Initial Configuration

Before the multicast facility can be used effectively, a couple of minor configuration chores must be carried out.

The reliable multicast protocol operates through the creation of "RMP groups". These are basically class D IP addresses which are used by RMP as host references to collect hosts together into a token ring. For example, the IP address "rtd.eavesdrop.domain" may be used to refer to a particular set of remote RMP hosts which may communicate as a token ring.

In the remote eavesdropping scheme, RMP group names have been mapped to unique CCD names, such that each camera has a corresponding group name. In order to be meaningful, this mapping has to be adopted consistently and across the entire system; i.e. each eavesdrop server and eavesdrop client must be aware of the correspondence between RMP group names and CCDs in order that the correct camera can be connected to by any client. This information is held as a resource file, .rmp_config , which must be stored in the home directory of the server operator and any client users. An example of such a file is shown below:

 
test.domain1->CAMERA1

 
rtd.eavesdrop->RTDSIMULATOR

The syntax shown here must be adhered to, as this file is read by a Tcl script. The name on the left of the arrow (->) symbol is the RMP group name for the CCD name on the right. The simulator will always be assigned the RMP group "rtd.eavesdrop", even if there is no entry in the configuration file, and so this name should otherwise be avoided.

A sample configuration file is contained in the RMP distribution.

In addition, the server operator is responsible for which clients are allowed to connect to the eavesdrop server. A file called .rmp_security should be held in the home directory of the server operator containing a list of allowed client IP addresses (not RMP addresses) or client names. An example of such a file is shown below:

 
# This is a comment 
# Here are some missing lines. 
 
rlshp2 
130.246.32.3 
130.246.32.2 
130.246.32.36 
rlspc2

Comments are preceded by a "#" character. Blank lines in the security file are allowed. The numerical IP addresses listed in this file are the addresses of clients that have permission to connect to and receive data from the server. In the absence of a security file, there will be no restrictions on client connections.

The security file can be edited while a server process is active, and can be accessed from the server application front end. See below for more information.