🚀 go-pugleaf

In article <1990Oct19.220518.3658@mdivax1.uucp> mdivax1!zintel () writes:
>
>Can a PU2.1 node support multiple SDLC lines to another PU2.1 node or 
>a PU4 with 2.1 B/F in the same manner that transmission groups are 
>supported between PU4 and PU5 nodes ? The SSI LU6.2/2.1 package seems
>to allow this to be defined, but I don't if it will work. I don't know
>how to define this to NCP/VTAM either. Any help would be appreciated.
>-- 

According to the IBM Info/MVS system, and GG66-0299 "VTAM V3R2
and NCP V4R3/V5R2: SNA Type 2.1 Node Support using the System/36
as an Example", the IBM units do not allow same.  

By the way, get that document if you don't already have it...

Also GG66-0283 "VTAM V3R2 and NCP....A Technical Overview"
and  GG66-3102 "   ""   ""           Installation
Considerations"

If you can get their Token Ring Architecture Reference Manual,
it has an excellent insite into the inner workings of a 2.1
Node...


(Say anything you want about IBM, they DO produce good docs...)
]

The SSI package does appear to allow multiple links even though
most IBM sources specifically state that the LEN architecture
prohibits more than a single link station connection (LAN or
WAN) between any two adjacent peer nodes.

You cannot have more than a single link between a 2.1 Node and a
Subarea--nor can you have a link between two 2.1 Nodes which are
both connected to the SAME subarea.......
(Really, the rules are too complex for E-Mail....get the
manual....)

Don't try to put more than a single link between the SSI based
Node and ANY real IBM Node...you will get in trouble routing
RU's...        

It looks like the SSI Nodes may allow multiple links, but I
would be very nervous about doing this to any other vendor's 2.1
software until either the IBM Prototype Architecture or APPN
become available architectures.... why buy trouble...?

By the way, the term NODE actually means a Path Control
Instance...   you can have as many links as you want between
BOXES as long as each has a distinct Physical Unit (approx=Node)
definition in the partner...  and nothing stops the Application
from providing the line group function...

IBM's restriction is actually noted in the 2.1 FAPL in the
section noting the LEN architecture...it has to do with the
available routing info in the LFSID which is backward compatible
with the old Boundary FID-2 Local Form Address,.,,'