Browse Prior Art Database

Flow Control Technique for Local Networks of Interconnected Token Rings

IP.com Disclosure Number: IPCOM000041400D
Original Publication Date: 1984-Jan-01
Included in the Prior Art Database: 2005-Feb-02
Document File: 3 page(s) / 60K

Publishing Venue

IBM

Related People

Bux, W: AUTHOR

Abstract

In communication networks of multiple rings with token-regulated access, interconnected by a backbone ring and by bridges between rings, message frames must be buffered in the bridges when being transferred from one ring to another. In this article, a flow control technique is suggested by which congestion can be avoided which otherwise could occur because of limited buffer capacity in the bridges. Fig. 1 shows the multiple ring network. It comprises several token rings R1...R3 to which a multiplicity of stations ST is attached. The three outer rings are interconnected by bridges B1...B3 and a backbone ring BB. The backbone ring has interface circuitry 11 in each bridge, and each outer ring has interface circuitry 13 in one bridge.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 3

Flow Control Technique for Local Networks of Interconnected Token Rings

In communication networks of multiple rings with token-regulated access, interconnected by a backbone ring and by bridges between rings, message frames must be buffered in the bridges when being transferred from one ring to another. In this article, a flow control technique is suggested by which congestion can be avoided which otherwise could occur because of limited buffer capacity in the bridges. Fig. 1 shows the multiple ring network. It comprises several token rings R1...R3 to which a multiplicity of stations ST is attached. The three outer rings are interconnected by bridges B1...B3 and a backbone ring BB. The backbone ring has interface circuitry 11 in each bridge, and each outer ring has interface circuitry 13 in one bridge. Bridge control units 15 control the transfer of messages from one ring to another and contain the necessary buffer storage. Messages are transferred in the form of frames, as shown in Fig. 2. This format corresponds to existing standards proposals. Contents of each frame are: SD = Start Delimiter, AC = Access Control Field, DA = Destination Address, SA = Source Address, INFO = Information Field, FC = Frame Check Sequence, ED = End Delimiter. The end delimiter consists of two bytes of which one carries a code violation and an error indication. The other eight bits are used for signalling. The two bits "A" and two bits "C" can be set to "1" by the destination station to indicate that it recognized the address and copied the frame, respectively. It is now suggested to use the two bits designated "X" as flow control bits FC1 and FC0. The proposed flow control mechanism operates as follows: 1) In every bridge, the buffers for frames to be forwarded from an outer ring to the backbone

ring are grouped into individual pools,

one for each destination ring (or, equivalently,

destination bridge). Note that this separation

need not result in a separation in hardware,

but would rather be done in microcode by the

message processors in the bridge ring adapters. The

frames received from the backbone ring are being buffered

in an extra buffer pool. 2) For transmission on the backbone ring, bridges schedule frames waiting in their buffer pools

in a cyclic fashion. 3) The FC0 and FC1 bits are set to zero by the

transmitting "source" bridge. A "sink" bridge,

i.e., a bridge copying a frame from the

backbone ring which is destined to a station on its

outer ring, indicates its level of congestion by

setting, in the frame passing

through, the flow control indicators FC0 and FC1 to an

appropriate value. FC0 = FC1 = 0 means n...