[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference azur::mcc

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

3869.0. "MCC : monitoring more than 300 Internet nodes ?" by ZTOIS1::VISTA (Renato VISTA, SIS Strasbourg, France) Wed Oct 07 1992 07:37

    Hi,
    
    I'm working in a DECmcc/ULTRIX BMS V1.2 environment.
    
    My customer wants to monitor Internet systems (ICMP/ping level).
    
    As he has to monitor more than 300 U*X stations, he has firstly defined
    more than 300 rules, in polling mode, on ipReachability attribute.
    
    Althought the polling interval is not too small (00:06:00 minutes), he
    finds that this way monitoring (and the functional need covered) is
    very weak versus the hardware configuration (DECsystem 5900 +
    DECstation 5000/133) expected to get performances (especially when using 
    Iconic_map).
    
    Asynchronous Traps are not seen as efficient way to monitor
    "internet-end-node" (because of Warm/Coldstart traps are sent when reboot
    is done, and not when break down happens...) ; it is possible when 
    sophisticated SNMP agents can sent enterpriseSpecific traps... (
    which will be completely displayed by DECmcc...V1.3...).
    
    Futhermore, he knows that HP Openview has the ability of monitoring
    ip datagrams in listening mode to determine whether an Internet
    station is talking on the "network" or not. If not, and after given timing
    out, an Internet station is declared as DOWN.
    
    I have no detailed information about such mechanism and I don't know if
    it is really efficient... 
    
    Does anybody have any information about that kind of tools ?
    
    
    Regards
    Renato

T.R	Title	User	Personal Name	Date	Lines
3869.1	V1.3 should answer some of your prayers.	YAHEY::BOSE		`Wed Oct 07 1992 14:04`	11
	Renato, For DECmcc V1.3 we have planned an ip reachability poller which will run as a separate process and channel events to the SNMP AM. With notification services set up, this will allow the icon colors to change when a node comes up or goes down. We have the poller running on a VAXstation 3100 M38 with 20 Mb memory and the resources used is minimal. More details regarding the poller will be provided later. Rahul.
3869.2	What about DECstation/ULTRIX platform ?	48283::VISTA	Renato VISTA, SIS Strasbourg, France	`Thu Oct 08 1992 10:05`	16
	Hi Rahul, Thank you for your reply. I think it seems to be an interesting solution. 1) Is there any SNMP poller planned running on a DECstation/ULTRIX/RISC ? 2) Is there any interoperability from SNMP-both ULTRIX & VMS based-poller with both DECmcc V1.3 ULTRIX & VMS platforms ? Regards, Renato
3869.3	Poller will be both VMS and ULTRIX	CHRISB::BRIENEN	DECmcc LAN and SNMP Stuff...	`Thu Oct 08 1992 17:58`	10
	RE: > 1) Is there any SNMP poller planned running on a DECstation/ULTRIX/RISC ? The poller will run on both VMS and ULTRIX. Not sure what you mean by "interoperable". The first version of the poller will not be distributed (i.e. there will be one poller per MCC system). We are looking at distribution in a subsequent release...
3869.4	Wonderful ipReachability news; how about DECnet Phase IV?	CUJO::HILL	Dan Hill-Net.Mgt.-Customer Resident	`Mon Oct 12 1992 00:18`	5
	This is great news, and something that has been needed all along. How about something for DECnet nodes now? What about bridges and terminal servers? -Dan
3869.5	Definitely needed, but...	CHRISB::BRIENEN	DECmcc LAN and SNMP Stuff...	`Tue Oct 13 1992 12:31`	11
	Hi Dan, We're doing an IP Reachability Poller for V1.3 to gain some experience with pollers in general (learning about integrating pollers with MCC, performance gains possible, etc). The IP poller we have now SCREAMS... We want to follow up (in a short development cycle release beyond V1.3) with efficient pollers for other technologies, along with technology specific Auto Discovery. Chris
3869.6	VAXstation 4000 Model 90: Polling and Events	CUJO::HILL	Dan Hill-Net.Mgt.-Customer Resident	`Wed Oct 14 1992 01:27`	29
	I demonstrated DECmcc V1.2 on a VAXstation 4000 Model 90, 80MB memory, to two of my customers today. I was polling 300+ ip nodes & 50+ DECnet nodes (5min intervals), 20+ stations (30sec intervals), 30+ bridges (5min intervals), and the odd terminal server, while at the same time receiving 1200 events generated by 6 nodes in 30 seconds. I had two map windows up along with the Notification Services window. I had roughly 10-15 alarms firing every minute, logging was enabled on most of them. This node was also part of a 19 node cluster, and was serving its disks to other cluster members. The performance of this system was VERY IMPRESSIVE. I could navigate the maps with no problem at all (no noticable pauses). It was a true pleasure. I can't wait to see what you guys have done with ipReachability polling. The DECnet reachability problem (polling is a customer requirement) still needs some serious attention. DECmcc is really beginning to move forward to become a more useful and (dare I say) fun product to manage multi-protocol networks. My customers commented on the multi-protocol capabilities once again today as something that is virtually untouchable by other network management platforms. Why don't we advertise this more? Keep up the good work. Also, try one of our $11,995 base-priced, Alpha-Ready, 43 VUP, 33 SPECmark, VAXstation 400 Model 90 workstations with a second-generation Ethernet controller (9.5Mbit/sec sustained). What a SWEET system. Try one. The faster, the better.