This adaptor implements the file functionality for the Sector distributed file system.
For more information about the Sector file system and how to get it up and running, please read:
trunk/adaptors/opencloud/codeblue2/docs/. 

This adaptor is implemented and maintained by Saurabh Sehgal ( saurabh.r.s@gmail.com ). 

Once the Sector/Sphere system is configured on the system, the adaptor can be used to perform 
general file operations on Sector. The application needs to authenticate with the Sector system
before it can execute these operations. 

**** Building ****

Note: The adaptor does not build automatically with the buildsystem. It must be manually built and installed. 
To build the adaptor, make sure to do "make" in:
    opencloud/opencloud_file/
    opencloud/codeblue2/
    opencloud/external/sectorsphere/

and, "make install" in:
    opencloud/external/sectorsphere/
    opencloud/opencloud_file/


****Authentication****

1) You can encode all the information in the codeblue2/conf/client.conf file. 
2) You can encode the information dynamically in the opencloud context attributes:
     saga::attributes::context_userid 
     saga::attributes::context_userpass
     saga::attributes::context_server ( in the form hostname:port )
     saga::attributes::context_usercert
3) Or, you can encode all this information in the ini file. 

The authentication logic will give precedence to a given authentication parameter in the following order:
    1. Context
    2. Ini File
    3. Sector client.conf file

**** File ****

The opencloud_file adaptor supports all file operations except the ones listed below:

1. Permissions. To change permissions for a particular user, change the settings in the master server installation directory: codeblue2/conf/users/<username> file. 
2. Links are not implemented in Sector.  
3. Async file operations are not supported. 

Example code located in trunk/examples/misc/opencloudfile.cpp


****Job ****

The syntax for saga::job::attributes::description_arguments attributes is:
       input_file_dir output_file_dir function_name num_rows param_address<optional> param_size<optional>
where, 

1. Input file directory can be a local directory or a directory already in Sector
2. Output directory, if does not exist, is created on Sector and the results of the computation are stored here
3. The function name is the name of the function in the supplied DLL to process the input files
4. The number of rows to be processed at one time by the DLL function
5. The param address is the address of the parameter you would like to pass to the DLL function in string format
6. The param size is the size of the param stored in the memory address supplied in 5

For a working example, please see trunk/examples/misc/opencloudjob.cpp