The algorithm for distributed semaphore by broadcast given in the textbook (Fig. 9.13).
Design and Implement the Framework
As described by the algorithm, each semaphore user has a helper process which is in charge of
communications with other helper processes. These helpers are solely working for the semaphore
and are independent to any specific application. These helpers form a framework of the
distributed semaphore. A user of the framework should be able to activate a helper process
(class DisSemHelper) on each participant computer. An essential requirement of the
broadcast-based algorithms is to establish a connection to every other node from every node.
Therefore, every DisSemHelper object needs to know the IP addresses and port numbers of
all the other nodes. You are suggested to have an Initiator process (running on any one of
the nodes) and collect the IP addresses and port numbers of all the other nodes. In doing so, the
constructor of Initiator takes the port number that the process listens to. The constructor of
each DisSemHelper has four parameters: the node ID number, the port number of the
DisSemHelper, Initiator‘s IP address, and Initiator‘s port number.
In order to achieve flexibility, the framework should allow each application process to
communicate with its helper in either of the two ways: the application program instantiates the
helper as a thread, or the application program communicate with its helper process by messaging.
You should minimize application programmers’ involvement in messaging. One way to achieve
this is to write a class DisSem that has two public methods P() and V(). Conceivably, its
constructor takes information regarding the semaphore name (unique), the node address and the
port number of its helper process.
The distributed semaphore was originally proposed for a distributed database system in which
multiple database servers have dedicated network connections. You are encouraged to devise an
application of the distributed semaphore framework such as a controller of a floating software
license (resource sharing), online auction, and collaborative editing. The purpose of your
application is to illustrate your framework.
You already prepared the Java socket-based simulation of channels in your Programming
Your implementation of the algorithm in Figuire 9.13 must be divided into two parts, (1) the
distributed semaphore framework that includes the helpers (class Helper as process Helper); and
class DisSem which extract the messaging part in process User. (2) The application programs
that utilize distributed semaphore. When execute your program, each Helper object runs in a
computer. The user application programs can either be deployed in the computer that runs its
Helper object, or run in yet another computer. Every user application program communicates
with the framework by instantiating an DisSem object and then calls its methods P_op and V_op.
Method P_op first send semop[i](I, reqP, lc); …; then receive go[i](ts); … Method V_op just
send semop[i](I, reqV, lc); …
Be sure to have an adequate test plan for the implementations of your framework, and to prepare
four test cases for your application.