer-to-Peer Computing
Trusted dynamic storage for dunbar-based p2p online social networks
Proceedings of the OTM Confederated International Conferences” On the Move to Meaningful Internet Systems”
Privacy-preserving data allocation in decentralized online social networks
Distributed Applications and Interoperable Systems
DECENT: A decentralized architecture for enforcing privacy in online social networks
Pervasive Computing and Communications Workshops (PERCOM Workshops), 2012 IEEE International Conference on
Lifesocial. kom: a secure and p2p-based solution for online social networks
Proceedings of the IEEE Consumer Communications and Networking Conference (CCNC)
Confidant: protecting OSN data without locking it up
Proceedings of the 12th International Middleware Conference
Supporting cooperative caching in ad hoc networks
IEEE Trans. Mobile Comput.
A distributed internet cache
Proceedings of the 20th Australian Computer Science Conference
An efficient scheme for dynamic data replication
Technical Report
Gemstone: empowering decentralized social networking with high data availability
Proceedings of the IEEE Global Telecommunications Conference (GLOBECOM 2011)
Measuring user behavior in online social networks
Netw. IEEE
Understanding latent interactions in online social networks
Proceedings of the 10th ACM SIGCOMM Conference on Internet Measurement
Reducing maintenance overhead in DHT based peer-to-peer algorithms
Proceedings of the Third International Conference on Peer-to-Peer Computing, 2003.(P2P 2003).
The impact of user’s availability on on-line ego networks
Comput. Commun.
An empirical study of availability in friend-to-friend storage systems
Proceedings of the P2P 2011, IEEE International Conference on Peer-to-Peer Computing, 2nd, 2011
Understanding availability
Peer-to-Peer Systems II
F2box: cloudifying f2f storage systems with high availability correlation
Proceedings of the IEEE 5th International Conference on Cloud Computing (CLOUD)
Developing the cloud-integrated data replication framework in decentralized online social networks
J. Comput. Syst. Sci.
A decentralized online social network with efficient user-driven replication
Proceedings of the IEEE International conference on Social Computing (SocialCom 2012)
Revisiting content availability in distributed online social networks
CoRR
Towards robust and scalable peer-to-peer social networks
Proceedings of the Fifth Workshop on Social Network Systems
F2F: reliable storage in open networks
Proceedings of the 5th International workshop on Peer-To-Peer Systems, IPTPS 2006
Distributed coverage of ego networks in f2f online social networks
Proceedings of the International IEEE Conferences Ubiquitous Intelligence & Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Cloud and Big Data Computing, Internet of People, and Smart World Congress (UIC/ATC/ScalCom/CBDCom/IoP/SmartWorld)
Towards the realization of decentralized online social networks: an empirical study.
Proceedings of the ICDCS Workshops
On user availability prediction and network applications
IEEE/ACM Trans. Netw.
Supernova: Super-peers based architecture for decentralized online social networks
Proceedings of the COMSNETS
The many faces of publish/subscribe
ACM Comput. Surv.
Content-based publish-subscribe over structured p2p networks
Proceedings of the 4th International Workshop on Distributed Event-Based Systems
Cited by (69)
Blockchain solutions for trustworthy decentralization in social networks
2024, Computer NetworksA decentralised messaging system robust against the unauthorised forwarding of private content
2023, Future Generation Computer SystemsInformation consumption and boundary spanning in Decentralized Online Social Networks: The case of Mastodon users
2022, Online Social Networks and MediaBridging separate communities with common interest in distributed social networks through the use of social objects
2022, Future Generation Computer SystemsCitation Excerpt :Recent scandals such as the one involving Cambridge Analytica have clearly shown that technology giants (Facebook, Google & Co, etc.) have serious problems in protecting user data. In order to guarantee users privacy protection and a greater control over their data, researchers have recently proposed several distributed platforms to implement decentralized online social networks (DOSNs) [1] either based on peer-to-peer (P2P) architectures [2–5] or with a Web-based nature [6,7]. In this context, an interesting decentralized platform for social Web applications, named Solid, was proposed by Tim Berners Lee in [8].
Assessment of Wealth Distribution in Blockchain Online Social Media
2024, IEEE Transactions on Computational Social Systems
Barbara Guidi is currently a postdoctoral researcher at the Department of Computer Science of the University of Pisa. She received her B.Sc. and M.Sc. in Computer Science from the University of Pisa, Italy, in 2007 and 2011, respectively. She received her Ph.D. degree in Computer Science from the University of Pisa, in 2015. In 2014, during her Ph.D., she was a visitor at the Heinrich Heine University of Dusseldorf. She was a Co-Chair for the conference EAI GoodTechs 2017, and she is a workshop Co-Chair for LSDVE 2018, in conjunction with EUROPAR 2018.
She received two Best Paper Awards: at the International Conference DCNET 2013 and at the workshop LSDVE 2017.
She has been involved in the TPC of conferences and workshops, such as IEEE CloudCom, and has been a reviewer for journals, such as Mobile Networks and Applications (MONET), Multimedia Tools and Applications (MTAP), and Concurrency and Computation: Practice and Experience (CCPE). Her current research interests include distributed systems, P2P networks, complex networks, Decentralized Online Social Networks, dynamic community detection, and the blockchain technology.
Marco Conti is a research director and scientific counselor, for information and communication technologies, of the Italian National Research Council. He has published in journals and conference proceedings more than 400 scientific papers related to design, modelling, and experimentation of Internet architecture and protocols, pervasive systems and social networks. He has published the books “Metropolitan Area Networks (MANs)” (1997), “Mobile Ad Hoc Networking” (2004), “Mobile Ad hoc networking: the cutting edge technologies” (2013) and “Online Social Networks: Human Cognitive Constraints in Facebook and Twitter Personal Graphs” (2015). He has received several awards, including the Best Paper Award at IFIP TC6 Networking 2011, IEEE ISCC 2012 and IEEE WoWMoM 2013. He is Editor-in-Chief of Computer Communications and Online Social Networks and Media, and Editor-in-Chief for special issues of Pervasive and Mobile Computing. He served as TPC chair for several major conferences, such as IFIP Networking 2002, IEEE WoWMoM 2005, IEEE PerCom 2006, and ACM MobiHoc 2006, and he was general chair (among many others) for IEEE WoWMoM 2006, IEEE MASS 2007 and IEEE PerCom 2010.
He is the founder of successful conference and workshop series, such as IEEE AOC, ACM MobiOpp, and IFIP SustainIT.
Andrea Passarella (PhD 2005) is currently a Researcher at the Institute for Informatics and Telematics (IIT) of the National Research Council of Italy (CNR). Prior to join IIT he was with the Computer Laboratory of the University of Cambridge, UK. He has published 130+ papers on various topics, including human-centric data management for self-organising networks, Online and Mobile social networks, opportunistic, ad hoc and sensor networks. He received four best paper awards, including at IFIP Networking 2011 and IEEE WoWMoM 2013. He is General Co-Chair for IEEE WoWMoM 2019 and workshops co-chair for IEEE INFOCOM 2019. He was the PC co-chair of IEEE WoWMoM 2011, Workshops co-chair of ACM MobiHoc 2015, IEEE PerCom and WoWMoM 2010, and the co-chair of several IEEE and ACM workshops.
He is co-author of the book “Online Social Networks: Human Cognitive Constraints in Facebook and Twitter Personal Graphs” (Elsevier, 2015), and was Guest Co-Editor of the special issue Online Social Networks in Elsevier Computer Communications, and several special sections in ACM and Elsevier Journals and of the book “Multi-hop Ad hoc Networks: From Theory to Reality” (2007). He is the chair of the IFIP WG 6.3 “Performance of Communication Systems”.
Laura Ricci received the M. Computer Science from the University of Pisa in 1983 and the Ph.D. from the University of Pisa in 1990. Currently, she is an Assistant Professor at the Department of Computer Science, University of Pisa, Italy. Her research interests include parallel and distributed systems, peer-to-peer networks, cryptocurrencies and blockchains.
In this field, she has co-authored over 100 papers in refereed scientific journals and conference proceedings. She has served as a program committee member of several conferences and has been a reviewer for several journals.