PGM Reliable Transport Protocol Specification

Abstract

Pragmatic General Multicast (PGM) is a reliable multicast transport protocol for applications that require ordered or unordered, duplicate-free, multicast data delivery from multiple sources to multiple receivers.

IP Multicast Requirements for a Network Address Translator (NAT) and a Network Address Port Translator (NAPT)

“…A NAT SHOULD forward IP multicast non-UDP protocols (e.g., PGM [RFC3208], RSVP [RFC2205]) from its 'outside' interface to IP multicast receivers on its 'inside' interface(s).…”

Overview of the Internet Multicast Routing Architecture

“…Most of the mechanisms are host-based and as such out of scope of this document, but one relevant from multicast routing perspective is Pragmatic Generic Multicast (PGM) [RFC3208]. It does not require support from the routers, bur PGM-aware routers may act in router assistance role in the initial delivery and potential retransmission of missing data.…”

Reliable multicast routing for software-defined networks

“…Nevertheless, it has been manifested that this simple sourcerecovery strategy is not scalable [11], [12] for multicast to support a large number of destinations, because the source node will be overwhelmed by the acknowledgement from the destinations and the required recovery traffic. To effectively address this issue, by exploiting the idea of local services via proxy servers [22], a hierarchical architecture [11], [12] with on-tree recovery nodes placed between the source and the destinations is proposed to facilitate local loss recovery, while the deployment of the recovery nodes has not been considered as an issue in the previous works since the recovery nodes only need to support temporary caching, and the literature thus mostly focused on the development of a low-overhead protocol with acknowledgement (ACK) or negative acknowledgement (NAK). Specifically, Reliable Multicast Transport Protocol II (RMTP-II) [11] properly designed the interval between subsequent ACKs to reduce the number of ACK messages from a destination to its recovery node, or from a child recovery node to its parent recovery node.…”

“…Specifically, Reliable Multicast Transport Protocol II (RMTP-II) [11] properly designed the interval between subsequent ACKs to reduce the number of ACK messages from a destination to its recovery node, or from a child recovery node to its parent recovery node. Pragmatic General Multicast (PGM) [12] proposed NAK elimination and NAK suppression to further reduce the number of NAK messages delivered. NACK-Oriented Reliable Multicast (NORM) Transport Protocol [23] incorporated forward error corrections (FEC).…”

“…To support reliable transmissions in multicast, source-based reliable multicast [10] is first introduced to directly recover the loss packets from the source, but it suffers from the scalability problem since the source is inclined to be overwhelmed because it needs to provide loss recovery for a large number of destinations [11]. To effectively address this crucial issue, a hierarchical architecture [11], [12] with on-tree recovery nodes placed between the source and the destinations is proposed to facilitate local loss recovery. The idea is similar to the web and multimedia cache/proxy servers widely deployed today, with the goal to facilitate local services and avoid load concentration on the source node.…”