IBM Networking

A. Use this configuration:

dspu host host1 xid 05dffff1 rmac 4000.3745.0001
dspu host host2 xid 05dffff2 rmac 4000.3745.0002

dspu pool hostpool host host1 lu 2 254
dspu pool hostpool host host2 lu 2 254

You have now defined 2 distinct upstream PUs and a pool of 506 logical units (LUs): 253 LUs from host1 and 253 LUs from host2.

A. Use this configuration:

dspu host HOST1 xid-snd 05dffff1 rmac 4000.3745.0001 rsap 4 lsap 12 
dspu pool POOL1 host HOST1 lu 32 36

dspu host HOST2 xid-snd 05dffff2 rmac 4000.3745.0002 rsap 4 lsap 12 
dspu pool POOL2 host HOST2 lu 42 46

dspu pu PU1 xid-rcv 05d00001 
dspu lu 2 6 host HOST1 12

dspu pu PU2 xid-rcv 05d00002 
dspu lu 2 6 host HOST2 22

dspu pu PU3 xid-rcv 05d00003 
dspu lu 2 6 pool POOL1

dspu pu PU4 xid-rcv 05d00004 
dspu lu 2 6 pool POOL2

You have now defined two distinct upstream PUs (HOST1 and HOST2) and four distinct downstream PUs (PU1, PU2, PU3, and PU4).

• PU1 is associated with HOST1 using dedicated logical units (LUs). Host LUs 12, 13, 14, 15, and 16 will be mapped to downstream LUs 2, 3, 4, 5, and 6 respectively.

• PU2 is associated with HOST2 using dedicated LUs. Host LUs 22, 23, 24, 25, and 26 will be mapped to downstream LUs 2, 3, 4, 5, and 6 respectively.

• PU3 is associated with HOST1 using pooled LUs. Host LUs 32, 33, 34, 35, and 36 are in POOL1. One LU from POOL1 will be mapped to each downstream LU (2, 3, 4, 5, and 6), but exact maps can not be determined.

• PU4 is associated with HOST2 using pooled LUs. Host LUs 42, 43, 44, 45, and 46 are in POOL2. One LU from POOL2 will be mapped to each downstream LU (2, 3, 4, 5, and 6), but exact maps can not be determined.

Note: You are really associating LUs instead of PUs, and you could configure LUs for a downstream PU associated to LUs from different hosts. For example:

dspu host host1 xid 05dffff1 rmac 4000.3745.0001
dspu host host2 xid 05dffff2 rmac 4000.3745.0002

dspu pu pu-test xid 05d0000f
dspu lu 2 2 host host1 2

!--- Downstream LU02 is matching (or corresponding)
!--- to the LU02 on the HOST1 (mainframe).

dspu lu 3 3 host host2 2

!--- Downstream LU03 is matching (or corresponding)
!--- to the LU02 on the HOST2.

A. The dspu enable-host and dspu enable-pu commands enable local SAP addresses for real interfaces. For example:

int TokenRing0
 dspu enable-pu lsap 8

int TokenRing1
 dspu enable-host lsap 12

Because remote source-route bridging (RSRB) is a virtual interface (that is, there is no interface rsrb 0), you can issue the dspu rsrb enable-host and dspu rsrb enable-pu commands to enable local SAP addresses for the RSRB virtual interface.

Decide which to use based on where your end-stations are located. There are only four choices:

1. If your upstream host is across the RSRB interface:

   

   You must issue this command:

   dspu rsrb enable-host [lsap lsap-address]
   

2. If your downstream physical unit (DSPU) is across the RSRB interface:

   

   You must issue this command:

   dspu rsrb enable-pu [lsap lsap-address]
   

3. If your upstream host is directly connected through the Token Ring:

   

   You must issue these commands:

   interface TokenRing
   
   dspu enable-host [lsap lsap-address]
   

4. If your downstream PU is across the RSRB interface:

   

   You must issue these commands:

   interface TokenRing
   
   dspu enable-pu [lsap lsap-address]
   

A. Yes, you only need to enable local-ack on the source-bridge remote-peer statement. The downstream physical unit (DSPU) determines if local-ack is enabled during exchange identifications (XIDs).

A. Cisco uses UNBIND and DACTLU. DACTPU is not sent to downstream physical units (PUs) when the upstream PU goes down. Logical Link Control, type 2 (LLC2) and system services control points (SSCP) to PU sessions between the router and the upstream or downstream PUs are maintained independently.

A. The host must set a bit in its ACTPU request corresponding to Dynamic Definition of Dependent LUs (DDDLU) support. After ACTPU +RSP is sent, the downstream physical unit (DSPU) router sends an NMVT for each defined host LU. Then, the NMVT specifies the LOCADDR for which DSPU wants to receive an ACTLU.

A. Yes, DSPU supports XID3, but DSPU does not negotiate XID3 parameters.

DSPU sends a static XID3 to an end-station and chooses the proper role: primary for downstream physical units (PUs), secondary for upstream PUs.

If the end-station does not accept the static XID3 parameters, then the end-station rejects the XID, and connection is not established.

A. Yes, DSPU supports dependent LU 6.2 sessions; that is, system services control points (SSCP) to PU and SSCP-LU sessions must be established.

DSPU does not support independent LU 6.2 sessions; independent LU 6.2 session routing is an Advanced Peer-to-Peer Networking (APPN) Network Node function.

DSPU can route dependent LU 6.2 BINDs because the BIND is received on an SSCP-LU session.

DSPU maps between downstream SSCP-LU sessions and upstream SSCP-LU sessions; and it routes the BIND accordingly.

DSPU does not really care what LU type the BIND is. The BIND is always routed based on the SSCP-LU session map.

APPN Network Nodes can route independent LU 6.2 BINDs, because it builds maps to other nodes based on LU names.

In conclusion:

• Dependent LU 6.2 can be routed by DSPU.

• Independent LU 6.2 must be routed by an APPN Network Node.

A. The client software is not really an issue. If it establishes system services control points (SSCP) to PU and SSCP-LU sessions with the host, it should work with downstream physical unit (DSPU). So far, all 3270 emulator software packages have been found to work with DSPU.

A. DSPU supports only dependent LUs—regardless of type—including dependent LU 6.2. DSPU can not process a BIND without a prior ACTPU and ACTLU sequence.

You must use Advanced Peer-to-Peer Networking (APPN) or Cisco SNAswitch to handle independent LUs.