I want to creates a layer between any other OpenGL-based application and the original OpenGL library. It seamlessly intercepts OpenGL calls made by the application, and renders and sends images to the display, or sends the OpenGL stream to the rendering cluster.
I have completed my openg32.dll to replace the original library, I don't know what to do next,
How to convert OpenGL calls to images and what are OpenGL stream?
For an accurate description. visit the Opengl Wrapper
First and foremost OpenGL is not a libarary. It's an API. The opengl32.dll you have on your system is a library that provides the API and acts as a anchoring point for the actual graphics driver to attach to the programs.
Next it's a terrible idea to intercept OpenGL calls and turn them into something different, like multiple viewports. It may work for the fixed function pipeline, but as soon as shaders get involved it will break the program you hooked into. OpenGL is designed as an API to draw things to the screen, it's not a scene graph. Programs expect that when they make OpenGL calls they will produce an image in a pixel buffer according to their drawing commands. Now if you hook into that process and wildly alter the outcome, any graphics algorithm that relies on the visual outcome of the previous rendering for the following steps will break. For example any form of shadow mapping will be broken by what you do.
Also things like multiple viewport hacks will likely not work if the program does things like frustum culling internally, before making the actual OpenGL calls. Again this is because OpenGL is a drawing API, not a scene graph.
In the end yes you can hook into OpenGL, but whatever you do, you must make sure that OpenGL calls as made by the application get executed according to the specification. There is a authorative OpenGL specification for a reason, namely that programs rely on it to have predictable results.