In the layered architecture, the presentation layer has a key impact on performance because the processing overhead of the presentation conversion is over 90% in the processing time of the total protocol stack. When the presentation layer receives a packet from the network layer, it starts presentation conversion with an original presentation data in the conventional approach. In the receiver side, the presentation layer can be processed after network, transport, session protocols are performed sequentially. But the processing time can be reduced if the presentation conversion is performed immediately processing as the data arrives from the network.
In this thesis, I suggest a parallel processing model for the Open Systems Interconnection(OSI) protocol stacks. Unfortunately, OSI protocol stack often impose ordering constraints that limit the opportunities to parallelize protocol layers. By using fixed packet, a parallel approach to protocol processing can be achieved between the network layer and the presentation layer. Thus It can process four layers simultaneously for the same data packet. As a result, the performance is improved as much as 61% for the conventional approach.