upload a file and send to service layer i.e. c# class library

I am trying to upload a file and send it to the service layer to save, however I keep finding examples on how the controller gets the HTTPPostedFileBase and saves it directly in the controller. My service layer has no depencies on the web dll hence do I need to read my object into a memory stream/ byte? Any pointers on how I should go about this is greatly appreciated...

Note: Files can by pdf, word so I may need to check content type also (maybe within the domain-service layer...

Code:

   public ActionResult UploadFile(string filename, HttpPostedFileBase thefile)
{
//what do I do here...?


}

EDIT:

public interface ISomethingService    
{
  void AddFileToDisk(string loggedonuserid, int fileid, UploadedFile newupload);    
}
    public class UploadedFile
    {
        public string Filename { get; set; }
        public Stream TheFile { get; set; }
        public string ContentType { get; set; }
   开发者_运维百科 }

public class SomethingService : ISomethingService    
{
  public AddFileToDisk(string loggedonuserid, int fileid, UploadedFile newupload)
  {
    var path = @"c:\somewhere";
    //if image
     Image _image = Image.FromStream(file);
     _image.Save(path);
    //not sure how to save files as this is something I am trying to find out...
  } 
}

You could use the InputStream property of the posted file to read the contents as a byte array and send it to your service layer along with other information such as ContentType and FileName that your service layer might need:

public ActionResult UploadFile(string filename, HttpPostedFileBase thefile)
{
    if (thefile != null && thefile.ContentLength > 0)
    {
        byte[] buffer = new byte[thefile.ContentLength];
        thefile.InputStream.Read(buffer, 0, buffer.Length);
        _service.SomeMethod(buffer, thefile.ContentType, thefile.FileName);
    }
    ...
}

Can't you create a method on the service layer accepting a Stream as a parameter and pass the theFile.InputStream to it ? Stream does not requires any web related dependency, and you avoid to duplicate memory by copying the data in some other data structures to consume it.