Reusable Server Kernel

Problem 41 (Sug): APPC Worker LD

=============== 2000-11-07 11:25:32 (-05.00) ===============
 
Name:        ** Suppressed **
Company:     ** Suppressed **
Telephone:   ** Suppressed **
E-mail:      ** Suppressed **
 
Text:
 
Hi, Brian.
Yes, please give me a call at your convenience and we can
discuss this in more detail.....
dj
=============== 2000-11-07 10:23:21 (-05.00) ===============
 
Name:        ** Suppressed **
Company:     ** Suppressed **
Telephone:   ** Suppressed **
E-mail:      ** Suppressed **
 
Text:
 
Hi Dave,
You make some good points.  Do you have a product in mind
that is being barred by lack of this support?  I'd like to
hear more about your business case.
Brian
=============== 2000-11-04 12:13:17 (-05.00) ===============
 
Name:        ** Suppressed **
Company:     ** Suppressed **
Telephone:   ** Suppressed **
E-mail:      ** Suppressed **
 
Text:
 
The RSK's support of remote worker machines is proving itself
to be a very useful concept. However, the communcation channel
between the wrorker and the RSK master, IUCV, is a source of
some problems. Would it be possible to replace the IUCV worker LD 
with one based on APPC?
Our reasons for asking for this are the following:
1) APPC is supported across a multitude of IBM platform, while
IUCV commincation is supported only on VM. APPC would allow
RSK-controlled worker machines to execute on, say, OS/2
workstations.
2) APPC would allow worker machines to reside on other, separate
VM systems, from the RSK master. IUCV also allows for comminication
betwwen separate VM images, but requires extra software like
PVM to do the actual communication function. This adds an extra layer of 
cost and complexity to such applications.
3) VM already provides a well documented and tested CSL library
for coding APPC aware applications. There is no such library
for IUCV, forcing each user to either develop their own CSL
for IUCV or code the worker application in Assembler, using native 
IUCV macros. 
4) While there are many valid reasons for requiring that RSK
servers be coded in assembler (for performance and other reasons),
RSK worker machines can be coded in whatever language best meets
the requrements the worker machine is to address. Using a standard,
well-defined and tested, communications method like APPC (in the
form of it's standard CSL), would make this task much easier for vendors
and end users alike, imho.