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 [url removed, login to view], 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.