The Seven Layers

The seven layers of the OSI model may be divided into two groups (Figure 6.5). The upper three layers are concerned with applications - the synchronisation of activity within a set of distributed applications, the representation of data, the management of associations, concurrency etc.. These are called the 'upper' or 'application oriented' layers. The lower four layers are concerned with the technology in use - error and flow control, routing etc.. These are the 'transport oriented' layers. Note that the Transport Service interface forms the boundary between these two groups. The idea is that the application oriented layers should not have to worry about the vagaries of the technological differences in the lower layers - the Transport layer should mask all these. Thus the functionality provided by the Transport layer has a great significance in determining the nature of the communications service available to the applications. Generally speaking, the boundaries between the lower three layers reflect boundaries that already existed when the OSI effort began. The upper layers are more interesting as they attempt to separate out common functionality required by applications which had not previously been the subject of standardization. This should be a 'good thing' as it should prevent new applications from re-inventing the wheel. It is these upper layers that are of most interest from the point of view of distributed system designers. Notice that you can separate out two aspects of a protocol layer:
  1. Signaling
  2. Data
In the telecom world, this separation is made obvious, since almost all signaling is done at the beginning' and end of a communications session. In the data communications world, it is less obvious, since signaling type activities may be required almost every time some data is send or received. In summary, the functions of the layers are: