I'm sorry to be a bit late with this response. I look into this group about
once a week only.
A few explanations and adjustments to the themes raised here:
The number of SSCP-dependent LUs which may be subordinate to an
SSCP-dependent PU is determined by the Format IDentifier (FID) type 2
Transmission Header (TH) addressing. This is laid out for you in the SNA
Formats manual which, in case you don't have it already can be accessed
using the following URL:
http://www.elink.ibmlink.ibm.com/publications/servlet/pbi.wss?CTY=US&FNC=SRX&PBL=GA27-3136-20
See Chapter 4 Transmission Headers (THs), FID2.
It is, however a little complicated since the simple addressing used with
peripheral nodes to subarea SNA nodes got overlaid with a more efficient but
complex scheme for LEN and APPN architectures.
In essence, one byte is used and one byte can have 256 values. Value "0" is
used in order to address the (necessarily SSCP-dependent) PU and so there
are 255 values remaining for SSCP-dependent LUs subordinate to the PU.
Thus we explain Paul's "255 Dependent LUs per PU" and "They use NAU/LOCADDR=
1 to 255."
I expect Juanma got his "254" from the traditional implementation of SNA
3270 controllers where the LU numbering of the SSCP-dependent LUs assigned
to the attached display and printer devices start at number "2".[1]
"LOCADDR" is the operand used for the LU statement in VTAM (and NCP) as you
may or may not know. "NAU" is "network accessible unit" ("network
addressable unit" when SNA was only subarea SNA). Strictly this abbreviation
should be used to describe the LU, PU and SSCP entities. Formal SNA
architecture likes to extend it also to the CP entity but I maintain this is
unnecessary since a CP, as a NAU, is only an LU which happens to use a
limited range of special mode names.
It's a while since I used the AIX SNA software but I expect that "NAU" is
used as the prompt for the LU address. In the AIX SNA software I used to
know - long, long ago - the prompt was simply "local LU address".
Now, I need to make an "adjustment" to Paul's explanation:
"There is no limit to the number of Independent LUs that can use a single
PU" because there is no PU! Simple really!
Actually, there is a limit for simultaneous use because of the requirements
of FID2 addressing as expanded for the needs of LEN and APPN - but it's a
very large number.
As initially supported by VTAM, the representation of an SSCP-independent LU
was restricted to one particular type 2.1 node adjacent link station which
was permitted to be either the source or destination of a BIND request
following LEN protocol rules. The definition syntax which the VTAM
developers used in order to represent the SSCP-independent LU simply
specified LOCADDR=0 on the LU statement following the PU statement which
defined the type 2.1 node adjacent link station.
Note that the type 2.1 node need have no SSCP-dependent resources associated
with the FID2 THs used over the link.[2] Thus there need be no actual SNA PU
entity represented by the PU statement.[3]
Later VTAM allowed the representation of the SSCP-independent LU within VTAM
also to be independent of any particular adjacent link station - a necessary
prerequisite for the dynamic selection of routes when using APPN protocols.
Thus the representation of an SSCP-independent LU within VTAM is now by
"cross-domain resource", whether created dynamically, if permitted, or by
definition of a CDRSC statement. The ALSLIST, adjacent link station list,
operand of the CDRSC statement can specify the names of PU statements - note
multiple - representing adjacent link stations from which a BIND from the
SSCP-independent LU can appear or to which a BIND to the SSCP-independent LU
can be sent.
Because a definition technique, once offered, cannot be retracted, VTAM now
shakes its head whenever it encounters an LU statement with LOCADDR=0 and
patiently converts it to a CDRSC statement with an ALSLIST operand
specifying a single name, the name of the PU statement, supposedly a
hierarchically superior SSCP-dependent PU but which we know now if we didn't
before represents only a type 2.1 node adjacent link station.
Try a VTAM DISPLAY of the name of an LU statement with LOCADDR=0 specified
and you will see what I mean.
Chris Mason
[1] I believe I have seen some probably 3174 documentation where the LU with
number 1 had a role related to file transfer (to and from ?) the control
unit.
[2] I originally wrote "associated with it" but then I realised that it
could be supporting the DLUR function.
[3] The only operand of the PU statement which actually relates to the SNA
PU entity is actually not even a full operand. It is the "F/NF" suboperand,
the second suboperand of the DISCNT operand. And all that the "F/NF"
suboperand does is to determine the setting of a bit in the DACTPU request.
Post by l***@attglobal.netPost by zhangychwhen i make a pu on a machine,how many lu can i defined max?
255 Dependent LUs per PU.
They use NAU/LOCADDR= 1 to 255.
NAU/LOCADDR=0 means an Independent LU6.2 LU.
There is no limit to the number of Independent LUs
that can use a single PU.
If you need more that 255 dependent LUs in a CS/AIX node
you can use multiple DLUR PUs or parallel links.
The details of parallel links depends on what link type
is being used (802.2, sdlc, etc).
Paul Landay