What is Rapid Transport Protocol?

Rapid Transport Protocol (RTP) is a transport layer protocol that provides the following benefits:

Flow control.
End-to-end error recovery.
Selective retransmission of lost packets.
Nondisruptive rerouting of sessions.
Segmenting and reassembly of packets.
Translation of APPN to HPR protocols (boundary function) for sessions that have both HPR and non-HPR portions.
Translation of subarea to HPR protocols for sessions that have subarea and HPR portions. This translation is limited to routes that have a:
- Non-HPR VR-based TG adjacent to an HPR route.
- Subarea portion adjacent to a one-hop HPR route where the HPR endpoints are an interchange node and an adjacent APPN node. (Some restrictions apply; for details, see Sessions traversing subarea and APPN networks.)

RTP uses an adaptive rate-based (ARB) congestion control algorithm that makes efficient use of network resources by providing a congestion avoidance and control mechanism. The basic approach used in this algorithm is to regulate the input traffic during changing network conditions. When the algorithm detects that the network is approaching congestion, it reduces the input traffic rate until these indications go away. The algorithm also allows more traffic to enter the network without exceeding the rate the receiver can handle.

At LU-LU session establishment, if a node providing the RTP function determines that at least some portion of the session route traverses one or more HPR-capable links and eventually terminates in another RTP node, it will establish an RTP connection between the two RTP nodes. Subsequent sessions can reuse existing RTP connections if the HPR portion of the route, the APPN Class of Service, and the network connection endpoints (NCEs) are the same. CP-CP sessions over APPN node-to-node connections and ATM native connections also use RTP connections.

Tip: You can use the HPRSESLM start option or the MODIFY VTAMOPTS,HPRSESLM= command (if HPRSESLM is not currently set to DISABLED) to limit the number of LU-LU sessions assigned to each RTP connection. Limiting the number of sessions on a single RTP connection can result in performance improvement by allowing concurrent traffic on multiple RTP connections.

When a VTAM® establishes a logical connection between itself and another RTP node, VTAM creates a dynamic PU to represent the connection. These PUs are easily identified because their names are always in the form CNRxxxxx (unless the DYNHPPFX start option has been used to change this default prefix).

Note: A model PU definition can be created to customize the characteristics of dynamically created PUs. Use the DYNTYPE=RTP operand for the model PU definition in the model major node.

RTP transmits data in a form known as a network layer packet (NLP). The NLP begins with a header that specifies the automatic network routing (ANR) information necessary for routing the packet through the network.

Note: The maximum data transmission size referenced by VTAM to set the maximum size of data transmitted is calculated from the maximum packet size returned by the Route_Setup signal exchanged during RTP initialization.

To prevent excessive data segmenting within VTAM's data transmission processing, carefully evaluate the TCP/IP MTU size (for Enterprise Extender) and the maximum received data size for VTAM and each node in the network.

For Enterprise Extender, VTAMs maximum BTU length the sender can receive is calculated as follows:

 
Max_Receive_Size=
MAX((MIN(TCP/IP MTU SIZE -3),32767),768)

Because excessive segmenting can occur in the effort to conform to maximum data size requirements, you should optimize the maximum received data size. Failure to do so can result in the following situations:

Inefficient TIPAC (I/O) buffer allocation
Excessive TIPAC buffer pool expand operations
Degraded data transmission performance