Skip to content

Latest commit

 

History

History
56 lines (56 loc) · 3.95 KB

File metadata and controls

56 lines (56 loc) · 3.95 KB

Scalable-Message-Passing-Bank-Client-Java

Your objective is to model a simple banking environment. Specifically, you will be given a small number of customers, each of whom will contact a set of banks to request a number of loans. Eventually, they will either receive all of the money they require or they will end up without completely meeting their original objective. The application will display information about the various banking transactions before it finishes. That’s it. So now for the details. To begin, you will need a handful of customers and banks. These will be supplied in a pair of very simple text files – customers.txt and banks.txt. While Erlang provides many file primitives for processing disk files, the process is not quite as simple as Clojure’s slurp() function. So the two files will contain records that are already pre-formatted. In other words, they are ready to be read directly into standard Erlang data structures. An example customers.txt file would be: {jill,450}. {joe,157}. {bob,100}. {sue,125}. An example banks.txt file would be: {rbc,800}. (bmo,700}. {ing,200}. In other words, each file simply contains a set of erlang tuples. You will see that each label is associated with a number. For customers, this is the total funds that they are hoping to obtain. For banks, the number represents their total financial resources that can be used for loans. To read these files, all you have to use is the consult() function in the file module. This will load the contents of either file directly into an erlang structure (a list of tuples). Note that NO error checking is required. The text files are guaranteed to exist and contain valid data. In addition, however, please keep in mind that these are just samples. The test files will have different customer/bank names and may also contain a different numbers of customers or banks (but no more than 10 of either). So your job now is to take this information and create an application that models the banking environment. Because customers and banks are distinct entities in this world, each will be modeled as a separate task/process. When the application begins, it will therefore generate a new process for each customer and each bank. Because you do not know how many customers or banks there will be, or even their names, you cannot “hard code” this phase of the application. The customer and bank tasks will then start up and wait for contact (you may want to make each new task sleep for a 100 milliseconds or so, just to make sure that all tasks have been created and are ready to be used). So the banking mechanism works as follows:

  1. Each customer wants to borrow the amount listed in the input file. At any one time, however, they can only request a maximum of 50 dollars. When they make a request, they will therefore choose a random dollar amount between 1 and 50 for their current loan.
  2. When they make a request, they will also randomly choose one of the banks as the target.
  3. Before each request, a customer will wait/sleep a random period between 10 and 100 milliseconds. This is just to ensure that one customer doesn’t take all the money from the banks at once.
  4. So the customer will make the request and wait for a response from the bank. It will not make another request until it gets a reply about the current request.
  5. The bank can accept or reject the request. It will reject the request if the loan would reduce its current financial resources below 0. Otherwise, it grants the loan and notifies the customer.
  6. If the loan is granted, the customer will deduct this amount from its total loan requirement and then randomly choose a bank (possibly the same one) and make another request (again, between 1 and 50 dollars).
  7. If the loan is rejected, however, the customer will remove that bank from its list of potential lenders, and then submit a new request to the remaining banks.
  8. This process continues until customers have either received all of their money or they have no available banks left to contact.