立即与支持人员聊天
与支持团队交流

MessageStats 7.8 - Deployment Guide

MessageStats Architecture Overview Planning for Deployment Deployment Recommendations and Solutions

Between Task Execution Servers and Database

The task processors on the Task Execution Servers gather data from the Exchange environment and store the data in the database. Since MessageStats provides information in an optimized manner, Task Execution Servers retrieve information from the MessageStats database for use in an aggregation decision. However, most communication is from the Task Execution Server (through the task processors) to the MessageStats database.

Communication between these components is through OLE DB, an RPC-based database communication vehicle. Performance of RPC-based communication is sensitive to network obstructions. Therefore, a clear connection should be maintained between these components.

Between Task Execution Servers and Exchange Servers

Though the Exchange environment is not part of MessageStats, it is important to understand communication between Exchange and the Task Execution Servers. Typically, you locate a Task Execution Server near the MessageStats database. However, Exchange servers are often grouped geographically. You might locate a Task Execution Server closer to the Exchange servers to achieve overall gathering performance.

To determine whether to locate the Task Execution Server closer to the database or the Exchange environment, compare gathering times using both models.

Between Scheduler Service and a Task Execution Server

The MessageStats Scheduler Service and the Task Execution Server operate like a manager and a worker. The Scheduler Service tells the Task Execution Server what to do and when to start. Once the Task Execution Server is active, it reports its current state and progress to the Scheduler Service.

The volume of communication between the Scheduler Service and Task Execution Server is low and these components need not be close. However, since the communication is RPC-based, the communication path should not contain network obstructions. If a firewall is present, you must open ports to enable communication.

You can install the MessageStats Scheduler Service on its own server but it is more efficient to install it with the initial task processor (Task Execution Server). Additional task processors can be installed on additional servers as needed.

If the amount of allocated work becomes too much for one Task Execution Server to perform, the Scheduler Service requires additional workers. If the daily gathering of information takes close to 24 hours, it is time to add another Task Execution Server.

Additional Task Execution Servers improve the gathering time in two ways:

There are many variables that affect the time required to collect data. The biggest variable is network usage; performance can vary greatly at different times and days of the week. Close observation of collection times is the best way to determine when additional Task Execution Servers are required.

Between Console, Scheduler Service, and Database

The administration function of the MessageStats MMC console requires it to communicate with the MessageStats Scheduler Service and with the MessageStats Database.

The MCC console communicates with the Scheduler Service to manage gathering tasks and schedule new work. Communication is through DCOM, which is RPC-based.

The MMC console reports the real-time progress of active gatherings using information it retrieves from the MessageStats database. Communication is through the OLE DB driver, which is RPC-based. The greater communication volume between the console and database requires a good connection between these components.

相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级