🚀 go-pugleaf

Response to Andrew McRae posting of 30 August 92:

Thank you for restating my position more eloquently.  Your point about
the fact that TCP/IP doesn't address the same types of issues that MMS
does is what I was trying to get across.  The reason I am writing about
these issues is that I see many people who are hopelessly negative about
all aspects of OSI and forget that there is some value.  The OSI process
resulted in an application standard that addressed manufacturing issues.
This is one area of OSI where OSI solves real problems and is useful.
TCP/IP, and its associated application layer protocols (FTP, etc.) do not
address these issues.  Therefore, OSI has value in spite of the fact that
Novell and FTP provide better file sharing mechanisms.

You are right in pointing out that there is no technical reason why MMS
cannot be run over TCP/IP (RFC1006 specifies how to run OSI application
protocols over TCP).  However, all MMS implementations to date have been
using OSI transport stacks.  We have a version of MMS that will run over
TCP/IP in the Sun SPARC environment (SunNet OSI).  Unfortunately, we are
the only ones offering this....there is no one else to talk to.  On the
OSI side there are many products available with MMS.  This fact alone
makes MMS over OSI more suitable for most systems.

I will also concede that OSI transport technology requires more processing
and memory than TCP/IP and this can be a problem in many systems that are
short of these resources.  I also believe that over time this will improve
to the point where it may not be an issue.  The improvement will come in 
the form of better OSI implementations and better computing platforms.
This is where we are putting our efforts because we just do not see any
support for widespread use of MMS over TCP/IP (I apologize in advance for
the use of the word "better" here).

Ralph Mackiewicz, SISCO Inc.
2943439@mcimail.com or ics.ralph@ics.org