[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

764.0. "MCC hardware config sizing" by RTPSWS::BRILEY (Are you a rock or leaf in the wind) Fri Mar 01 1991 09:26

    I am looking for some assistance in hardware configuration sizing.
    
    My customer is interested in implementing MCC to manage their network. 
    The network currently consists of approximately 400 VAX's, 40 routers,
    100 Vitalink bridges, 10 DECnet/SNA Gateways, and serveral 100 PC's. 
    They are expecting growth of at least 50% in the area of VAX's and
    bridges over the next two years, and the potential of thousands of
    PC's.  The PC's will be running under both Pathworks and Novell.
    
    We are currently looking at a hardware configuration consisting of a
    VAX 4000 Model 300 clustered with a VAXstation 3100 Model 38 SPX.  They
    are also looking at tossing in some windowed terminals.  Is this config
    sufficient to manage a network of this size?  What amount of disk space
    should we look at getting?  What amount of memory would be reasonable
    for each system?  I know all of these questions are dependent on the
    type and amount of management that will be done.  For the sake of this
    discussion assume near worst case load, and good response time to the
    network manager or operator.
    
    Also could someone fill me in on how the sizing process takes place?
    
    Thanks,
    
    Rob

T.R	Title	User	Personal Name	Date	Lines
764.1		BSYBEE::EGOLF	John C. Egolf LKG2-2/T02 x226-7874	`Sat Mar 02 1991 13:18`	19
	Rob, I'm going to be out next week, and I don't much time now to answer your note. Your configurations sounds good. The major things that effect performance are polling and event operations. You poll for data when you do alarm functions and also when you record historical data. You need to make some determinations as to how many polls per minute or per hour you expect to be doing. We have some data on polls per time-frame and how much cpu gets chewed up. As for disk storage, again what will you be recording, for how many entities how ofter and how long will you keep the data on disk? From this we can give some ball park figures. Got to run... JCE
764.2	More info	RTPSWS::BRILEY	Are you a rock or leaf in the wind	`Wed Mar 06 1991 11:00`	12
	John, What we would like to do is monitor circuit and node availability on a close to real time basis (any suggested intervals). Next on a less frequent basis, but often enough to capture prolonged traffic peaks we would like to monitor traffic. Finally we would like to gather configuration information on a daily basis. This is just an initial guess, any guidance would be greatly appreciated. Thanks, Rob
764.3	Any "rate of Consumption" of disk space?	DENVER::THURY	Rocky Mountain High	`Wed Apr 17 1991 18:55`	12
	I too am looking for some information on system sizing, specifically disk space. I am working with a customer and Sales to get MCC for this customer. One of my concerns is how long can I expect the minimal configuration to remain functional (I got burned with VAXcluster console -- too small disks, although met minimal config). Does anyone have any numbers for "rates of consumption" of disk space? No one locally has any MCC experience, and the SPD give no clues. Any emperical data would be greatly appreciated! Denny
764.4	recordings....	TOOK::CALLANDER		`Thu Apr 18 1991 14:56`	36
	Part of the problem we have had with giving specifics to work from is that the amount of space you use is very heavily dependent upon how you use MCC, and what asynch modules you have. If you have a site where they want to be recording sizeable amounts of data (many partitions), on a significant number of entities, you will need to take this into consideration when determining what you require. You also need to think about how you will use the PMs. Will you have the Alarms FM send all rule firings to log files, or send mail to manager accounts, or simply turn an icon a color? Some of these require disk space, especially if you want to keep the log for an extended time. Here is a quick list of some items that can take disk space, consider what you want to do and which of these will effect you. Then you can play around with MCC and even make some estimates yourself. FCL use of TO FILE qualifier that saves commands on disk (like show bridge forwading database physical entries to a file) IM PM Map files, number and size of maps (how many nodes are you talking about in the network) Alarms FM use of the alarms log command procedure to log a rule firing or exception to a log file, or the mail procedure to send data to a mail account Alarms FM number of rules stored in the alarms MIR file Historical FM recording of partition information on entities Exporter, exporting of data to disk files for use in reports Reports package, and how and where reports are to be generated, and stored DNS namespace, is the MCC node the server or is it a client, and what requirements does DNS have This is definitely not everything, but the ones I could think of right off the top. Hate to do this to you, but the ball is in your court. To get an answer you have to quantify the MCC usage requirements. Good luck