Creating an In-process Adapter in C/C++

This section reviews how to write an in-process adapters in C/C++ to communicate with the Sybase CEP Server.

In-process adapters are compiled into dynamically linkable libraries that are called "shared object" libraries on most UNIX-like operating systems. These libraries are linked to the server, and the code in them runs as part of the server process.

This chapter explains how to create simple adapters for reading from and writing to a Sybase CEP Server in the C and C++ programming languages (this API allows you to write both input adapters and output adapters). Since this example is provided in source form, modifications are easily performed.

In-process adapters tend to have lower overhead than out-of-process adapters. They also give you access to some internal features in the engine; for example, you can call a function to find out the schema of a stream.

For an adapter to function, it must be able to attach to its Sybase CEP stream and convert data to (or from) the Sybase CEP Server representation. These two tasks - communication and conversion - are independent of each other. The primary purpose of the In-process Adapter API is to provide the functions needed to communicate with the Sybase CEP Server. You must write your own conversion functions that are appropriate for the source (or destination) of the data.

• The Components of an In-process Adapter
  A description of the components of an in-process adapter, including the library and ADL files.
• Algorithm Overview
  The C/C++ subfunctions in an in-process adapter run only when called by Sybase CEP Server. This section describes the Sybase CEP Server callback functions.
• Session State Information and Persistent State Information
  This section describes session state and persistent state information, which enables users to use the same information in more than one invocation.
• API Interface
  This section provides an introduction to the various API interfaces.
• Session States vs. Persistent States
  This section provides more detailed information about session states and persistent states.
• Signatures of User Functions
  Each user-defined adapter must provide three or four entry points in order to function. Introduces these adapter functions, their purposes, and parameters.
• In-process Input Adapters
  This section describes the intialization, execution, and shutdown subroutines for in-process input adapter, and provides an outline of the user code for an input adapter.
• In-process Output Adapters
  This section describes the intialization, execution, and shutdown callbacks for in-process output adapters, and provides an example of a typical callback pattern for output adapters.
• In-process Adapters
  This section provides more information about the initialize(), execute(), and shutdown() routines required in a C/C++ in-process adapter.
• Useful Utility Functions
  This section contains utility functions that you may find useful.
• Requirements for the C/C++ File
  This section describes function and syntac requirements for the C/C++ file.
• Step-by-Step Instructions for Creating an In-process Adapter
  This section describes the steps required to create an in-process adapter, from writing a C or C++ program to copying the ADL file to the Sybase CEP Server and Sybase CEP Studio plugins directories.
• Compiling an In-process Adapter
  This section describes the steps required and settings you need to modify in order to compile an in-process adapter.
• Practical Tips for Using the In-process Adapter SDK
  This section provides tips for debugging adapters and improving the performance of adapters.