Configuring RoundTrip for monitoring Exchange server response time
There are several ways that a RoundTrip session can be set up to monitor Exchange server response times.
Outward star
In the outward star configuration, a single server is used as the source server, and it initiates RoundTrip messages to many target servers. Use this configuration to monitor the response time of a central office and several remote offices. The outward star configuration is shown in the following figure.
Outward star RoundTrip configuration 
Inward star
In the inward star configuration, many servers act as source servers to a single target server. This configuration is similar to the outward star but has the advantage of reducing the resources used on the single target server. The practical limit to the number of servers acting as targets from a single source is 10 to 15 target servers per source server. When this limit is reached, the resources needed to process messages from all the target servers interferes with PATROL for Exchange Servers parameter updates. The inward star configuration is shown in the following figure.
Inward star RoundTrip configuration 
Circular
In a circular configuration, RoundTrip sessions are established between servers in a series. Each server acts as a source server to a different target server, eventually looping back to the original source server. This is a useful configuration if mail must pass through an intermediate server while in route to another server. In the following figure, mail from Server A to Server C must first pass through Server B.
Circular RoundTrip configuration 
Two-way
In a two-way configuration, Server A acts as the source server to Server B, and Server B acts as the source server to Server A. There is practically no advantage to performing two-way RoundTrip monitoring because there should be no difference between a RoundTrip session from Server A to Server B versus a RoundTrip session from Server B to Server A. In both cases, mail is sent from one to the other and a reply is sent from one to the other. If the Send interval, Warn after, and Alarm after time periods are short, this could be the quickest method for being notified that an Exchange server is experiencing problems. However, if the RoundTrip intervals are shorter than the polling cycles for service and process monitoring, then RoundTrip will go into alarm first. It is not always advisable to set RoundTrip parameters to very short intervals because you could be prone to false alarms. The two-way configuration is displayed in the following figure.
Two-way RoundTrip configuration 