Site home page
(news and notices)

Get alerts when Linktionary is updated

Book updates and addendums

Get info about the Encyclopedia of Networking and Telecommunicatons, 3rd edition (2001)

Download the electronic version of the Encyclopedia of Networking, 2nd edition (1996). It's free!

Contribute to this site

Electronic licensing info



HPR (High-Performance Routing)

Related Entries    Web Links    New/Updated Information

Search Linktionary (powered by FreeFind)

Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.

HPR is an internetworking protocol designed by IBM as an upgrade to its APPN (Advanced Peer-to-Peer Networking) protocol. It was originally referred to as APPN+, but is now officially referred to as APPN-HPR or simply HPR. IBM designed the protocol to provide the same internetworking functionality as TCP/IP. HPR was designed as a transport for legacy data over frame- and cell-based networks. In fact, IBM designed it to fully exploit frame relay and ATM. In 1999, Cisco bought IBM's networking operations and has since been pushing TCP/IP to the SNA community.

Note: RFC 2353 (APPN/HPR in IP Networks, May 1998) defines a method with which HPR nodes can use IP networks for communication.

IBM released HPR in late 1996 to work with its aging APPN protocol. HPR provides high-end routing features that its predecessor, ISR (Intermediate Session Routing), did not have. The primary difference between HPR and ISR is that HPR provides a dynamic routing service that can route around link failures. Load balancing and rate-based congestion control are also supported. HPR also provides better performance than ISR because it eliminates the need for intermediate routers along a path to get involved with detecting and recovering errors or managing the flow of data. Only the end systems perform these functions in HPR, thus eliminating excessive traffic and overhead.

HPR consists RTP (Rapid Transport Protocol), which safely reroutes data around failed nodes, and ANR (Automatic Network Routing), a routing protocol similar to source routing. ANR uses a label-swapping technique that avoids the need to examine the contents of packets to determine how the packet should be routed, as is the case with IP. The label is read by ANR routers and used to forward the packet. The techniques do not require a table lookup. Labels are determined when a session starts and attached to each packet until the session ends.

A congestion avoidance technique called ARB (Adaptive Rate-Basedadaptively meters data at the highest sustainable rates and automatically manages peak loads. HPR fully supports SNA Class of Service and priorities, allowing applications with special bandwidth requirements to have their needs met.

In 1999, Cisco bought IBM's networking operations and has since been pushing TCP/IP to the SNA community. Its SNASw (SNA Switching Services) carries SNA traffic directly across IP network without being encapsulated (DLSw requires encapsulation). Basically, SNA frames are integrated into UDP/IP packets in the routing layer of the protocol stack and traverse the network as IP packets. HPR still exists in this scheme. HPR sees the IP network as a logical link and that link can extend all the way to the SNA systems. HPR handles rerouting around failures and retransmission requests in SNASw.

Copyright (c) 2001 Tom Sheldon and Big Sur Multimedia.
All rights reserved under Pan American and International copyright conventions.