Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Providing a Single-Node Image for Connection to a Multi-System Network Node

IP.com Disclosure Number: IPCOM000111341D
Original Publication Date: 1994-Feb-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 4 page(s) / 193K

Publishing Venue

IBM

Related People

Truty, GL: AUTHOR [+3]

Abstract

A method is disclosed for defining a network node as a single node, even when the node is internally comprised of multiple systems, each of which owns alternative communication paths to the rest of the distributed network. Without this method, a multi-system node must be defined in other nodes as a complex set of connection records describing the various constituent systems of the multi-system node.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 29% of the total text.

Providing a Single-Node Image for Connection to a Multi-System Network
Node

      A method is disclosed for defining a network node as a single
node, even when the node is internally comprised of multiple systems,
each of which owns alternative communication paths to the rest of the
distributed network.  Without this method, a multi-system node must
be defined in other nodes as a complex set of connection records
describing the various constituent systems of the multi-system node.

      Each member system of the multi-system node appears as member
1.  This eliminates the need for member connections by not
externalizing the fact that it is a multi-system node.  In order to
accomplish this, there are three basic problems that need to be
solved:

1.  Determining primary trunks for connections from the multi-system
node to the rest of the network - When no single-node image is
needed, the individual member numbers in the connection records
distinguish the connections from different members.  It is therefore
only necessary to track primary/secondary connections within a single
member; no coordination among the members is necessary.  With this
method, it is more difficult to discover which trunks are to be used
when connecting to this multi-system node.

2.  Sequencing events within the multi-system node - Again, if a
single-node image is not needed, individual member numbers
distinguish events.  So a simple counter within each member sufficed
for sequencing events.  This is more difficult to provide when a
single-node image is to be  provided.

3.  Processing failed members within the multi-system node - When a
member failed, each member sent out a "subtract" record indicating
that the inter-member connection failed.  Such an indication is no
longer meaningful if a single-node image is to be provided.

      The connection between two nodes in a Network Job Entry (NJE)
network is based on an exchange of sign-on records between the two
nodes.  One node sends an initial sign-on ("I") record to the other
node, which then responds with a response sign-on ("J") record.
These records are used to exchange information about the nodes, such
as the node names and the features supported by each.  They are also
used to communicate information about the connection upon which both
nodes must agree, such as a sequence number (Connect Event Sequence,
or CES), and whether the connection is a primary or secondary trunk.

      To ensure that the primary/secondary trunk indicator is set
properly and that the sign-ons are sequenced properly, the member
systems within each node must internally exchange records to
determine the status throughout the multi-system node.  The rules for
this record exchange are as follows:

o   If the member already knows about an active trunk for this
    connection, either from this member or another member, the
    connection should be secondary.  Since sequence numbers are not
    used for secon...