Hardware requirements to support small, medium, and large environments

Upgrade considerations for small to medium and medium to large environments

Based on whether you are upgrading from a small to medium or medium to large, you must update the following parameters in the pronet.conf file:

Hardware requirements for the Infrastructure Management Server

Consider the following hardware requirements when you are planning the size of your environment.

With an external Oracle database

¹ Note: The free swap space is needed even after the installation is complete and while the Infrastructure Management is up and running.

Hardware requirements for an Oracle database server

Consider the following hardware requirements for an Oracle database server when you are planning the size of your environment. For information about supported databases, see Supported-databases-for-Infrastructure-Management.

Database considerations

Infrastructure Management uses the Oracle database in any deployment. The Oracle database must run on a different system within the same subnet network.

Disk subsystems and database performance best practices

The performance of the disk subsystem used for the database can significantly affect the scalability of your system. To optimize performance, BMC recommends the following guidelines:

• Use RAID 10 or 01 on larger Infrastructure Management installations to better satisfy the higher demands on write I/O operations. BMC does not recommend using RAID 5.
• Use SAN on medium and large deployments.
• Use a larger number of physical disk drives (spindle count) than capacity disks. Because disk drive throughput performance has not increased at the same rate as disk capacity, you can achieve better performance with smaller disks.
• For large environments, run a defragmentation process on the database file every six months.
• Avoid extending raw data retention past recommended guidelines because this can impact I/O performance. For recommended guidelines, see the following topics:
  • Sizing-charts-and-guidelines-for-data-and-event-management-environments
  • Sizing-charts-and-guidelines-for-event-and-impact-management
  • Frequently-asked-questions-FAQ-about-performance-and-scalability
• BMC performance testing for sizing was performed on UNIX and NTFS file system. Performance problems were observed using Veritas file systems, and BMC does not recommend those file systems for large deployments. For information about database performance benchmarks and tuning recommendations, see the following topics:
  • Performance-benchmarks-and-timing-for-data-and-event-management-operations
  • Performance-tuning-recommendations-for-data-and-event-management

CPU guidelines

Parallel CPUs usually improve user response time better than a single CPU with a higher clock speed.

Guidelines and limitations for deployment

Follow these guidelines when planning the hardware deployment:

• You can run multiple Integration Service processes on the same computer. Each Integration Service process on a computer requires 3 GB to 4 GB of memory.
• The computers on which the Administration Console or the Operations Console are installed and the computer in which the TrueSight Infrastructure Management Server is installed should be running on the same network. To avoid a performance issue on the TrueSight Infrastructure Management Server, do not open the Administration Console or the Operations Console on the the TrueSight Infrastructure Management Server. Open these consoles on a computer other than the computer on which the TrueSight Infrastructure Management Server is running.
• BMC recommends that the TrueSight Infrastructure Management server be a dedicated server with no other products installed on it.
• BMC recommends a maximum of three Integration Service instances running on a single Integration Service host computer. For more information, see Sizing chart for multiple Integration Services running on the same Integration Service host.