Online Meta-Gaming Networks are Internet-based communities of (computer) game players that extend in-game functionality by focusing on the relationship between game sessions, on what happens in the meantime between game sessions, and on the relationship between games. Motivated by the popularity of these game-related networks and by the growth of their importance---many major game developers have started to operate their own online meta-gaming network---, in this work we analyze a long-term observation of XFire. Using long-term, large-scale data that we have collected, we present a high-level, marginal distribution- and time-based analysis of XFire: its global network, player activity, user-generated content, and social structure. We find that XFire is a slowly growing network whose players spend collectively in-game over 100 years, every hour. We quantify the ``hardcore''-ness of XFire players, and find that a significant fraction of them have played over 10,000 in-game hours. We also find that XFire community members are routinely engaged in the creation and consumption of game-related media, such as screenshots and videos.

The expedient design of resource allocation mechanisms, particularly in light of dynamically varying user densities, is an important building block for the provision of peer-to-peer massively multiuser virtual environments. In case of their decentralized operation, a meaningful analysis of a given mechanism is difficult to achieve. The most realistic and convincing way of carrying out an evaluation certainly - if not performed in vivo - is against real data. This paper provides a comparative evaluation of two self-organized resource allocation schemes against Second Life avatar traces. Collected from regions with different usage characteristics, the traces allow for a thorough analysis of the performance and behavior of the individual schemes under real churn and mobility which, so far, have been simulated for assessment. We analyze both the effectiveness of the resource allocation and the respective effects on the topology, concluding some general guidelines for the design of such decentralized mechanisms on geometric overlay networks.

It is well-known that networked multi resolution contents are designed and optimized to be exchanged between a provider and one or more requesters while allowing the best possible reproduction quality to all requesters. According to participants�� terminal computing resources, push or pull modes can be used to distribute the reproduction load (i.e., server scalability versus light weight clients). In the same way, peer-to-peer techniques are available to distribute the transmission load. By focusing on these two ideas, this paper raises the following question: unless it is admitted that peer-to-peer architectures allow to increase the scalability and the responsiveness of the services they support, is there a best tradeoff between multi resolution content reproduction quality and multi resolution content distribution efficiency? In other worlds, can we accurately take into account the distribution strategy at multi resolution content generation levels? The present paper does not provide an answer, but aims at achieving a first step in this direction by proposing a new dedicated C++ GPL v3 tool that serves as a NVE distributed emulation platform. This tool includes a modular multi resolution 3D histogram that can be distributed on top of centralized or peer-to-peer network architectures.

In this paper we discuss that the specific consistency and latency requirements of world state updates in Massively Multi-User Virtual Environments (MMVEs) are determined by several characteristic factors of the virtual world, i.e. their Interaction Context. As this context is highly application-dependent, MMVE developers require a simple specification scheme which allows them to specify these requirements for their specific Interaction Contexts. In this paper we present such a scheme, which we believe represents an important step towards dynamic and flexible consistency management for MMVEs.

Supporting real-world activities online has been one of the main goals in applied computer science. However, the gathering of a large crowd for a specific purpose (e.g., lecture, concert, rally), has not yet moved online. The recent concept of crowdcasting (i.e., broadcasting to an audience while engaging them with real-time feedback) may pave the way towards it, but the audience scale and feedback fidelity are still limited by bandwidth and latency constraints. Increasing the availability of such resources thus can be important to support new forms of online activities. In this concept paper, we propose Interactive Multicast Overlay Network (IMON), a scalable approach to support crowdcasting where the number of participants can be potentially large, while the information content delivered can be rich. IMON is a peer-to-peer (P2P) overlay multicast that relies on idle peer bandwidth, and self-organizing structures to support scalable crowdcasting. The main design and challenges of IMON are described for future evaluations.

Audio communication for massively multiuser virtual environments (MMVEs) comprises the provision of hundreds of thousands of users with their personal, position dependent compositions of audio streams, in real-time, at low latency, using non-realtime components connected through a heterogeneous network with diversely limited bandwidths and varying delays. In lack of functioning comprehensive solutions, users resort to detached and limited conferencing tools. In this paper, we present a peer-to-peer approach with proven latency bounds, a suggested architecture, and best practices for its implementation. We test-drove our prototype using the OMNeT++ simulator in conjunction with the DELFOI delay model. First latency results indicate feasibility for domestic scenarios.

The field of state persistency for peer-to-peer massively multiuser virtual environments is explored and categorised into super peer storage, overlay storage, hybrid storage and distance-based storage. The storage types are briefly evaluated against the identified storage system requirements of scalability, fairness, reliability, responsiveness and security. All storage types are found to be lacking in some respects and therefore a novel storage system called Pithos is introduced. The design of Pithos, which is based on a multi-tiered hybrid architecture making use of grouping is presented. Preliminary results show that Pithos addresses all of the identified requirements.