PM Sizing Guidelines
See Single Server Sizing for sizing parameters. The following are general recommendations for sizing
Fault Management--Assumes sustained traps do not exceed 1200/sec when using a single 7200RPM drive. Faster drives or disk arrays can manage higher trap throughput.
Performance Management--Sizing depends on the monitor intervals, number of monitors and attributes within each monitor. Redcell Synergy supports this data at 300 to 600 inserts per second (7200 to 15000 rpm drives respectively). Every attribute is considered 1 insert. For and example an ICMP monitor has 3 attributes so 3 inserts
Traffic Flow Analysis--Database lags limit performance for this Redcell Synergy capability (the software is faster than the hardware). Traffic Flow Analysis supports a maximum of 300 to 600 inserts per second (7200 to 15000 rpm drives respectively). Actual Insert rates are highly variable and depend on how devices are configured for flows. Current licensing limits Traffic Flow Analysis exporters to 20 but this is arbitrary since one exporter can generate so much data it overwhelms the system. Alternatively100 exporters can generate minimal data and still have good performance.
15k drives can support 3000 inserts / 5 sec on a Mysql db. In this scenario network lag and device response are not considered. So, actual results will likely be slower.
Types of deployment available: single-server, distributed, clustered, HA (see Clustering). Consider the following in these architectures:
• Hardware sizing & selection
• Database sizing & selection
• Scalability considerations for each Monitor type (monitor protocols – SNMP, Telnet, WMI, etc.)
• User tunable/configurable parameters within Redcell and/or APM
• Examples of monitored environments (central-centric, branch office, etc.) – basic and advanced, polling intervals, roll-up intervals, etc.
• PM data export (how to extract needed PM data for 3rd-party use)