CSE3OSA Assignment 2019

1. Home
2. CSE3OSA

LaTrobe University
Department of Computer Science and Computer Engineering
CSE3OSA Assignment 2019

Objectives

Gain hands on experience with multi-threaded concurrency programming with threads.

Assignment Assessment notes

This assignment is worth 20% of your overall assessment. There are two tasks. Task 1 has 50 marks. Task 2 has 40 marks. Programming Style has 10 marks. Total marks are the sum of the above three. Note the marking guide is at the end of each task. Make sure you target areas that are worth marks.

IMPORTANT: Onsite Assignment Execution Test will be conducted on Friday (9:30am-11:50am; 14:10pm-16:00pm) of 18th October. If you do not do the execution test, you can not get marks of Task 1 and Task 2. Please see the end of this document for more details.

Submission details

Submit Your Coursework via LMS. Also, you need to compile and run your solutions and show me the results during the onsite assignment execution test. Please see the end of this document for more details.

This is an individual assignment. You are not permitted to work as a part of a group when writing this assignment.

System Crashes

Since the nature of the assignment requires multi-threaded and multi-process programming (in bonus task). There is a higher probability of system crashes, especially near the deadline due to many people working at the same time. If a system crashes on the weekend it will be unavailable until Monday. So PLEASE schedule your time so that you don’t leave it all to the last minute.

Thread Implementation

Please do this assignment using C, C++ or Java Threads.

Copying, Plagiarism

Plagiarism is the submission of somebody else’s work in a manner that gives the impression that the work is your own. For individual assignments, plagiarism includes the case where two or more students work collaboratively on the assignment. The Department of Computer Science and Computer Engineering treats plagiarism very seriously. When it is detected, penalties are strictly imposed.

Programming Task

Your task is to implement a memory manager that manages page reads and writes from multiple client threads. In the first programming task the memory manager keeps all the pages in a memory buffer large enough to fit all the pages and the server processes a list of requests and sends them to clients to log. In the second task you extend the solution for task 1 so that requests are sent from the client to the server instead of all the requests originating in the server.

Please look at the Assignment Diagram file in the assignment folder to gain a better understanding of the assignment requirements.c Sample input and output files are also provided in the Sample Files file.

Please do not use sleep anywhere inside your code. Or something that simulates a sleep system call. Eg. while () { }. You are also not allowed to do some kind of busy wait for a condition like ( while (some condition not true) { } ) However you are allowed to use the pause(), pthread_cond_wait() system calls and their Java equivalents to block a thread or process.

Please submit separate source files for each of the two tasks.

Task 1 [50 marks]:

There should be 1 server thread and N client threads, where N is supplied by the user as a command line argument. The server opens a file called “all_requests.dat”, the file has the following format:

<client id><space><read/write><space><page id><space><contents of page>

<client id><space><read/write><space><page id><space><contents of page>

...

client id – refers to the id of the client thread which is between 0 and N – 1. Note this is not the thread id that is assigned by the system.

read/write – refers to whether the request is a read or write request.

page id – refers to id of the page being requested.

contents of page – this field only exists if the request is a write request. It contains the contents of a page that is to be written into the server which is a sequence of alphabet characters (a-z, A-Z) with no spaces or anything else in between and has a maximum size of 4096 characters (size of page for most machines).

Here is a sample “all_requests.dat” file. In this example there are 2 clients with ids of 0 and 1 respectively. Note the read requests do not have any associated page contents.

All_requests.dat file:

0 read 20

1 write 20 helloKKAAA

0 read 1

1 read 21

0 read 20

0 write 8 peter

The server thread also reads from the file called “init_buffer_pages.dat” which contains the initial contents of ALL the buffer pages (you can be sure read and write requests are to pages existing in this file). It has the following format:

<page id><space><contents of the page>

<page id><space><contents of the page>

Here is a sample init_buffer_pages.dat file:

1 abcdefghigkhhLLK

2 ABjjjjjsskskskskksk

20 hellohowareyou

21 HHBGHH

8 JHjhjjsjjkwje

The program should do the following:

Function 1) The server thread starts and reads in the init_buffer_pages.dat file to initialise the memory buffer.

Function 2) Create N client threads, where the number N is taken from the command line.

Function 3) The server thread reads from all_requests.dat file to get the page read or write requests. The server must process these requests in order from the 1^st to the last. For example the 2^nd request on the list must be processed before the 3^rd request.

Function 4) Each client thread should keep a log of the pages that it has read inside a file called “client_log_n” where n is its thread id. The pages read must be written into the log file in the order that they are received from the server. Please note you only need to log read requests. In the case of a write request the server does not need to contact any of the clients. This will let us know if server is processing the requests in the correct order. The beginning of each entry should contain the page id followed by the page contents. Here is the format:

<page id><space><contents of page>

<page id><space><contents of page>

Here is an example (note in the example below the page 20 has changed between the first and last reads):

20 hellohowareyou

1 abcdefghigkhhLLK

20 helloKKAAA

Function 5) Please note that the server must wait for the current client thread to have finished writing its log entry before getting the next thread to log its read request.

Function 6) When all the clients finish their requests the server thread must end and cause the entire process to end.

Marking Guide

Task 1 (50 Marks):

Please note：1) you need to run your program based on if client log files contain the correct data on the assignment execution test;

2) Please add necessary comments in your program to indicate Function 1 - 6.

Task 2 [40 marks]

Extend task 1 so that the requests are issued by the clients instead of the server. The server now do not use the “all_requests.dat” file. Instead every client thread reads from a different input file called “client_requests_n.dat”, where n is the thread id. Note this is not the thread id that is assigned by the system. Note thread id goes from 0 to N-1.

The client request files have the following format:

<request no.><space><read/write><space><page id><space><contents of the page>

<request no.><space><read/write><space><page id><space><contents of the page>

Apart from request no. field the other fields have the same meaning as in the “all_requests.dat” file in task 1.

Note request no. is a number starting from 1 2 3 ... to the total number of requests. You can assume that when you combine all the client request files that all request number starting from 1 to the total number of requests exists. Your job is to process all the requests in ascending request number order. Also note the request numbers within each client request file are always sorted in ascending order.

Here is a sample pair of client request files. In this example there are 2 clients.

client_requests_0.dat file:

1 read 20

4 write 21 hellohowareyou

5 read 21

client_requests_1.dat file:

2 read 21

3 write 20 petermaryjohn

The program should do the following:

1) The server thread starts and reads in the init_buffer_pages.dat file to initialise the memory buffer (same as task 1).

2) Create N client threads, where the number N is taken from the command line (same as task 1).

3) Each thread reads from its own client_request_n.dat file to get the page read or write requests. The requests are sent to the server in the order that they appear in the file. The server must process these requests in ascending request number order.

4) The requests are blocking requests, that is if a request from client x is to read a page client x should not send the next request until the server has provided the requested page to the client x. Similarly for a write request the client x should not issue the next request until it has received confirmation from the server that its write has completed. Requests from different clients can be issued concurrently.

5) Each client thread should keep a log of the pages that it has read inside a file called “client_log_n” where n is its thread id. The pages read must be written into the log file in the order that they are received from the server. (same as task 1)

6a) The server must process the requests in ascending order according to request number specified in the client_requests_n.dat files. For example request number 2 must be processed before request number 3, although request number 3 arrived at the server before request number 2. This can happen since the thread issuing request number 3 may have executed before the thread issuing request number 2.

6b) The server is NOT allowed to open the client_requests_n.dat file or the all_requests.dat file. The server must get the request number and all other request information from the client when the client sends its read or write request.

6c) The clients are NOT allowed to send all their requests at once at the beginning. The client can only send a new request after its previous request has been processed. See point 4 for more details.

6d) Not following the above rules will result in very significant reductions in your mark (you may loose up to 100% of the marks for task 3). The reason is if you don’t follow these rules then this task will be almost the same as task 1.

7) When all the clients finish their requests the server thread must end and cause the entire process to end. (same as task 1)

Marking Guide

Task 2 (40 Marks):

Please note：1) you need to run your program based on if client log files contain the correct data on the assignment execution test;

2) Please add necessary comments in your program to indicate Function 1 - 7.

Programming Style [10 marks]:

Programming style will be judged on the following:

1. Correct use of synchronization techniques such as mutex, semaphores, monitors, etc. [5 marks]
2. Comments / functional decomposition [5 marks]

Total Marks = Marks of Task 1 + Marks of Task 2 + Marks of Programming Style

LMS Submission Details

• Please ensure that your name and student number is on every file that you submit.
• Prepare two separate zip files, one for Task 1, and one for Task 2. And then zip up the two files into one file.
• Submit your zip file via LMS.
• You can submit the same filename as many times as you like before the assignment deadline; the previously submitted copy will be replaced by the latest one.

Onsite Assignment Execution Test

• Execution Test will be conducted on Friday (9:30am-11:50am; 14:10pm-16:00pm), 18th October (week 11), BG116. Each student will have around 5 minutes to show the solution. An Assignment Execution Test Signup table will be available from lab 5- lab 9. You can select a time slot from the table for your execution test.
• You need to compile and run your solutions and show me the results during the assignment execution test.
• Everyone is required to do the assignment execution test. Please arrange your schedule in advance. If you do not do it, you can not get marks of Task 1 and Task 2.