none
Is there a limitation on how many PU and Local APPC LU per Host integration Server ? RRS feed

  • General discussion

  • Is there a limitation on how many PU and Local APPC LU per Host Integration Server ?
    Thursday, November 22, 2012 6:16 AM

All replies

  • Take a look at the following as it should answer your questions:

    246685 Maximum number of supported objects that are allowed in SNA Server and in Host Integration Server
    http://support.microsoft.com/kb/246685/EN-US

    A host connection in HIS Server can be equated to a PU since the connection is configured to connect to an IBM Physical Unit (PU).

    Thanks...


    Stephen Jackson - MSFT

    Thursday, November 29, 2012 8:46 PM
  • Thanks for the link. I am still confused by the difference between a node and a connection. What are their difference ?

    Thursday, November 29, 2012 11:54 PM
  • A Node is an instance of the SNA Server service (snaservr.exe). A connection equates to a PU in VTAM (for host system connections) on which you have the LUs defined. For a Peer Systems connection, the connection defines the entry point into the APPN network.

    Thanks...


    Stephen Jackson - MSFT

    Friday, November 30, 2012 4:01 PM
  • Thanks for the link. I am still confused by the difference between a node and a connection. What are their difference ?

    Peter

    > I am still confused by the difference between a node and a connection.

    Of course you are!!!

    Unfortunately these Microsoft types utterly refuse the dribs an drabs of education I throw their way from time by means of posts in this forum and so they disseminate
    confusion about SNA entities (and VTAM entities) to their poor long-suffering customers. What particularly riles me is the scorn they pour on my attempts to persuade them - indirectly since I will have been addressing one of these long-suffering customers - to think about what they pretend to be describing!!!

    I'll try to help you - and, if Stephen Jackson and his colleagues were not so arrogant, I'd be helping them also. But they are beyond help.

    > I am still confused by the difference between a node and a connection. What are
    their difference?

    We talk a lot about networks do we not? Think of a real net. It is made up of a
    flexible material which I will simply describe as a thread. One strand of tread is
    tied to another strand of thread repeatedly in order to crate the network. The
    points where two strands of thread are tied together are usually called knots.

    Turning to SNA and IP (and OSPF RFC 2328 if you want a more rigorous academic
    description although again, the terminology is different), a "node" is the knot and
    a "connection" (or "link") is the thread between adjacent knots. Where the thread
    enters the knot is a "link station" in SNA and an "interface" in IP.

    One possible confusion with this description is that this net analogy does not
    cover LANs (or the IP network used as a medium for SNA when using Enterprise
    Extender or IP-DLC). A LAN (and IP network) allows an SNA connection to be
    established dynamically - as if a piece of thread was tied between two knots when
    one of other of the knots decided it would be useful and untied when it was not
    longer needed perhaps. Recall that SNA, unlike IP, is "connection-oriented".

    If you read RFC 2328 you will see that OSPF has to abandon its simple principles
    and perform some twists and turns in order to cope with LANs.

    -

    >> A host connection in HIS Server can be equated to a PU since the connection is
    configured to connect to an IBM Physical Unit (PU).

    Believe me, when I read this I immediately thought "rubbish". If only Stephen Jackson had paid the slightest attention to my frequent clarifications over what "PU" in SNA and VTAM really means, he wouldn't be setting this trap for you. But arrogance and a mocking attitude has its price!

    Thus I wasn't in the least surprised when reading the next post and discovering that you had become confused.

    Let's simply try and explain what this statement is trying to say in terms which
    make SNA and VTAM sense.

    A connection between HIS and VTAM is represented in VTAM using a PU statement for historical reasons dating back to the first versions of VTAM which supported only peripheral nodes - such as 3270 controllers. It was an unfortunate design decision that the VTAM developers made back in the mid-1970s to load onto the PU statement a dozen or so operands which had nothing whatsoever to do with the SNA PU entity - but everything to do with the SNA "adjacent link station" entity. In other words, these operands described to one node, VTAM (indirectly to NCP where appropriate, always in the early days), how to establish, manage and possibly terminate the connection (or link) to the adjacent peripheral node.

    Strict honesty - and I try always to be strictly honest! - there is one, not even
    full, operand value which relates to a bit which can appear in the DACTPU request.
    This is the value of the second suboperand of the DISCNT operand of the PU
    statement which can have the value "F" or "NF" for "final" and "not final". It is a
    matter for the implementation of whatever machine supports the SNA PU entity how it
    reacts the the "final" bit. I know only of the 3770 range of multimedia
    workstations - some supporting RJE - of the late 1970s and early 1980s which turned
    themselves off.

    This "final" bit is the only connection - if you'll excuse the play on words - in the operands of the PU statement between the VTAM PU statement and the SNA PU entity. Now you know why I condemned the VTAM developers for their gross mistake in being lazy in not inventing an "ALS" - for "adjacent link station" - statement. We - particularly your Microsoft specialists - have been suffering ever since VTAM and NCP used PU statements in order to provide parameters for connections between NCPs. These were the first cases of there being no sign of an SNA PU entity when setting up connections using PU statements.

    Today, in the context of HIS, the use of PU statements which have nothing whatsoever to do with SNA PU entities applies to the connections between the APPN
    "Branch Network Node" concealed within the "IP-DLC Link Service" and VTAM. On
    the other hand the VTAM PU statement (and the definitions which create an SNA PU
    entity within HIS) to be used on the logical connection supported by the DLUR-DLUS
    "pipe" sessions do actually involve an SNA PU entity, necessarily SSCP-dependent
    just like the SNA PU entity I described initially, with its subordinate SSCP-dependent SNA LU entities defined with VTAM LU statements (and the definitions which create SNA LU entities within HIS). However, it is still only the "final" bit which might - it doesn't, as far as I know, I should have tested it once just to be sure - provide a parameter relevant to the SNA PU entity, the remaining operands, essentially just the ANS and DISCNT operands, refer to a operational characteristics of the logical DLUR-DLUS connection.

    Please post again if there is any of this with which you need more help. For my sins, I used to teach this stuff from the mid-1980s to the millennium having studied it intensively from just about the dawn of SNA. I remember clearly the moment I realised that the PU statement had become a thorough misrepresentation and the time I worked out that, in some cases, there was still that "final" bit as a totally inadequate excuse!

    Chris Mason

    Friday, November 30, 2012 5:30 PM
  • Thanks

    Saturday, December 1, 2012 6:48 AM