Connection Manager Code Branch

I have been working on a connection manager implementation for quite some time and I have finally got round to making it compatible with the existing controller.xml without any modifications needed to the controller.xml (although to get the most out of this branch then you should look at modifying the controller.xml).

What does it do?

This code, uses a connection manager for each connection to a server.

In the existing code each time a command needs to be sent to the server or a sensor value needs to be updated a new connection is made to the server and after the command has been executed the server is then disconnected. If you have lots of sensors that connect to the same server you quickly run into issues as multiple simultaneous connections are made to the same server and commands fail to execute.

The connection manager introduced here, does exactly what it says on the tin. It manages the connection to the individual servers and then all commands are routed through this connection. At present a permanent connection is made to the server so when the controller starts it opens the connection and when the controller is stopped it closes the connection.

The connection manager also supports command scripting (currently JavaScript) and multiple command actions (a single command can consist of any number of command send, read or script Actions; send actions send a message to the server, read actions read data from the server and script actions run a script that allows you to process the data retrieved from the server and convert it into the correct format for updating a sensor value for example.

Making use of the scripting functionality requires manual changes to the command definition in the controller.xml; if you are interested in trying this then let me know and I'll give you some pointers, this is all still in development but it does give you a lot of power to write more complicated commands.

What Protocols are supported?

Currently supported protocols: -

Telnet
UDP
HTTP
TCP

Any supported protocols in your controller.xml file will be filtered through the connection manager, other command protocols will continue to work as they did before using the old mechanism.

Where can I find it?

http://openremote.svn.sourceforge.net/svnroot/openremote/branches/feature/Controller_2_0_0_Alpha11_Manager/

How do I compile it?

Building code from the repository has been explained in the page below, just checkout the branch listed above: -

http://openremote.org/display/docs/Building+OpenRemote+2.0+Developer+Releases

How do I use it?

Once compiled it should run just like the existing controller but you will notice in the console window some different messages appearing. When one or more commands are found that use a protocol supported by the gateway manager a new gateway connection is created for each unique server (for Telnet it is the host IP address and port that make it unique), when this gateway connection is started a message is output to the console window: -

-------- GATEWAY: Started gateway for (Telnet 192.168.1.71:6601) Thread[Thread-22,5,main]

If a connection is lost then the gateway manager will try to re-establish the connection at an increasing interval from 5-60seconds and a message like below will appear in the console window: -

Cannot establish Gateway connection [Telnet 192.168.1.71:6601] so going to sleep for 5 seconds.

Additional Gateway Parameters

The connection manager also accepts some additional gateway parameters that will apply to all commands that use that gateway: -

Parameter	Description	Value
defaultPollingInterval	This sets the default time between polling requests for the sensors that use this gateway.	Seconds: 1-86400 (Default = 1)
connectTimeout	Sets the time allowed to make the connection to the server.	Milliseconds: 1-60000 (Default = 1000)
readTimeout	Sets the time allowed to read data from server.	Milliseconds: 1-5000 (Default = 200)
sendTerminator	Sets the characters used at the end of a send command.	String (Default = "\n")

These parameters are optional and if you need them they should be set within the first command element in the controller.xml that the gateway will load, using the standard property Element format, for example: -

    <command id="1" protocol="telnet">
      <property name="port" value="6601" />
      <property name="command" value="OK|status" />
      <property name="ipAddress" value="192.168.1.71" />
      <property name="defaultPollingInterval" value="10000" />
      <property name="connectTimeout" value="20000" />
      <property name="readTimeout" value="1000" />
      <property name="sendTerminator" value="\r" />
    </command>

Additional Command Parameters

The connection manager also accepts some additional command parameters, that need to be set for each command if desired.

Parameter	Description	Value
pollingInterval	This sets the time between polling requests for just this command.	Seconds: 1-86400 (Default = 1)

Building on the command element example above if you wanted to also change the polling interval just for that command then: -

    <command id="1" protocol="telnet">
      <property name="port" value="6601" />
      <property name="command" value="OK|status" />
      <property name="ipAddress" value="192.168.1.71" />
      <property name="defaultPollingInterval" value="10000" />
      <property name="connectTimeout" value="20000" />
      <property name="readTimeout" value="1000" />
      <property name="sendTerminator" value="\r" />
      <property name="pollingInterval" value="1000" />
    </command>

Why use this branch?

In it's basic form you will not notice any performance differences with the existing controller, but if you start using scripting and multiple command actions then it opens up another world of functionality. Some reasons you may want to use this code: -

If you have been experiencing connection issues with the existing controller because you have lots of sensors that use the same server.
You need the connection to be kept open at all times.
If you need to format the response from the server before feeding the value back into a sensor.
You want to change how often all the sensor values are updated.
You want to change how often a single sensor value is updated.

Can this help in my situation?

It you have a particular need that can't be met by the existing code and would like to know if this branch could help, feel free to ask, if it's not currently possible but your situation is not too specific and can benefit others then I may be able to incorporate it for you.