| SMP could become SNMP V2 if the IETF adopts the specs. It's a new
spec based on three years of real-world experience with SNMP. New
features include security, better error messages and traps, bulk
data retrieval, security, and several other features I can't recall.
Also, there are dozens of new MIBs and a new RFC compiler to combine
MIBs, include exceptions and changes, and call the edited mib(s) a
'macro' that can be referenced in a MIB2 attribute on an agent. There
is also a common agent kind of process called "EMANATE" (aka Son of SMUX).
All neat stuff that will most likely be accepted by the SNMP market and
need to be supported by SNMP vendors (like Digital). The specs have not
been made public yet - however Jeff Case visited Digital to conduct a
class for MSU and MCC engineers (mostly on the security stuff). When
SMP takes off - those products that support it from an agent and manager
point of view will be more valuable in the customer's eyes.
As for the CMIP market - well - it's a bit on the leading edge of
market demand right now. Some telecom vendors are making use of it,
and the government GOSSIP requires it - so CMIP should start it's
movement into the growth phase of demand - but it doesn't have the
years of experience or hundreds of supporting vendors - so CMIP will be
used where it's required or where it can perform things SNMP can't.
Filtering and scoping are mentioned as things CMIP can do that SNMP
can't, but I've yet seen the application that demonstrated those
features being used in an obivous way. The DME-RPC will confuse the
need for CMIP when comes out. The real market for CMIP should develop
as the real market for OSI develops. SNMP over OSI with OSI MIBS will
be an issue - but I suspect that those willing to put up with the
complexities of OSI will enjoy the compexities of CMIP and use it.
bill
|
| Bill,
Do you have an idea for what benefits CMIP. In my country, the
DGT is going to have several network management platform with network
management. Most of the vendors' product are SNMP matured but without
solid implementation of CMIP. Althougth the trend for
Telecommunication is CMIP/OSI, do we have any comments on what reason
benefits CMIP? Or do you recommend any document for this issue?
Thanks and Regards,
barry
|