Monitoring, Version 6.2
This section presents deployment example for small, medium, and large environments. In the interest of readability and understanding, these examples are simplified schematics rather than actual real-world solutions. They serve as a convenient starting point for deployment planning.
Note on terminology:
The term hub environment or hub installation refers to
a single hub monitoring server and the management components with which it
is associated: remote monitoring servers, portal server, Warehouse Proxy agents,
Summarization and Pruning agent, and the Tivoli Data Warehouse. The monitoring
environment of an enterprise can be described with one of the following terms:
You can deploy a small IBM Tivoli Monitoring environment on a single computer. This type of environment might be useful for a teaching environment, for advanced prototyping and testing of solutions, or for monitoring a small, but critical, server environment. A single server environment with failover support might be built and dedicated to monitoring heartbeats from the mail servers within a medium size company.
Figure 3 shows a small monitoring environment. If you install all components on a Windows computer with a DB2 or Microsoft SQL Server database, the installation procedure described in Installing IBM Tivoli Monitoring on one computer automates the tasks involved in setting up data warehousing.
The small environment is limited to 100 to 200 agents. It can perform minimal historical data collection and does not use the IBM Tivoli Enterprise Console. Adding significant data collection or enterprise-wide event correlation requires additional servers.
The advantages of a small environment are that it is simple, the most reliable, and the least expensive.
The disadvantages of a small environment are that you can support only a small number of agents. Also, more than minimal historical data collection or a significant number of situation events could overwhelm capacity.
If you are going to use a small monitoring environment, choose a multiprocessor (2 or 4-way) for the monitoring server with at least 2 GB of memory. Configure the software and begin monitoring CPU and memory usage during periods of both normal and high volumes of situation events. If CPU or memory usage is constrained, consider deploying a separate server for the Tivoli Data Warehouse, the Warehouse Proxy agent, and the Summarization and Pruning agent.
You can build a medium monitoring environment with five to ten computers. This level of deployment might be used in a medium-sized company, or it can be used within a large company to monitor different server groups independently according to organizational boundaries. For example, one hub installation might monitor the mail servers, while another installation monitors the HTTP servers, and a third monitors transaction servers. The departments responsible for each server group maintains control over the monitoring and historical data for that group.
Figure 4 shows a medium monitoring environment.
The medium monitoring environment supports as many as 1000 to 1500 monitoring agents using a single hub monitoring server, a single Warehouse Proxy agent, and a few remote monitoring servers. Tivoli Enterprise Console is used to correlate events, and a Tivoli Data Warehouse server collects and saves historical data for debugging and reports.
The medium monitoring environment offers a number of advantages over the small monitoring environment: significant historical data collection and event correlation is possible across a large number of agents, complex database queries do not interfere with the performance of real-time monitoring tasks, and complex correlation is possible using Tivoli Enterprise Console.
The disadvantage of the medium monitoring environment is that if multiple independent hub installations are deployed, enterprise-wide historical database and event correlation is not available.
Follow these guidelines for the medium monitoring environment:
The large monitoring environment with a single hub monitoring server, shown in Figure 5, is similar to the medium environment. The difference is that the large single hub installation contains multiple Warehouse Proxy agents and additional remote monitoring servers to support a larger number of agents. (The difference in the number of remote monitoring servers and agents is not represented in the diagram.) As with the medium monitoring environment, you can deploy a single hub installation or multiple independent hub installations to monitor different server groups according to organizational boundaries.
A large single-hub monitoring environment can support thousands of monitoring agents. The number of monitoring agents that can be managed from a single-hub environment depends on a number of factors, which are described in Determining the number of monitoring servers needed. Use Table 5 to help determine the number of remote monitoring servers that are appropriate for a large environment. Install multiple Warehouse Proxy agents to spread the work of receiving historical data from the monitoring agents and inserting it into the warehouse database.
The advantage of the large single-hub monitoring environment over the medium environment is that it supports a larger number of monitoring agents.
The disadvantages are limitations of scale and, if multiple independent hub installations are deployed, enterprise-wide historical database and event correlation is not available. The multi-hub deployment described in the following section alleviates these disadvantages.
Follow these guidelines for the large single-hub monitoring environment:
Figure 6 depicts a large multi-hub installation that is designed to handle enterprise-wide historical data collection and event correlation.
A large multi-hub environment can monitor a larger number of agents than a large single-hub environment. Tivoli Enterprise Console is used to correlate events across all hub monitoring servers. A single Tivoli Data Warehouse server collects and saves historical data for debugging and reports.
The advantage of the large multi-hub monitoring environment is that historical data and event correlation is enterprise-wide. You can use complex database queries to access all data.
A disadvantage of this environment is that it is more difficult to debug. The possibility of failure increases as the number of servers grows. Failure recovery is also more complex.
Follow these guidelines for a large multi-hub monitoring environment:
For example, consider the following scenario: A multi-hub installation consists of two hubs (as in Figure 6).
Install application support for all those agent types (Windows, UNIX, Linux, z/OS, and CICS) on the hub monitoring server, remote monitoring servers, portal server, and portal desktop clients that belong to Hub A (the primary hub installation). Install application support for the z/OS and CICS agents on the hub monitoring server, remote monitoring servers, portal server, and portal desktop clients of Hub B.
For information about application support, see Installing and enabling application support.)
[ Top of Page | Previous Page | Next Page | Contents | Index ]
