This document describes the various layers and control blocks of SNA and BCOM buffers. The intent of this document is not provide a complete SNA course, it is far too complex for the scope of this document. It concentrates on the basic underpinnings of SNA and their relationship with BCOM.
SNA Fundamentals
The following diagram will be used in the course of the discussion.
Table 1 - SNA layers and Data Structures
Note 1: We can say that the RU corresponds to BCOM's MAXBLK. When there is an FMH, the buffer space available for BCOM data will be less than the RU size.
Note 2: The diagram depicts an SDLC frame at the DLC level. For other medium, there is an equivalent DLC-level frame.
The SNA Layers
SNA, like most networking protocols, is implemented in a layered manner. The seven layers implemented in SNA are:
Physical Layer. It represents the physical interface with the transmission medium. It deals with cables, connectors and modems. SNA does not define this layer;
Data Link Control. Deals with the management of the link and the error free transmission of bit-streams across the link. SDLC, DLC (for Token-Ring) and S/370 I/O channels are implementations of this layer;
Path Control. Routes data through the network, manages routes and congestion. Segmentation (see later) is handled at this level;
Transmission Control. Provides for checking of messages sequence numbers. Pacing is implemented at this layer. SNA encryption and decryption is also performed here;
Data Flow Control. Controls the flow of data within a session. This layer implements chaining (see later), session responses, session flow control and session sequence number assignment;
NAU Services. Network Addressable Unit Services provide session presentation services (data formatting considerations), programming services, transaction services and the like;
End User. End Users of the network or SNA applications (TPs). BCOM is an SNA application;
SNA Data Structures
Request/Response Unit (RU). For the purpose of this basic discussion, we can assume that the RU is the data (BCOM buffer). The RU relates to the unit of communication between LUs;
Function Management Header (FMH). The header is an optional part of the RU and is used to convey control information between session partners. BCOM uses FMH-5 (LU 6.2 ATTACH) with the HANDSHAKE (x'01') message to initiate execution of BCOM's Server on another platform. FMH-7 can also return error information from a LU 6.2 partner.
Request/Response Header (RH). The Request/Response Header describes the following type of RU: user-data request, SNA command, response RU. The RH contains such information as Request/Response, Change Direction (CDI), Bracket Protocol (BB, EB, CEB) and indication of the presence of an FMH. The RH contains also the information on chaining (see later). The combination of the RH and the RU is called a Basic Information Unit (BIU).
The size of the RH is always 3 bytes.
Transmission Header (TH). Contains routing information used for transport purposes. The Transmission Header (TH) contains segmentation information (see later). The combination of the TH, the RH and the RU (including the FMH, if there is one) is called a Path Information Unit (PIU). The PIU relates to the unit of communication between PUs. When PIUs are passed down to the Data Link Control layer, they could be blocked (many PIUs) to form a Basic Transmission Unit (BTU).
The size of the TH varies in from 2 to 29 bytes depending on the TH type.
SDLC Header and Trailer. A BTU is enclosed within an SDLC header and a trailer. This forms an SDLC frame. The same principle applies to non-SDLC (Token-Ring, Ethernet, etc.) medium.
The length of the SDLC headers and trailers is irrelevant to the following discussion and can be ignored.
Some SNA concepts
Chaining. Chaining is an LU to LU protocol that is used to size the messages down to the destination LU buffer size (RUSIZE). Chaining is done automatically by SNA. Request Headers (RH) contain the chaining information. There are restrictions on the usage of Chaining with BCOM: As BCOM does its own chaining, MAXBLK must always be less or equal to the RUSIZE to prevent automatic chaining by SNA.
Segmentation. Segmentation is a protocol that is made to ensure that PIUs do not exceed the size of the buffer in the receiving PU. Segmentation is done automatically by SNA. Transmission Headers (THs) contain the segmentation information.
Pacing. Session pacing supplements the Chaining Protocol. Its objective is the same: prevent overflowing of the receiving LU's buffer. The pacing parameter (PACING=m) indicates the number of chain elements that will be sent by a node before requesting a pacing response. This regulates the flow of data transfer and makes sure the receiving LU is not overflows. PACING=m applies to data transfer between the NCP and the PU. VPACING=n applies to data transfer between VTAM and NCP.
HOME |
ABOUT US |
PRODUCTS | SUPPORT
| CONTACT US | SITEMAP
Copyright 2003. ETI-NET. All Rights Reserved. All content copyright
their respective owners.
Site Design by
MissionECommerce. Last Updated:
03/23/2006 01:21 PM MST (GMT -0500)