Turin 2006 Winter Olympics

Let us consider a possible deployment scenario for the provisioning of wireless entertainment services to practically show the motivations of the REDMAN middleware. Suppose to be involved in the organization of a wide-scale sport event, such as Turin’06 Winter Olympics, with several thousand supporters and official delegates distributed in the Olympic area, many of them carrying personal wireless-enabled devices, from Wi-Fi digital assistants to Bluetooth phones. Suppose to be interested in providing that large number of people with entertainment services while they are waiting for their local sport event, e.g., to enable them to access results, pictures, and short multimedia streams about other competitions, either occurred in the previous days or occurring at the same time in other Olympic areas, while in the interval between the first and the second manche of the giant slalom. In other words, the goal is to augment the fruition of the whole Olympic event to the paying public that attends a specific competition and can access additional entertainment services only because of its presence in the event area.

In deployment scenarios such as stadiums, a huge number of possibly mobile users are co-located in a relatively small area, where it is possible to think about the installation of an infrastructure of wireless Internet access points. In this case, entertainment service provisioning represents a technical challenge mainly because of scalability issues. The exploitation of multimedia streaming/adaptation servers running in fixed Internet hosts and wireless access points to provide spectators with Internet connectivity is not viable. Only to mention a basic example of scalability issue, research activities have shown that, to achieve usable performance for multimedia distribution, the number of concurrent clients for a single access point should be largely less than one hundred for IEEE 802.11 and less than 6 for Bluetooth [2]. In Turin’06 alpine skiing stadium (about 10,000 seats), this would require installing several hundreds/thousands of Wi-Fi/Bluetooth access points, no longer useful after the end of the Olympics.

In other kinds of deployment scenarios such as alpine ski run environments, several mobile users are distributed in a quite large spatial region, typically with no possibility to cover the whole area with wireless access points. In this case, in addition to scalability concerns, the large and mountain-type area makes definitely unpractical the installation of even a few access points with the required power lines.

In both categories of environments, spectators could be interested in requesting different types of “official” information provided by the Olympic organization (competition results, moving/still images related to preferred athletes, …) and also in sharing “unofficial” multimedia data directly collected on the competition fields by other onlookers (digital pictures, audio registrations, …), possibly by having these data dynamically adapted to the hardware/software characteristics of requesting clients, e.g., their screen resolution and color depth.

The REDMAN lightweight middleware aims at supporting service provisioning, in the above challenging environments, to wireless clients that do not need any statically deployed network infrastructure. REDMAN should be in charge of autonomously disseminating replicas of common interest resources among the spectators’ wireless nodes and of dynamically ensuring their accessibility regardless of user mobility, seamlessly from the point of view of application developers. For instance, the Olympic organizers should have the possibility to distribute new still/moving images of just-ended events at any time by injecting one copy and by specifying the desired replication degree; spectators’ wireless devices should access the requested resources by efficiently retrieving them among their peers. Similarly, the middleware should support the dissemination of spectator-provided resources of potential interest when other spectators agree on offering their devices for collaborative replication. The potential advantages of such a middleware are manifest: it would enable the provisioning of scalable services that do not require any statically deployed network infrastructure and are entirely wireless network-centric by only exploiting the computing/communication resources of spectators’ portable devices. The technical challenge is to design and implement original replication solutions that fit the above wireless environments, i.e., lightweight, highly decentralized, with very limited network/computing overhead, and with feasible performance.