Content

1. Overview
2. Design and Implementation
3. Derby as backend
4. MySQL as backend
5. PostgreSQL as backend
6. Connect to the Database with Matlab
7. Appendix

Overview

This guide will give a short overview to design and implementation of the LLRP message repository. The message repository provides two different datastructures:

• Repository storing and exporting all incomming and sent LLRP messages in full together with control structure (when the message was received, from which reader and adapter, etc.). Stored LLRP messages can be reconstructed completely from this datastructure.
• ROAccessReportsRepository saving RO_ACCESS_REPORT LLRP messages in a serialized form. This repository gives a more convenient way, how to access the message content of LLRP messages with tag reports.

  NOTICE: The ROAccessReportsRepository is optional!

Design and Implementation

The Repository and ROAccessReportsRepository both implement a Strategy Pattern allowing to exchange the underlying repository implementation transparently. At runtime, the LLRP Commander automatically invokes the "Concrete Strategy" depending on the configuration specified by the user.

// how to get a handle to the Repository
Repository repository = ResourceCenter.getInstance().getRepository();

// how to get a handle on the ROAccessReportsRepository:

// first get a handle of the repository:
Repository repository = ResourceCenter.getInstance().getRepository();

// now obtain a handle of the ROAccessReportsRepository
ROAccessReportsRepository roAccessReportsRepo = repository.getROAccessRepository();

Derby as backend

By default, the LLRP Commander uses the derby interface as its primary logging backend (zero configuration). In the eclipse preferences page select the option "Use internal standalone Derby Database" (should be checked by default).

MySQL as backend

In order to log the LLRP messages to an external MySQL database, the LLRP Commander needs access to a database of the MySQL server. Please be aware that MySQL currently is not able to log fractional time-parts (milliseconds and smaller). So, if you plan to inspect the timestamp on such a fine grain level, you should probably consider using the PostgreSQL database.

We assume, that you have MySQL installed and running on your system.

Log into the MySQL Command Line Client as root and perform the following steps:

1. Create the database (preferably use llrpMsgDB as the name).

   mysql> CREATE DATABASE llrpMsgDB;
   

2. Create a user that is allowed to access the database. We use username llrp and password llrp.

   mysql> GRANT CREATE,SELECT,INSERT,UPDATE,DELETE,DROP ON llrpMsgDB.* TO llrp IDENTIFIED BY 'llrp';
   

3. Update the settings in the eclipse preferences page:
   1. Ensure to uncheck "Use internal standalone Derby Database".
   2. Specify the DB implementor

      For the Fosstrak MySQL-connector this equals to:
      org.fosstrak.llrp.client.repository.sql.MySQLRepository
      

   3. Set the JDBC Connect URL.

      Syntax: jdbc:mysql://<HOSTNAME_OR_IP>:<PORT>/<DATABASE>
      
      If you set the database to llrpMsgDB and the MySQL server runs on the localhost, put:
      jdbc:mysql://localhost:3306/llrpMsgDB
      

   4. Set the username to llrp and the password to llrp.
   5. Click on the button "Test". If no error occurs, the "Switch" button is enabled and you can switch your backend at runtime.
   6. Although the LLRP Commander supports "on-the-fly" exchange of the repository, we recommend you to restart eclipse in order to load the new configuration from scratch.

PostgreSQL as backend

In order to log the LLRP messages to an external PostgreSQL database, the LLRP Commander needs access to a database of the PostgreSQL server.

1. Open the PostgreSQL admin tool (pgAdmin) and connect to your server. On *nix you can alternatively use the command-line to create the database and the roles.
   1. Create a new Login Role with Role-name llrp and password llrp. In Role Privileges check Can create database objects (see Appendix).

      Command-line command to create the role in *nix
      > sudo -u postgres createuser -D -A -P llrp
      

   2. Create the database (preferably use llrpmsgdb as the name) and set the owner to llrp.

      Command-line command to create the database in *nix
      > sudo -u postgres createdb -O llrp llrpmsgdb
      

2. Update the settings in the eclipse preferences page:
   1. Ensure to uncheck "Use internal standalone Derby Database".
   2. Specify the DB implementor

      For the Fosstrak PostgreSQL-connector this equals to:
      org.fosstrak.llrp.client.repository.sql.PostgreSQLRepository
      

   3. Set the JDBC Connect URL.

      Syntax: jdbc:postgresql://<HOSTNAME_OR_IP>:<PORT>/<DATABASE>
      
      If you set the database to llrpmsgdb and the PostgreSQL server runs on the localhost, put:
      jdbc:postgresql://localhost:5432/llrpmsgdb
      

   4. Set the username to llrp and the password to llrp.
   5. Click on the button "Test". If no error occurs, the "Switch" button is enabled and you can switch your backend at runtime.
   6. Although the LLRP Commander supports "on-the-fly" exchange of the repository, we recommend you to restart eclipse in order to load the new configuration from scratch.

Connect to the Database with Matlab

Make sure to configure the LLRP Commander to use either the MySQL or the PostgreSQL database as backend!

Connect from Matlab:

1. You need to add the MySQL or PostgreSQL JDBC-Connector to the Java classpath of Matlab before you can establish a connection to the SQL server.
   1. Either obtain a copy of the MySQL JDBC-Connector (download) or PostgreSQL (download - please select JDBC-4 8.3.*).

      Store the file to a location of your choice.

   2. Add a reference in the Matlab java classpath.txt. You can find the file in the folder matlabroot\toolbox\local\classpath.txt.

      file: matlabroot\toolbox\local\classpath.txt
      Example: add reference to the JDBC Connector
      ...
      c:/sqlDrivers/mysql-connector-java-5.1.6.jar
      c:/sqlDrivers/postgresql-8.3-603.jdbc4.jar
      

   3. Restart Matlab!
2. Establish a connection to the database:

   Example MySQL:
   >> conn = database('llrpMsgDB', 'llrp', 'llrp', 'com.mysql.jdbc.Driver', 'jdbc:mysql://localhost:3306/llrpMsgDB');
   
   Example PostgreSQL:
   >> conn = database('llrpmsgdb', 'llrp', 'llrp', 'org.postgresql.Driver', 'jdbc:postgresql://localhost:3306/llrpmsgdb');
   

   1. Example how to fetch all the entries already logged into the database:

      >> result = get(fetch(exec(conn, 'SELECT * FROM table_ro_access_reports')), 'Data');
      

   2. Example how to fetch all the EPCs:

      >> result = get(fetch(exec(conn, 'SELECT count(EPC) FROM table_ro_access_reports')), 'Data');
      

   3. Example how to display the results:

      >> disp(result)
      

Appendix

User-settings for the PostgreSQL: