| Title: | DECmcc user notes file. Does not replace IPMT. |
| Notice: | Use IPMT for problems. Newsletter location in note 6187 |
| Moderator: | TAEC::BEROUD |
| Created: | Mon Aug 21 1989 |
| Last Modified: | Wed Jun 04 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 6497 |
| Total number of notes: | 27359 |
A question from a customer : "cCan super-NEs with up to 300 000 objects be managed ? (Siemen talks about managing something like their next generation ATM switch, but even today complex NEs can have up to 100 000 objects, they say)." apart from the DBm limitations and event flows from these "objects" any other point worth to be highlighted ? Thank's for help, JJ
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1909.1 | use multiple DECmcc instances | TOOK::MATTHEWS | Tue Dec 10 1991 15:34 | 14 | |
It really depends on whether they are 300,000 global entities or
300,000 child entities of a single global entity. In other words
how are you going to model the ATM network. If it is 300,000
global entities, you would have many instances of MCC with
some strategy of partitioning the 300,000 entities into
some meaningful domain structure with each mcc instance only
handling a small number of the domains.
Many customers continue to think of a large centralized mono-lithic
management system. Any such system will always break at some boundary
of scaling. Distributing the load over multiple instances resolves
the issues.
wally
| |||||