Practical 1: Non-blocking Web server
Due Friday by 17:00 Points 35 Available until Jun 21 at 23:59
Please Note: Before attempting this practical please study the material on Sockets. pdf
and the tutorials under Week 2 lecture materials I will not be covering socket
programming in the lecture and this is left for you as a learning exercise and part of
practical 3. If you are having a lot of issues with coding, please see the helpful staff at the
Computer Science Learning Center
Learning Outcomes
1. Reading a Protocol Specification
Most of the application layer protocols are relatively simple and text-based. This exercise gets you
to first read the protocol specification(s). The RFC documents are not very typical of a standards
document - they are very compact. They are also somewhat incomplete - they come with a
standard reference implementation which is used to resolve ambiguities in the specification. So
the first learning outcome is the ability to read a protocol specification.
2. Practical Experience with Socket API
The second outcome is some practical experience with using the socket API. This is quite
straightforward in reality, but there are quite a few "not immediately obvious" tricks.
3. Understanding the HTTP
The third learning outcome is a detailed knowledge of the protocol itself. HTTP is simple and not
dissimilar to many other application layer protocols. Once you see how this works, it becomes
much easier to implement other protocols (Simple Mail Transfer Protocol, SMTP, for example) and
to add functionality to existing applications ("send this information to a friend")
Note that the focus is not the programming but the protocol. There are generally libraries
available in high-level languages that support HTTP and other protocols. If you were
building a service you would most likely use these libraries rather than working at the
socket layer. However, those libraries hide much of the protocol behavior we are trying to
study, which is our motivation for working at the socket layer.
Acknowledgement: This assignment is Adapted from Kurose & Ross - Computer Networking: a
top-down approach featuring the Internet. It is not the same though so make sure you are
following the specifications and using the provided code from this practical sheet and not the
textbook or website.
CBOK categories: Abstraction, Design, Data & Information, Networking, and Programming.
Submitting your assignment
You will need to keep your files in SVN. If you aren't familiar with SVN, the following SVN notes
will assist you.
The handin key for this practical is <year>/<semester> <year>/<semester>/cna/webserver
You can submit and automatically mark your assignment by following this link
(https://cs.adelaide.edu.au/services/websubmission/) to the web-submission system. Note that this is
just a preliminary test to make sure your Web server meets the basic functionality. It's your
responsibility to ensure it meets the specifications for the methods your are asked to implement in
this practical.
Template files
This practical comes with a skeleton web server and helper methods to reduce complexity.
Download and extract the following zip file into your repository.
prac1_files.zip
We will go through these files in the next section.
The practical is set up so that all code you need will belong inside the comment tags below, where
X ranges from 1 to 11.
Introduction
In this practical, we will develop a web server. In the end, you will have built a non-blocking web
server that is capable of processing multiple simultaneous service requests in parallel. You should
/* START CODE SNIPPET X */
...
/* END CODE SNIPPET X */
1
2
3
be able to demonstrate that your Web server is capable of delivering your home page to a web
browser and that your server returns standards compliant responses.
Resources to study
To understand the data that is sent to and from a web server you will need to study the HTTP. Use
the following resources and sections to help you understand the key data points that need to be
sent and received.
RFC 1945 (https://www.cs.adelaide.edu.au/users/third/cn/RFCs/rfc1945.txt)
Wireshark HTTP lab quiz
Verbose cUrl command
For the purposes of this practice we will only be looking at HTTP/1.1 and older requests. HTTPS
requests are out of the scope of this practical.
RFC 1945
Most of the application layer protocols are relatively simple and text-based. The RFC documents
are not very typical of a standards document - they are very compact. They are also somewhat
incomplete - they come with a standard reference implementation which is used to resolve
ambiguities in the specification. It is important to be able to read a protocol specification.
For this practical, you will need to understand the lines (text) that gets sent to the server on an
HTTP request, and (at the very least) the first line that is returned. Scan the contents page for the
keywords; request and response.
Wireshark HTTP Lab Quiz
In doing the Wireshark HTTP lab, you would have seen the data that is sent to and from a web
server. There is a lot of extraneous data that is sent which is beyond the scope of this practical.
Identify the key data bits of data.
Food for thought: What data would you change to cause the HTTP request to fail?
Verbose cUrl command
The cUrl command is a command line web request tool. By enabling the -v (verbose) option you
can see the data sent to and from a web server.
The following command is to the URI http://jsonplaceholder.typicode.com/todos/1
(http://jsonplaceholder.typicode.com/todos/1)
1 $ curl -v http://jsonplaceholder.typicode.com/todos/1
In the output above > denotes data sent to the server and < denotes data received from the
server. Study lines 4, 5, 6 and 10 from the above carefully.
Compile and Run
To compile the Web server run:
To run the Web server run:
When you run the program for the first time it will terminate with no output. That is ok for now, we
* Trying 104.24.106.213...
* TCP_NODELAY set
* Connected to jsonplaceholder.typicode.com (104.24.106.213) port 80 (#0)
> GET /todos/1 HTTP/1.1
> Host: jsonplaceholder.typicode.com
> User-Agent: curl/7.54.0
> Accept: */*
>
< HTTP/1.1 200 OK
< Date: Mon, 25 Feb 2019 23:07:55 GMT
< Content-Type: application/json; charset=utf-8
< Content-Length: 83
< Connection: keep-alive
< Set-Cookie: __cfduid=d4765bb5a80e476f748c1625a4b41109d1551136075; expires=Tue, 25-Feb-20
< X-Powered-By: Express
< Vary: Origin, Accept-Encoding
< Access-Control-Allow-Credentials: true
< Cache-Control: public, max-age=14400
< Pragma: no-cache
< Expires: Tue, 26 Feb 2019 03:07:55 GMT
< X-Content-Type-Options: nosniff
< Etag: W/"53-hfEnumeNh6YirfjyjaujcOPPT+s"
< Via: 1.1 vegur
< CF-Cache-Status: HIT
< Accept-Ranges: bytes
< Server: cloudflare
< CF-RAY: 4aedd4b8b9ce1d44-MEL
<
{
"userId": 1,
"id": 1,
"title": "delectus aut autem",
"completed": false
* Connection #0 to host jsonplaceholder.typicode.com left intact
1 $ gcc helpers.c web_server.c -o web_server
1 $ ./web_server [port]
will fix them in as we go along.
In the following steps, we will go through the code for the first implementation of the Web server in
C. Follow along with the web_server.c file.
Non-blocking
Our first implementation of the Web server will be non-blocking, where the processing of each
incoming request will take place inside a separate process. This allows the server to service
multiple clients in parallel, or to perform multiple file transfers to a single client in parallel.
In the code below, this is done with fork() . This creates a new process with a copy of the original
process' variables. Threads can also be used to serve multiple requests in parallel. In this case,
both threads share variables. This is more efficient as a copy doesn't have to be made, but it also
requires care to make sure that both threads don't try to change values at the same time and that
changes happen in the right order. If you are keen to use threads instead, you are welcome to
change the code.
The basic structure of the code is:
The parent code will listen for new connection requests and each time a new connection is
received a child is created and will run code to handle the new connection: reading the HTTP
request and sending an HTTP response on the connection.
The parent will continue separately and go back to the start of the loop accepting further
int main(int argc, char *argv[])
{
while (true)
{
if ((pid = fork()) == 0)
{
/*----------START OF CHILD CODE----------------*/
/* We are now in the child process */
...
/* All done return to parent */
exit(EXIT_SUCCESS);
}
/*----------END OF CHILD CODE----------------*/
/* Back in parent process
* if child exited, wait for resources to be released
*/
waitpid(-1, NULL, WNOHANG);
connections.
Variables
The main program starts by defining variables that will be used in the code.
It is important to understand the purpose of each variable.
The server will have two sockets:
1. The listen_socket variable is for the socket that listens and accepts new requests
2. The connection_socket variable is to hold the connection to a client.
It is very important that you send and receive requests on the correct socket.
The response_buffer is just a variable to hold the response that your server will send back to the
client. It is set to the maximum response size which can be changed in the config.h file.
status_code and status_phrase will be used to hold the HTTP response code and phrase that
should be sent back to the client.
Step 1: Create a socket
The first step for the server is to create a socket. You will need to write the code to create the
socket. Be sure to use the correct variable (either listen_socket or connection_socket ).
If you are not sure how to create a socket referer to:
The lecture notes
socket man page $ man 2 socket
int main(int argc, char *argv[])
{
/* structure to hold server's and client addresses, respectively */
struct sockaddr_in server_address, client_address;
int listen_socket = -1;
int connection_socket = -1;
int port = 0;
/* id of child process to handle request */
pid_t pid = 0;
char response_buffer[MAX_HTTP_RESPONSE_SIZE] = "";
int status_code = -1;
char *status_phrase = "";
Online socket tutorial (https://www.binarytides.com/socket-programming-c-linux-tutorial/)
Specifying the port
Normally, Web servers process service requests that they receive through the well-known port
number 80. You can choose any port higher than 1024, but remember to direct any requests to
your Web server with the corresponding port. The following code sets the port to either the value
typed on the command line, or if no port is given, then the DEFAULT_PORT , which is defined in the
config.h file.
Steps 2 & 3: Binding
/* 1) Create a socket */
/* START CODE SNIPPET 1 */
...
/* END CODE SNIPPET 1 */
1
2
3
4
/* Check command-line argument for port and extract
* port number if one is specified. Otherwise, use default
*/
if (argc > 1)
{
/* Convert from string to integer */
port = atoi(argv[1]);
}
else
{
port = DEFAULT_PORT;
}
if (port <= 0)
{
/* Test for legal value */
fprintf(stderr, "bad port number %d\n", port);
exit(EXIT_FAILURE);
}
Next, you need to bind the socket you created to the port and network interfaces that it should
listen to. You will need to write the code to set the correct values in the server_address structure
and call bind() to bind the address information to the socket.
The memset function makes sure that serv_addr doesn't have any values in it (i.e. it will clear the
data structure so that it contains all 0's).
See $ man 2 bind for how to use the bind() method
See https://linux.die.net/man/7/ip (https://linux.die.net/man/7/ip) on how to set the values in
serv_addr.
Steps 4 & 5: Listening
The server should now start listening for a TCP connection request on the socket. Once it is
listening, it can begin accepting connections. Because we will be servicing request messages
indefinitely, we place the accept request operation inside of an infinite loop. This means we will
have to terminate the Web server by pressing ^C (Ctrl-C) on the keyboard.
/* 2) Set the values for the server address structure */
/* START CODE SNIPPET 2 */
...
/* END CODE SNIPPET 2 */
/* 3) Bind the socket to the address information set in server_address */
/* START CODE SNIPPET 3 */
...
/* END CODE SNIPPET 3 */
You need to add the code to make the socket listen and to accept a connection.
the accept method will block until a connection is made.
See $ man 2 listen and $ man 2 accept for how to use the listen() and accept() methods
Handling a connection
When a connection request is received, we create a child process to handle the request in a
separate process and close any sockets the child is not using (remember the child gets a copy of
the sockets. Closing the child socket does not close the parent's copy only it's own copy).
After the child has started execution, the main (parent) process closes its copy of the connected
socket and returns to the top of the request processing loop to accept another connection. The
main process will then block, waiting for another TCP connection request, while the child continues
running. When another TCP connection request is received, the main process goes through the
same process of child process creation regardless of whether the previous child has finished
execution or is still running.
Note the wait() call below allows the system to free up resources from the child when a child
exits. If no child has exited, wait() with NOHANG returns immediately.
/* 4) Start listening for connections */
/* START CODE SNIPPET 4 */
...
/* END CODE SNIPPET 4 */
/* Main server loop
* Loop while the listen_socket is valid
*/
while (listen_socket >= 0)
{
/* 5) Accept a connection */
/* START CODE SNIPPET 5 */
...
/* END CODE SNIPPET 5 */
/* Fork a child process to handle this request */
if ((pid = fork()) == 0)
Step 6: Reading the request
Once we have closed the listen_socket in the child process, we create an http_request structure
which will be used to store the information about the HTTP request, such as it's method and URI
and pass this structure to the helper function Parse_HTTP_Request() . This helper will read the
request from the given socket and fill in the http_request structure for you. The interface
information about the helper functions is in the file helpers.h and the file helpers.c contains the
implementation. Note that we have provided these to help you and are not supposed to be an
exhaustive list of functions required to complete the practical. However, most of what you
need is given to you in these helper files so the helper files tries to make this practical
much easier for you by doing most of the hard work for you.
new_request will now be filled with the method and URI that the client sent.
/*----------START OF CHILD CODE----------------*/
/* We are now in the child process */
/* Close the listening socket
* The child process does not need access to listen_socket
*/
if (close(listen_socket) < 0)
{
fprintf(stderr, "child couldn't close listen socket\n");
exit(EXIT_FAILURE);
}
...
/* All done return to parent */
exit(EXIT_SUCCESS);
}
/*----------END OF CHILD CODE----------------*/
...
/* if child exited, wait for resources to be released */
waitpid(-1, NULL, WNOHANG);
/* See httpreq.h for definition */
struct http_request new_request;
/* 6) call helper function to read the request
* this will fill in the struct new_request for you
* see helper.h and httpreq.h
*/
/* START CODE SNIPPET 6 */
...
/* END CODE SNIPPET 6 */
Closing the connection
Finally, the child is finished and can close any sockets it still has open and exit.
Now we are ready to connect to a browser.
Connecting to a browser
After your program successfully compiles, run it with an available port number (e.g. 8080, 9000,
etc), and try contacting it from a browser.
Enter http://localhost:[port]/index.html into the address bar of your browser, replacing [port]
with the port number for your Web server and press enter. The browser will display an error
because our Web server has accepted the request, closed the connection and not sent anything
/* Back in parent process
* Close parent's reference to connection socket,
* then back to top of loop waiting for next request
*/
if (connection_socket >= 0)
{
if (close(connection_socket) < 0)
{
fprintf(stderr, "closing connected socket failed\n");
exit(EXIT_FAILURE);
}
}
/* if child exited, wait for resources to be released */
waitpid(-1, NULL, WNOHANG);
back.
In the terminal (that is running your Web server) the server should display the contents of the
HTTP request message.
Step 7: Parsing HTTP Request
Instead of simply displaying the browser's HTTP request message, we will analyze the request
and send an appropriate response. We are going to ignore the information in the header lines, and
use only the path contained in the request line.
You now need to write code that will decide, based on the results from Parse_HTTP_Request() , what
the status_code and status_phrase variables should be set to for the response.
Your server is required to handle and return the correct status code and phrase for Your server is required to handle and return the correct status code and phrase for
Waiting connection on port 8080...
received request: GET /index.html HTTP/1.1
Host: localhost:8080
Connection: keep-alive
Upgrade-Insecure-Requests: 1
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_3) AppleWebKit/537.36 (KHTML, like
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8
Accept-Encoding: gzip, deflate, br
Accept-Language: en-AU,en;q=0.9,en-US;q=0.8
Method is: GET
URI is: /index.html
version is: HTTP/1.1
Sending response line:
the following situations the following situations:
1. The requested resource is found
2. The requested resource is not found
3. The requested method isn't implemented (your server will only implement GET and HEAD
methods)
4. The client sent an invalid request
Look in the RFC 1945 (https://www.cs.adelaide.edu.au/users/third/cn/RFCs/rfc1945.txt) specification
to work out the most suitable status code and phrase.
You can use the -X [method] option in cUrl to run web requests on your Web server.
Fill in the code needed at:
Having determined the appropriate response, we can now send the response header. Check the
HTTP response format to make sure you fill this in correctly!
Step 8: Formatting the response
Step 9: Sending the response
Now we can send the status line and our header lines to the browser by writing to the socket. Be
sure you send to the correct socket.
1 $ curl -v -X GET localhost:8080/index.html
/* 7) Decide which status_code and reason phrase to return to client */
/* START CODE SNIPPET 7 */
...
/* END CODE SNIPPET 7 */
/* 8) Set the reply message to the client
* Copy the following line and fill in the ??
* sprintf(response_buffer, "HTTP/1.0 %d %s\r\n", ??, ??);
*/
/* START CODE SNIPPET 8 */
sprintf(response_buffer, "HTTP/1.0 %d %s\r\n", ??, ??);
/* END CODE SNIPPET 8 */
printf("Sending response line: %s\n", response_buffer);
/* 9) Send the reply message to the client
* Copy the following line and fill in the ??
* send(??, response_buffer, strlen(response_buffer), 0);
*/
/* START CODE SNIPPET 9 */
Steps 10 & 11: Preparing a HTTP Response
Now that the status line has been placed on the socket on its way to the browser, it is time to do
the same with the response headers and the entity body. We need to decide whether or not an
entity body should be returned.
You will need to understand the difference between GET and HEAD.
We can use the helper methods Is_Valid_Resource() to see if the requested file exists. If the
requested file does exist, we can call helper function Send_Resource() to send the Content-Length
header and the file contents on the socket (see helpers.h for how to use Is_Valid_Resource()
and Send_Resource() ).
After sending the entity body, the work in the child has finished, so we close the socket before
terminating.
Running the Web server
When you have completed all the steps correctly your web browser will be able to respond with a
message when running http://localhost:[port]/index.html
send(??, response_buffer, strlen(response_buffer), 0);
/* END CODE SNIPPET 9 */
8
9
bool is_ok_to_send_resource = false;
/* 10) Send resource (if requested) under what condition will the
* server send an entity body?
*/
/* START CODE SNIPPET 10 */
is_ok_to_send_resource = ...
/* END CODE SNIPPET 10 */
if (is_ok_to_send_resource)
{
Send_Resource(connection_socket, new_request.URI);
}
else
{
/* 11) Do not send resource
* End the HTTP headers
* Copy the following line and fill in the ??
* send(??, "\r\n\r\n", strlen("\r\n\r\n"), 0);
*/
/* START CODE SNIPPET 11 */
send(??, "\r\n\r\n", strlen("\r\n\r\n"), 0);
/* END CODE SNIPPET 11 */
Prac 1: Web Proxy (1)
Criteria Ratings Pts
1.0 pts
4.0 pts
You should also make sure that your Web server will work with HEAD and other requests. You
may use the cUrl command to test your web server.
Last Words
We have only coded the very basics of a web server. There is a lot of missing, for instance, we do
not return the content type of the file, we don't return an entity body when there is an error (which
means nothing will be displayed in the browser when an error occurs). If you have extra time, look
at implementing some of these.
This completes the code for the second phase of development of your Web server. Try adding
some web pages or text files to the RESOURCE_PATH directory (the RESOURCE_PATH directory is
defined in config.h ) Create this directory where you are running the server and try viewing your
files with a browser. Remember to include a port specified in the URL of your home page, so that
your browser doesn't try to connect to the default port 80.
When you connect to the running web server with the browser, examine the GET message
requests that the web server receives from the browser. You can use a tool called telnet to check
that your server responds correctly to HEAD requests.
Successful compilation of code based on the skeleton provided.
If your server crashes for reasons that are not due to a failure of the
websubmission system (for example submitting solutions that are not
tested on the University Linus image or with inadequate amount of testing
and with software bugs), we will perform no further manual inspections or
testing and you will receive an automatic ZERO for the whole assignment.
1.0 pts
Server started up successfully
If your server crashes for reasons that are not due to a failure of the
websubmission system (for example submitting solutions that are not
tested on the University Linus image or with inadequate amount of testing
and with software bugs), we will perform no further manual inspections or
4.0 pts
testing and you will receive an automatic ZERO for the whole assignment.
Server acceptance of client request for html file and correct response 5.0 pts
Correct server behaviour for a request for non existent file 5.0 pts
Full
Correct server behaviour to a HEAD method 5.0 pts
Correct server behaviour to an unimplemented method 5.0 pts
Correct sever behaviour for a bad request 5.0 pts
Server successfully handles multiple requests (non-blocking) 5.0 pts
版权所有：编程辅导网 2021 All Rights Reserved 联系方式：QQ:99515681 微信：codinghelp 电子信箱：99515681@qq.com
免责声明：本站部分内容从网络整理而来，只供参考！如有版权问题可联系本站删除。
