Due: Thursday, October 8th, 11:59 PM ET

Handy References:

Lab 2 Goals

  • Implement the server side of a non-persistent HTTP over TCP connection.

  • Understand how the socket system calls on the server side interact with the client side socket system calls.

  • Use threading to serve multiple concurrent clients.

  • More practice with sockets, send(), and recv().

Server Attention Span
Figure 1. Server Attention Span

Partnered Labs

This lab is a partnered lab. As stated on the course webpage, you are expected to follow Lab Parternship policies to:

  • be equally responsible in setting up a time to meet and for both partners to be present for office hours.

  • not delete or significantly alter code written by your partner when they are not present.

  • work on all aspects of the project together: initial top-down design, incremental testing and debugging, and final testing and code review.

You might find VSCode extremely useful to setup with your partner to LiveShare code.

Overview

Having built a Web client, for this lab we’ll look at the other end of the HTTP protocol — the Web server. As real web clients (e.g., browsers like Firefox) send requests to your server, you’ll be finding the requested files and serving them back to the clients.

Socket Programming. Step 1:Both Server and Client create sockets.The server creates a socket
Figure 2. The server runs socket, bind and listen on the serverSocket and accepts new clients, by creating a dedicated socket for each client. This dedicated socket, connectionSock, is used to send and recv to a client. The server closes connectionSock once it has completed sending the file requested by that client, and returns to accepting new connections. When we run a concurrent server in week2, the server will accept incoming clients and process them in parallel, by spawning a thread for each client.

Lab Requirements

Your server program will receive two arguments: 1) the port number it should listen on for incoming connections, and 2) the directory out of which it will serve files (typically called the document root). For example:

./lab2 8080 test_documents

This command will tell your web server to listen for connections on port 8080 and, serve files out of the test_documents directory. You should be able to connect to your server from a web-browser or from telnet using http://localhost:8080/path.

Why are we using port 8080? I thought HTTP ran on port 80.

On most UNIX systems, only users with administrative or root privileges, were allowed to run a web server on port 80. If users do not have root privileges, they can use any port greater than 4096 to run their own web server. Often, port 8080 and 8000 are used since they sound "close" to port 80.

But how would the client know that the web-server was using a different port? Isn’t it expecting that the web-server would be on port 80?

Great question! For the client to connect to a web-server on anything other than the default port 80, we would have to explicitly specify the port number in the URL as follows: http://host:port/path

Getting your Lab2 Starting Point Code

Log into CS43 Github for our class and get the ssh-URL to your lab git repository. Follow along with the prompts below to SSH, create a lab directory and clone your lab repos.

# ssh into our lab machines
ssh yourusername@lab.cs.swarthmore.edu

# cd into your cs43/labs sub-directory and clone your lab2 repo
cd ~/cs43/labs
git clone [your-ssh-URL]

# change directory to list its contents
cd lab2-user1-user2

# ls should list the following contents
ls
 Makefile README.md week1-worksheet.md week2-worksheet.md lab2.c
 test_stat.c thread_detach_example.c test_documents/

Setting up SSHFS

In preparation for this lab, you can setup SSHFS so you can access your directory structure on the CS machines as if they were on your local machine. Once you setup SSHFS you can view, modify and add files to your remote directories as if they are local.

To do so, follow the SSHFS instructions provided by Prof. Zach Palmer. MAC Users: If you have previously installed homebrew, your setup process might look a little different. Let me know if the scripts above do not work for you, and I can go over some of the tweaks to the SSHFS instructions.

Setting up Port-forwarding

With your server is running on the CS machines, it might be more convenient if you could connect to it from your local machine’s browser. To do so let’s setup a SSH port-forwarding. Let’s say you’re running your server on butter.cs.swarthmore.edu on port 8080.  You can select any local port on your laptop (8000) and tunnel the data to butter. You can "tunnel" your output with SSH such that you run your lab server on a CS machine and a local port on your desktop/laptop passes your client GET request through the ssh tunnel to the port of your choice.

ssh -L 8000:butter.cs.swarthmore.edu:8080 user@butter.cs.swarthmore.edu

The format it follows is below, read more here: https://explainshell.com/explain?cmd=ssh+-L

ssh -L [localport]:[host to tunnel to]:[port to tunnel to] [host to connect to]
  • The first port number (localport) is anything that’s available on your local machine. It can be the same as the second one (port to tunnel to) if you want, but I made it different in the example just to show that it need not match

  • The host to tunnel to and port to tunnel to should be the host/port your server is running on. Once you’ve made the connection, you can fire up http://localhost:8000 in your laptop’s browser, and it should behave as if you were connecting to butter on port 8080.

Workflow of your program

Roughly, your server should follow this sequence:

  1. Read the arguments, bind to the specified port, and find your document root (you will find the chdir() system call helpful).

  2. Accept a connection, and:

    1. hand it off to a function.

    2. week 2: pass the function to a new thread for concurrent processing .

  3. Receive and parse a request from the client.

  4. Look for the path that was requested, starting from your document root (the second argument to your program). One of four things should happen: You might want to make each of these cases a separate function!

    1. If the path exists and it’s a file, formulate a response (with the Content-Type header set) and send it back to the client.

    2. If the path exists and it’s a directory that contains an index.html file, respond with that file.

    3. week 2: If the path exists and it’s a directory that does NOT contain an index.html file, respond with a directory listing.

    4. If the path does not exist, respond with a 404 code with a basic HTML error page.

  5. Close the connection, and continue serving other clients.

Weekly Deliverables

Week 1

Serial web-server: Workflow items above 1 - 5, except 2.b and 4.c.

Week 2

Concurrent web-server: Implement 2.b and 4.c.

SimpleHTTPServer

To get an idea of how your webserver will function, try out a simple Python HTTP Server.

$ cd lab2/test_documents #cd into the directory from where you want to serve content
$ python -m SimpleHTTPServer 8000 #specify the port number you want to run your server on.
Serving HTTP on 0.0.0.0 port 8000....

Open a separate terminal window and run a client using telnet!

$ telnet localhost 8000
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.

GET /index.html HTTP/1.0
Host: localhost

(enter twice)

Week 1

Socket Programming

You will notice that the server side of socket programming has a few more system calls. Use man 2 bind, listen, and accept to read through each of these functions. Look through your starter code on github, and follow along with the description of each of the system calls.

  1. socket(): Like the client side we first create a socket and we will name it server_sock since it is going to serve a special purpose. We are going to use it to just receive new connections but we are never going to use server_sock to call send and recv.

  2. setsockopt: The default behavior of TCP (implemented by the OS) is that if you bind to a port and terminate your program, the OS makes you wait for a minute before anyone else can bind to that port again, so that there is no confusion as to who you are connecting to. This option turns it off and is really handy when you are trying to debug your webserver.

  3. bind(): On the server side, we instead want to call bind() to announce the port and IP address we are serving content from.

    • As the server, we stake our claim on a port and IP address, as our identifier, so that any incoming client can connect to us. A machine can have more than one network interface or IP address, usually if they connect to two different networks. We pass the INADDR_ANY macro, saying we are willing to serve content on all the IP interfaces we have on the server side.

  4. listen(): So far we’ve just established the port and IP address we want to be identified with. To start accepting client connections we have to call listen(). This is the equivalent of turning the socket "on".

  5. while(1): our server is always on: we enter an infinite loop, and this where the main body of the work is going to happen. We declare a second socket, that we will eventually use to represent a new client.

  6. accept(): finally, we call accept to connect to a new client. In accept you will notice that there are two sockets. One is a parameter, and the return value of accept is a new socket. Once we have a valid return value in sock from the accept call, we are ready to send and receive using sock. We will never call send and recv on the server_sock, this is just accepting new clients and handing them off to the new socket sock.

Open your week1-worksheet.md and answer Q1 on socket programming.

Accessing documents and directories on the Server

As mentioned in the lab requirements, your server program will take in the port number and the directory to serve files from. The test_documents directory in your git repository, is considered the document root path / when responding to requests. For e.g.,

  • if you are asked for /index.html, you should respond with the file that resides in test_documents/index.html.

  • If you are asked for /dir1/dir2/file.ext, you should respond with the file test_documents/dir1/dir2/file.ext.

  • If the path requested by the client is a directory, you should handle the request as if it was for the file index.html inside that directory if such a file exists.

Handling file types, and errors

In addition to serving requested files, your server should handle at least the following cases

  • A variety of file formats should all display properly, including both text and binary formats. You’ll need to return the proper HTTP Content-Type header in your response. Your code should be able to handle the following file extensions and return the appropriate Content-Type.

    • html: text/html

    • txt: text/plain

    • jpeg: image/jpeg

    • jpg: image/jpg

    • gif: image/gif

    • png: image/png

    • pdf: application/pdf

    • ico: image/x-icon

      You may assume that the file extension is correct (e.g., I’m not going to name a PDF file with a .txt suffix). You can assume. for example, that if the client asks for ./hello.pdf, then either it does not exist or if it does, then ./hello.pdf has the correct filename extension. In this case you would search for the trailing "dot" and say if the extension is pdf, then the content-type is application/pdf.

  • If asked for a file that does not exist, you should respond with a 404 error code with a readable error page, just like a Web server would. It doesn’t need to be fancy, but it should contain some basic HTML so that the browser renders something and makes the error clear.

Open your week1-worksheet.md and answer Q2 on structuring your HTTP response.

Determining if a path is a file or folder

  • chdir(): use this function to change your cwd to test_documents. You probably want to do this at the very beginning of your program so all paths can be relative to the document root.

  • stat(): use this system call to determine if a path is a directory or a file. Allocate a variable of type struct stat and pass it to stat().

    • Use the macro S_ISDIR() and pass in the st_mode field of your struct stat variable. S_ISDIR() will return a boolean telling you whether your path was to a directory or a file.

    • You don’t need to worry about all the other fields of struct stat including S_ISCHR, S_ISBLK, etc.)

Open your week1-worksheet.md and answer Q3 to write a top-down design for the control flow you plan to use in lab2.c

Week 2

Threading

Some clients may be slow to complete a connection or send a request. To prevent all other clients waiting on one slow client, your server should be able to serve multiple clients concurrently, not just back-to-back. For this part of the lab, we will use multithreading with pthreads to handle concurrent connections.

  • We will use pthread_create, and pthread_detach, after we call accept for each new client.

  • Unlike many of your prior experiences with threading (e.g., parallel GOL in CS 31), the threads in this assignment don’t need to coordinate their actions. This makes the threading relatively easy, and it’s something that can be added on once the main serving functionality is implemented. When starting out, organize your code such that it calls a function on any newly-accepted client sockets, and let that function do all the work for that connection. This will make adding pthread support quite simple!

  • In your starter code you should see a thread_detach_example.c. This is very similar to what you will be implementing. This function takes the number of threads as an input argument, and creates and detaches each thread. Each thread independently runs thread_function. We pass one argument to the thread in the example, an integer pointer. In your server, this will be the socket ID.

    • Inside of the thread_function, you just have to cast the input back from a generic void * unknown type pointer to be an integer pointer. And then you can dereference that pointer to get the value, and then you can free it. This is the main complexity in this part of the lab-- wrangling pointers!

  • Finally, we have a pthread_detach function. This basically says I am creating a thread, it is going to go do something in the background, and I don’t need the thread to return a result, but just exit once its done executing. Therefore the return value of our thread_function is NULL to satisfy a void * return value. By detaching a thread, we are telling the OS to just clean it up once its done executing our thread_function.

Open your week2-worksheet.md and answer Q1 on threading system calls.

Providing a directory listing

  • Your Web server should respond with a list of files when the user requests a directory that does not contain an index.html file.

    • Similar to opening a file with fopen and reading from a file with fread, you can read the contents of a directory using the opendir(), readdir() and closedir() calls.

    • That is, if you have a valid directory path, you can pass it to opendir and store the result in a (DIR * ) pointer. Just like a file pointer, everytime you open a directory, you should close the directory with closedir.

    • Next, you can keep calling readdir(), which returns info for one file, on that (DIR * ) pointer until it returns NULL. Look up man 3 readdir to figure out the return value of readdir and how can access the name of the file, and type of the file from the return value. DO NOT attempt to free that pointer — the man page makes it very clear that you should not attempt to free that pointer in the man page for readdir !.

    • Note that there should be no additional files created on the server’s disk to respond to the request.

    • You can follow the following html format to create your directory listing:

      <html>
      Directory listing for: #current path# <br/>
      <ul>
      <li><a href="your_dir_listing_with_slash/">"dir_name"</a></li>
      ....
      </ul>
      </html>
Open your week2-worksheet.md and write pseudo-code to provide your directory listing.

Hints / Tips / Assumptions

Always, always, always check the return value of any system calls you make! This is especially important for send, recv, fread, and fwrite calls that tell you how many bytes were read or written. * You may assume that file suffixes correctly correspond to their type (e.g., if a file ends in ".pdf" that it is a PDF file).

Client Request Assumptions

  • You may assume that requests sent to your server are at most 4 KB in length.

  • You may assume that if the user requests a path that is a directory, the path will end in a trailing /. When generating the list of files in a directory, make sure your server also sends back URLs that end in / for directories. This is for the benefit of your browser, which keeps track of its current location based on the absence or presence of slashes.

  • You can assume that you will always be passed a valid root directory that can be resolved by chdir().

  • You will only receive GET requests from the client. You can use a buffer size of 4096 for the incoming GET requests (and for that matter for the send buffer too).

  • If you receive an HTTP/1.1 request, please respond back with an HTTP/1.0 response.

String Parsing and File I/O

  • If you need to copy a specific number of bytes from one buffer to another, and you’re not 100% sure that the data will be entirely text, use memcpy() rather than strncpy(). The latter still terminates early if it finds a null terminator (\0).

  • If you are trying to do some sort of specific string or memory manipulation, feel free to ask if there’s a better/recommended way to do it rather than brute force. Often there may be a standard library function that will make things easier.

  • the file size you send to your client might be bigger than your buffer. You will probably want to read in the file from disk in a loop. And every time your buffer is full, you can call send (in an inner loop) - recall that one call to send with say 200 bytes, will not necessarily send all 200 bytes!). You can choose a buffer size of 4k - 4096, and read from disk using fread().

Responding to client requests

  • Call send and receive on int sock returned from the call to accept, but not on the server_sock.

  • If the file is found, the response HTTP headers you send to the client can be of two forms:

    • if the file is found then the header is of the form: HTTP/1.0 200 OK\r\nContent-Type: <content type>\r\n\r\n <data here>

    • if the file is not found, then the header is of the form: HTTP/1.0 404 File not found\r\nContent-Type: text/html\r\n\r\n <html error message here>

  • Remember to close the socket once you are done sending the response message (either file, or 404 not found error).

Testing

  • Take compiler warnings seriously. Unless it’s an unused variable, you should address the warning as soon as you see it. Dealing with a pile of warnings just makes things more difficult later.

  • Test your code in small increments. It’s much easier to localize a bug when you have only changed a few lines.

Week 1, testing

  • The response from your server should mimic result of running:

python -m SimpleHTTPServer
  • When testing, similar to Lab1 testing, you should be able to retrieve byte-for-byte copies of files from your server. Use wget or curl to fetch files and md5sum or diff to compare the fetched file with the original.

  • I will do this when grading. For full credit, the files need to be exact replicas of the original.

  • To test for valgrind errors you can use the following command valgrind --leak-check=full -v ./lab2. You might find that valgrind reports some still-reachable blocks. This comes from the "sock" variable that identifies the socket, which is not freed when you kill your webserver with ctrl-C since its running in an infinite loop waiting on "accept()". You should ideally have a LEAK Summary that has 0 bytes in definitely lost, indirectly lost, and possibly lost and a small number of reachable bytes when you kill your server.

Week 2, testing

In addition to the testing for week 1, your lab in week 2 should be able to handle clients in parallel. Testing simultaenous requests is next to impossible using two browsers, especiall since your webserver is hosted locally.

Instead, we are going to create a degenerate client with telnet. Run your webserver, and in a seperate window, run telnet and issue only the first line of the GET request. While your server is still waiting for the telnet client to complete its GET request, open a browser window to connect to localhost. You should still be able to receive and send requests from your browser.

Grading Rubric

Total: 8 points

  • 2 points for completing weekly-lab questions.

  • 1 point for correctly returning a 404 error code and HTML message when a request asks for a file that does not exist.

  • 1 point for serving index.html, if it exists, when asked for a file that is a directory.

  • 1 point for returning a file listing when asked for a directory that does not contain an index.html file. The listing should use simple HTML to provide clickable links.

  • 1 point for handling multiple clients concurrently. Your server should be able to have one or more open telnet connections, doing nothing, while still being able to handle browser/wget/curl requests.

  • 1 point for serving both text and binary files (that can be rendered correctly — set Content-Type) to a standard Web browser (e.g., Firefox).

  • 1 point for serving exact copies of text and binary files to command line clients like wget or curl. The MD5 sums should match!

Submitting

Please remove any debugging output prior to submitting.

To submit your code, simply commit your changes locally using git add and git commit. Then run git push while in your lab directory. Only one partner needs to run the final push, but make sure both partners have pulled and merged each others changes. See the section on Using a shared repo on the git help page.