Sentinel Rapid Deployment provides the ability to have different configurations depending on the needs of the environment. The following performance testing information is the result of Novell’s testing for specific configurations listed in the tables below.
The hardware recommendations for a Sentinel implementation can vary according to each implementation; therefore, we recommend that you consult Novell Consulting Services or any of the Novell Sentinel partner prior to finalizing the Sentinel architecture. The test information below can be used as a guideline.
Linux testing was performed to scale the maximum EPS with a different number of devices and to scale the maximum number of devices for a specific EPS. The following hardware configuration was used:
Number of CPU Cores: 4
CPU Model: Intel Xeon CPU X5770 @ 2.93 GHz
RAM: 16 GB
Hard Disk Size (+RAID type and number of disks in RAID): 1.7 TB (RAID 5, 6 disks)
NOTE:All testing was done with syslog-based event sources. Other connectors might offer different performance.
The following table shows the maximum EPS you can scale with a different number of devices on a SLES system:
Table 2-6 Maximum EPS on a SLES System
The following table shows the maximum devices you can scale at different EPS rates on a SLES system:
Table 2-7 Maximum Devices on a SLES System
NOTE:
If you want to scale more EPS or devices, install additional Collector Managers.
The maximum device limits are not hard limits, but are recommendations based on the performance testing done by Novell. They assume a low average events rate per second per device (less than 3 EPS). Higher EPS rates result in lower sustainable maximum devices. You can use the equation (maximum devices) x (average EPS per device) = maximum event rate to arrive at the approximate limits for your specific average EPS rate or number of devices, as long as the maximum number of devices does not exceed the limit indicated above.