Basics

File descriptors

• When Unix programs do any sort of I/O, they do it by reading and writing to file descriptors, which is an integer used to identify an open file.
• The open file can be a network connection, a FIFO, a pipe, a terminal, a real on-disk file, or anything else

Socket

<aside> 👉 Basics

</aside>

• A socket is an endpoint for communication, identified as a IP:port pair, and uses standard Unix file descriptors behind the scenes
• A pair of processes communicating over a network employs a pair of sockets, one for each process
• Sockets are incredibly fast, and considered to be a low-level form of communication b/w distributed processes

<aside> 👉 Types of Sockets

</aside>

There are a lot of sockets, such as Internet Sockets (DARPA INTERNET), UNIX sockets, X.25 sockets etc.

• There are a lot of Internet sockets out there, but the most common are Stream Sockets (SOCK_STREAM) and Datagram sockets (SOCK_DGRAM)
• SOCK_STREAM are reliable and use TCP behind the scenes, while on the other hand, SOCK_DGRAM are fast and use UDP behind the scenes
• With SOCK_DGRAM we don't have to maintain an open connection, like that in case of `SOCK_STREAM

<aside> 👉 How are sockets used?

</aside>

• Sockets are generally employed in a client-server architecture viz. the server awaits for incoming client requests by listening to a specific port and once a request is received, the server accepts a connection from the client socket to complete the connection
• If we assume two hosts with addresses (146.86.5.20 and 161.25.19.8) are communicating with one another, with ports 1625 and 80 respectively, then the sockets in communication can be said as 146.86.5.20:1625 and 146.86.5.20:80

<aside> 👉 Implementing sockets

</aside>

• We make a call to the socket() system routine, which returns a socket file descriptor in return. We can then communicate with the descriptor using send() and recv(), sentto() and revcfrom() for SOCK_DGRAM.
• We can also use the normal read() and write() system calls with the socket file descriptor, but the former ones offer much greater control over the socket.

Concepts

Byte Ordering

• The Big Endian byte ordering format is suitable for network programming and is known as the Network Byte Order. Since the byte order is different for different hardware, we can write a function to convert to the Network Byte Order (even if it already is, there’s no harm), and thus we can actually be sure we’re dealing with ordering the correct way.
• There are two types of numbers we can convert: Short (2 bytes) and long (4 bytes) and we must convert the numbers to the network byte order before they go out on the wire (standardized), and convert them to host byte order when they come off the wire (system dependent)

Important data structures

<aside> 👉 Socket file descriptor

</aside>

It’s simply an integer.

<aside> 👉 struct addrinfo

</aside>

struct addrinfo {
	int              ai_flags;     // AI_PASSIVE, AI_CANONNAME, etc.
	int              ai_family;    // AF_INET, AF_INET6, AF_UNSPEC
	int              ai_socktype;  // SOCK_STREAM, SOCK_DGRAM
	int              ai_protocol;  // use 0 for "any"
	size_t           ai_addrlen;   // size of ai_addr in bytes
	struct sockaddr *ai_addr;      // struct sockaddr_in or _in6
	char            *ai_canonname; // full canonical hostname
	
	struct addrinfo *ai_next;      // linked list, next node
};