| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 4402.1 | Children, lots of children | FARMS::LYONS | Ahh, but fortunately, I have the key to escape reality. | Mon Jan 18 1993 11:57 | 13 |
| 4402.2 | | WELTM1::CRIDDLE | Graham Criddle, DS Tech Consultant, 853-4015 | Mon Jan 18 1993 12:42 | 18 |
| Hi (again)
Thanks for the reply...
I understand that the Node entity has a lot of children, but find it strange
that MCC cannot ask a remote Node what children it has in one operation using
one logical link instead of using one logical link for each of them.
It would appear that when looking into a Node4 entity only one logical link
is used.
Why is it necessary to use many when talking to a Phase V node and only one
when talking to a Phase IV node?
Is it due to different capabilities within NML and CMIP?
Rgds,
Graham
|
| 4402.3 | QARed At priority = L | FARMS::LYONS | Ahh, but fortunately, I have the key to escape reality. | Mon Jan 18 1993 12:59 | 0 |
| 4402.4 | upgrade | TOOK::PURRETTA | | Mon Jan 18 1993 13:37 | 16 |
| The DNA5 AM *does* have a link cache as of V1.2.3, and also as of that
version the look-into algorithm was changed to essentially what Dave
described. Registration takes longer, but the look-into operation
only goes to the node for a few cases where the next level child entity
is a dna5 module which takes an instance.
You should upgrade to v1.2.3, or if you're an internal site you may want
to try out the 1.3 fieldtest kit.
To answer your question about why the AM doesn't just ask the node about
all the children, there are two reasons. First, there isn't a CMIP request
(that I know of) to do this in a single request. Even if there were,
the access module has no idea this is what is wanted. The PM or FM sends a
single request down to the AM about each child entity.
John
|
| 4402.5 | | WELTM1::CRIDDLE | Graham Criddle, DS Tech Consultant, 853-4015 | Tue Jan 19 1993 08:47 | 6 |
| John...
Many thanks for your reponse, I will look into getting 1.2.3
Rgds,
Graham
|
| 4402.6 | How to find out the names of my children | BLUMON::SYLOR | Architect = Buzzword Generator | Thu Jan 28 1993 09:22 | 31 |
| John,
All children of the node can be requested in a single CMIP operation. In NCL,
the command would be
NCL> Show Node DEC:.foo *
That would return all the attributes of all children of the node (a.k.a.
all modules).
In the CMIP message, the entity class would be
{emaprefix(1.3.12.2.1011.2) entities(1) node(2) known-class(127)}
the entity instance would have two parts,
first the internal form of the Full Name DEC:.foo
second the magic instance value known, which at the user interface is
represented by *, but which in ASN.1 is represented by
[APPLICATION 29] IMPLICIT NULL
This is described in the entity model, section 4.6 (pg 22-23) section 8.18
(pg 103-105) and in section A.1, Table 2 (pg 128-129).
Of course the problem of how we get the FMs to ask
"Who are your children?",
rather than asking,
"Do you you have a child named Alice?",
"Do you have a child named Alex?",
"Do you have a child named Ann?",
... etc ...
as it does now is a bit harder to solve.
Mark
|
| 4402.7 | All partitions too | TOOK::PURRETTA | | Thu Jan 28 1993 10:53 | 6 |
| Thanks for the info Mark. I didn't realize that.
It's also too bad that in MCC there's no concept of wilcarded
partitions. I'm aware that in CMIP we can ask in a single request
info for all partitions but can't express that across the call
interface down to the AM.
|