With simple server-based solutions, physical server placement is crucial for keeping latency low (1). This raises the cost of service deployment compared to a simple server setup. Traditional application layer multicast (ALM) solutions, e.g., [1], [2], [6], form multicast groups, which are not optimized with respect to (2) and (3). These are conceptually tackled by context-aware [3] or parametric [7] publish/subscribe. Such generic approaches, however, do not optimize topology for latency based on interest locality. Peer-to-peer approaches for networked virtual environments, such as VON [5] and pSense [12], focus on interest management, but do not employ sophisticated event dissemination strategies. Points (5)-(7) call for an adaptive solution.