Defining and communicating between subnets

A subnet is a set of nodes identified by name to simplify NJE routing tables by reducing NCC record traffic and path manager overhead. Networks consisting of many nodes become more manageable when you group nodes into subnets.

For example, when two corporations merge, it might be easier to manage nodes from each of the two corporations as separate subnets, with access to each other through nodes attached to each other through CONNECT statements. Such nodes are called gateway nodes. The path manager presumes that if one node in another subnet can be reached, all nodes in that subnet can be reached without additional resistance.

Figure 1 provides an example of two subnets named MD and NY. In this example, NYC and WASH are the gateway nodes for each of the subnets. Figure 2 shows the initialization statements that define these subnets.

Figure 1. Example of Defining Subnets

Figure 2. Example of Subnets Defined in the Initialization Data Set

/*******  CONNECT STMT TO SUBNETS  **********************/
/*******************************************************/
CONNECT NODEA=0005,NODEB=0006,MEMBA=1,MEMBB=1,REST=10
/*******  NJE NODE DEFINITIONS  DEFINING SUBNETS ********/
/*******************************************************/
NODE(1-10) AUTH=(NET=YES),TRACE=YES
NODE(01) NAME=WHETN,PATHMGR=YES,SUBNET=MD
NODE(02) NAME=GBURG,PATHMGR=YES,SUBNET=MD
NODE(03) NAME=BELAIR,PATHMGR=NO,SUBNET=MD
NODE(04) NAME=DEL,PATHMGR=NO,SUBNET=MD
NODE(05) NAME=WASH,PATHMGR=NO,SUBNET=MD
NODE(06) NAME=NYC,PATHMGR=NO,SUBNET=NY
NODE(07) NAME=POK,PATHMGR=NO,SUBNET=NY
NODE(08) NAME=KGN,PATHMGR=NO,SUBNET=NY
NODE(09) NAME=TRYTN,PATHMGR=NO,SUBNET=NY
NODE(10) NAME=ALBNY,PATHMGR=NO,RECEIVE=JOBS,SUBNET=NY
/*******************************************************/

These definitions must be in the initialization data set of each node involved in order to define a subnet. To prevent confusion, IBM® suggests using different naming conventions for subnets than those used for nodes. For example, if all nodes in a network begin with the prefix 'ND', use a different prefix for the names of subnets in your network.

Use the SUBnet= parameter on the NODE(nnnn) initialization statement to define the subnet to which a node belongs. Then, use one (or more) of the nodes in a subnet as the gateway to that subnet. That node communicates to the gateway node in another subnet. Any node with which you establish a connection to a node in another subnet becomes a gateway node. You might want to use more than one node as the gateway depending on factors such as the number of nodes in a subnet, or the communication needs between subnets. Note, however, that the fewer gateway nodes that are used, the more manageable the communication between subnets becomes.

Because you cannot define the subnets on all nodes in a network simultaneously, you must specify them carefully. Ensure that nodes outside your subnet have your subnet defined before nodes inside your subnet define it. Then define all subnets consistently throughout the network, so that all nodes have a valid view of routing.

You can exercise control over subnets through JES2 operator commands, which allow you to:

Add nodes to a subnet or change nodes from one subnet to another by using the $T NODE(nnnn) command.
Change the static connections using the $ADD, $DEL, and $T CONNECT commands.
Display a list of nodes in one or more subnets using the display subnet(s) ($D SUBNET) command.

$D SUBNET(M*)

$HASP816 SUBNET(MD) WHETN, GBURG, BELAIR, DEL, WASH

For more information about using these commands, see Displaying information about a network and z/OS JES2 Commands.