Paper Review:
A protocol for Packet Network Intercommunicataion

Reviewer: Oleg Elkhunovich

Problem

The paper targets the problem of protocol design that allows for the sharing of resources in different packet switching networks.

Contribution

The contribution of this paper is the beginning of a standard TCP protocol that allows communication and sharing of resources between different packet switching networks. While, TCP protocol it describes is different from today's design, it is a starting point for an important networking standard.

Main Ideas

  • Interface for the protocol should be as simple and reliable as possible and should primarily deal with passing data between various packet switching networks.
  • Common protocol should be used between hosts or processes in different networks. Interface between networks has the smallest role.

    Critique

    Significance:5
    This paper is of breakthrough significance as it paves the way for one of the most important networking standards - TCP, now employed in the world's biggest network - Internet. It addresses key issues for design of the protocol for communication between packet switching networks.

    Methodology:
    Methodology of the paper is not as impressive as its overall significance. The authors do not justify a lot of their decisions, and those that are justified are not always thorougly explained.

    Limitation:
    The biggest limitation of authors' approach is the lack of experimental data to test their design. The paper would be more convincing if protocol was described in more detailed terms, and some tests on implementation would be run. The authors leave that task for future research.

    Lessons:
    The lesson from this research is that oftentimes, the requirement for the protocol or standard from economic and technological point of view, is to have as simple interface as possible. Authors leave open specification and implementation deatail and do not provide any experimental backing.