Abstract-Online social networks (OSN) have attracted millions of users. This enormous success is not without problems; the centralized architectures of OSNs, storing the users' personal data, provides ample opportunity to privacy violation. These problems have raised the demand for open, decentralized alternatives. We tackle the research question: is it possible to build a decentralized OSN over a social overlay, i.e., an overlay network whose links among nodes mirror the social network relationships among the nodes' owners? This paper provides a stepping stone to the answer, by focusing on the key OSN functionality of disseminating profile updates. Our approach relies on gossip protocols. We show that mainstream gossip protocols are inefficient, due to the properties that characterize social networks. Therefore, we leverage these very same properties towards our goal, by appropriately modifying the forwarding rules of gossip protocols. Our evaluation, performed in simulation over a crawled realworld social network, shows that our protocols provide acceptable latency, foster load balancing across nodes, and tolerate churn.

Abstract-Peer-to-peer systems based on an overlay network that mirrors the social relationships among the nodes' owners are increasingly attracting interest. Yet, the churn induced by the availability of users raises the question-still unanswered-of whether these social overlays represent a viable solution. Indeed, although constraining communication to take place only among "friends" brings many benefits, it also introduces significant limitations when healing the overlay in the presence of churn.This paper puts forth two contributions. First, we show through simulation on real datasets that churn induces relevant delays in information dissemination, which may ultimately hamper the practical application of social overlays. Yet, identifying opportunities for improvement and evaluating design alternatives through simulation is impractical, due to the size of the target networks, the large parameter space, and the many sources of randomness involved. Therefore, in our second contribution we combine analytical and simulation techniques to enable the estimation of dissemination delays at a practical cost.

No abstract

Abstract-Decentralized social networks are an emerging solution to the privacy issues plaguing mainstream centralized architectures. Social overlays-overlay networks mirroring the social relationships among node owners-are particularly intriguing, as they limit communication within one's friend circle. Previous work investigated efficient protocols for peer-to-peer (P2P) dissemination in social overlays, but also showed that the churn induced by users, combined with the topology constraints posed by these overlays, may yield unacceptable latency.In this paper, we combine P2P dissemination on the social overlay with occasional access to the cloud. When updates from a friend are not received for a long time, the cloud serves as an external channel to verify their presence. The outcome is disseminated in a P2P fashion, quenching cloud access from other nodes and, if an update exists, speeding dissemination. We show that our protocol performs close to mainstream centralized architectures and incurs only modest monetary costs.

We present Sedano, a system for processing and indexing a continuous stream of business-related news. Sedano defines pipelines whose stages analyze and enrich news items (e.g., newspaper articles and press releases). News data coming from several content sources are stored, processed and then indexed in order to be consumed by Atoka, our business intelligence product. Atoka users can retrieve news about specific companies, filtering according to various facets. Sedano features both an entity-linking phase, which finds mentions of companies in news, and a classification phase, which classifies news according to a set of business events. Its flexible architecture allows Sedano to be deployed on commodity machines while being scalable and fault-tolerant.