Sirui Yang

The Content Pollution in Peer-to-Peer Live Streaming Systems: Analysis and Implications

There has been significant progress in the development and deployment of peer-to-peer (P2P) live video streaming systems. However, there has been little study on the security aspect in such systems. Our prior experiences in Anysee exhibit that existing systems are largely vulnerable to intermediate attacks, in which the content pollution is a common attack that can significantly reduce the content availability, and consequently impair the playback quality. This paper carries out a formal analysis of content pollution and discusses its implications in P2P live video streaming systems. Specifically, we establish a probabilistic model to capture the progress of content pollution. We verify the model using a real implementation based on Anysee system; we evaluate the content pollution effect through extensive simulations. We demonstrate that (1) the number of polluted peers can grow exponentially, similar to random scanning worms. This is vital that with 1% polluters, the overall system can be compromised within minutes; (2) the effective bandwidth utilization can be sharply decreased due to the transmission of polluted packets; (3) Augmenting the number of polluters does not imply a faster progress of content pollution, in which the most influential factors are the peer degree and access bandwidth. We further examine several techniques and demonstrate that a hash-based signature scheme can be effective against the content pollution, in particular when being used during the initial phase.

PKTown: A Peer-to-Peer Middleware to Support Multiplayer Online Games

Peer-to-peer (P2P) system can support service for lots of end users with little hardware investment, which suits for the efforts to lower down the service cost for multiplayer online games (MOG). In this paper, we introduce PKTown, a P2P middleware inserted into Star Craft and the network layer. PKTown captures all packets generated by Star Craft. We apply application layer multicast (ALM) to transmit the broadcast packets through overlay network. PKTown extends the LAN game experiences to players belong to different LAN. We design a k-regular random overlay network based on game specific requirements analysis. No change is made on binary code of Star Craft. Preliminary test results show that PKTown works well and supports other MOG in the same way. Our contribution is to provide a new game experience method with little hardware investment.

Anysee: Multicast-based Peer-to-Peer Media Streaming Service System

In this paper, we propose a peer-to-peer live media streaming system based on application layer multicast, called Anysee, which provides efficient live media streaming service to large scale number of users. We construct the overlay data delivery tree with high locality and low maintenance overhead. The simulation results with log traces gathered from almost 7,200 users have shown that Anysee performs excellent even in a dynamic user environment

A modeling framework of content pollution in Peer-to-Peer video streaming systems

Peer-to-Peer (P2P) live video streaming systems are known to suffer from intermediate attacks due to its inherent vulnerabilities. The content pollution is one of the common attacks that have received little attention in P2P live streaming systems. In this paper, we propose a modeling framework of content pollution in P2P live streaming systems. This model considers both unstructured and structured overlays, and captures the key factors including churns, user interactions, multiple attackers and defensive techniques. The models are verified with simulations and implemented in a real working system, Anysee. We analyze content pollution and its effect in live streaming system. We show that: (1) the impact from content pollution can exponentially increase, similar to the random scanning worms, leading to playback interruption and unnecessary bandwidth consumption; (2) content pollution is influenced by peer cooperation, peer degree and bandwidth in unstructured overlays, and topology breadth in structured ones; (3) the structured overlay is more resilient to content pollution; (4) a hybrid overlay result in better reliability and pollution resistance; (5) hash-based chunk signature scheme is most promising against content pollution.

Modeling Modern Social-Network-Based Epidemics: A Case Study of Rose

The social-network-based epidemics, such as email-based ones, have been long studied. However, few have noticed some newly emerging epidemics which especially based on portable devices. In this paper, we think of such viruses and take a representative, the Roseepidemic, for case study. We build a model with a system of differential equations and closed-form solutions for three propagation scenes correspondingly. With both theoretical and numerical analysis, we find out that (1) Rose is able to infect hosts as exponentially as the Internet-based worms do;(2) In the Internet cafe scene, it is difficult to contain Rose even with reactive recovery measures; (3) the most influential factors for Roses propagation are the amount of hosts and portable devices and the lifetime of Internet cafe machines, while the arrival rate of clients and the proportion of immune machines only affect in the print service office scene.

Worm Containment in Peer-to-Peer Networks

Recently there have been increasing attentions on the security aspects in Peer-to-Peer (P2P) networks, especially with respect to worm epidemics. However, existing worm containment mechanisms can be largely ineffective due to the long delay exhibited in worm identification and patch generation. In this paper, we propose a new worm containment algorithm based on the concept of overlay partition, which can effectively quarantine worms into small and more manageable sub-networks. This algorithm relies on a simple anomaly detection which takes effect much faster than existing schemes. With “bridge links”, the alert of an anomaly can be disseminated faster than random scanning and dedicated worms. Further, in order to deal with false positives, we propose a granularity based clustering algorithm, called CAGA, which, through extensive simulation, is shown to realize better service sustenance during unnecessary partition caused by false positives and can effectively contain fast worms. Finally, we examine the effects of topologies and network dynamics.

TOBAB: a trend-oriented bandwidth adaptive buffering in peer-to-peer streaming system

Multimedia streaming application is increasingly popular. P2P mode makes it much more suitable for large-scale users to participate into one single application. However, most effort is spent on issues such as overlay construction and content delivery. As a foundational aspect in P2P-based multimedia systems, buffer management needs more exploration and traditional measures should be refined. In this paper, a bandwidth adaptive buffer exchange strategy is proposed, which is similar to the slow start process in TCP protocol. A novel algorithm is applied in a live media streaming system, called Anysee, and has been proven resilient to the bandwidth fluctuation in P2P networks.

Ripple: Toward a Substrate Middleware for Peer-to-Peer Applications

Peers in multiple P2P applications execute overlapped operations with different implementation, like neighbour management, resource publish/lookup, delay investigation, which leads to poor interoperability and unnecessary network load. Peers contribute their resources only when they stay in same application. We propose a P2P middleware, called Ripple, to accomplish common functions in a layered architecture. Ripple manages all peers running various applications in a uniform substrate overlay network. The substrate overlay is designed with a hybrid topology. This hybrid topology does not aim at only a particular type of applications, but also enables the flexibility for various applications. Our work is an attempt to address common needs of P2P applications, which may simplify the development and improve the performance of P2P networks, as well as mitigate the sampling stress to physical networks.