This study provides some new insights into the scalability of reliable group communication mechanisms using overlays. These mechanisms use individual TCP connections for packet transfers between end-systems. End-systems store incoming packets and forward them to downstream nodes in the multicast tree using different unicast TCP connections. In this paper we assume that buffers in end-systems are large enough for the storage. It is shown that the throughput of the reliable overlay group communication scales in the sense that for all multicast tree sizes and topologies, the group throughput is strictly positive provided the saturation throughputs of all unicast connections are bounded away from 0. This is in contrast with the IP supported multicast paradigm where reliable protocols have vanishing throughput when the group size tends to infinity. The scalability of packet delay and buffer occupancy is then investigated. In the absence of additional control, the occupancy of the buffer and the latency in the end-systems explode with time. It is then shown that a proactive rate throttle mechanism implemented at the source leads to finite packet latency and buffer occupancy in any end-system of the network provided certain moment conditions are satisfied by cross traffic in the routers. Some of the theoretical scalability results are based on methods stemming from statistical physics which are called hydrodynamic limits. They are validated by a set of experiments on the Internet and by simulations based on the max-plus representation of TCP which allow the handling of very large overlay networks. The paper also discusses a few important practical matters pertaining to the shaping of the trees and the required control mechanisms.