Critique on the article by liuqingyan


									                                   Critique on the article

                                  Thejaswi Mogenahalli

This article appeared in the magazine COMMUNICATIONS OF THE ACM September
2003/Vol. 46, No. 9 and is written by Alfred W. Loo. The article describes the future of
peer to peer computing. It chronicles the evolution of the peer to peer from the traditional
mainframes, through client server to the present P2P. The author advocates that the P2P
will be the answer to most of the problems in today’s world, from economical to
technological viewpoints.

The author cites many ongoing project that utilize the idle time of processors for
beneficial work. One such being the huge and ambitious project undertaken by Intel
Corporation, United Devices, the University of Oxford and the National Foundation for
cancer research. This network is far more powerful than supercomputer available today.
This project requires the donors to download a piece of software which has to be
installed, on the local machine which will utilize the idle CPU cycles for computing the
values of for Cancer Research. The author cites a few concerns that a user might have
before he is willing to donate his idle CPU time. Among the many, the most glaring is the
concern for security. The participants have to completely trust the provider and download
the software which has to be installed. Some of the issues may be:

• Deleting files or directories on the computer;
• Reading from or writing to the files on the computer;
• Executing programs or commands such as making long-distance calls with a modem
and telephone line; and,
• Connecting to other computers and performing illegal operations such as hacking.
The next being motivation, the participants do not receive any benefit from such
philanthropic deeds, hence many organization and individuals will not come forward to
do such a thing.

Compatibility is a major issue, the Cancer project as an example works only on PC’s. The
software that has to be downloaded does not work on workstations running Unix. This
posses another major question, how to achieve compatibility in a heterogeneous
environment. This was over come by the seti@home project which distributed different
installation software for different platforms. The cost of maintenance of this kind of a
process is high. The solution proposed by the author is the implementation in Java which
can run on any platform.

My perspective
The article although talks about all the huge and ambitious project that are taking place,
he fails to portray the significant impact one normal user like me.
The article point to the rather obvious problem that is faced when two or more different
project compete for the CPU time or storage.
The author ignores the complexities involved in partitioning the task at hand into several
disjoint sub task.

To top