Networked Applications: Sockets

Assembled by Ossi Mokryn, based on Slides by Jennifer Rexford, Princeton,

And on data from beej’s guide : http://beej.us/guide/bgnet

1

Socket programming

Application Layer2

Socket API introduced in BSD4.1 UNIX,

1981 explicitly created, used,

released by apps client/server paradigm two types of transport

service via socket API: unreliable datagram reliable, byte stream-

oriented

a host-local, application-created,

OS-controlled interface (a “door”)

into whichapplication process can both send and receive messages to/from another

application process

socket

Goal: learn how to build client/server application that communicate using sockets

Clients and Servers

3

Client program Running on end host Requests service E.g., Web browser

Server program Running on end

host Provides service E.g., Web server

GET /index.html

“Site under construction”

Client-Server Communication

4

Client “sometimes on” Initiates a request to

the server when interested

E.g., Web browser on your laptop or cell phone

Doesn’t communicate directly with other clients

Needs to know the server’s address

Server is “always on” Services requests

from many client hosts

E.g., Web server for the www.cnn.com Web site

Doesn’t initiate contact with the clients

Needs a fixed, well-known address

Client and Server Processes

5

Program vs. process Program: collection of code Process: a running program on a host

Communication between processes Same end host: inter-process communication

Governed by the operating system on the end host Different end hosts: exchanging messages

Governed by the network protocols

Client and server processes Client process: process that initiates

communication Server process: process that waits to be contacted

Socket: End Point of Communication

6

Sending message from one process to another Message must traverse the underlying network

Process sends and receives through a “socket” In essence, the doorway leading in/out of the house

Socket as an Application Programming Interface Supports the creation of network applications

socket socket

User process User process

OperatingSystem

OperatingSystem

Identifying the Receiving Process

7

Sending process must identify the receiver Name or address of the receiving end host Identifier that specifies the receiving process

Receiving host Destination address that uniquely identifies the

host An IP address is a 32-bit quantity

Receiving process Host may be running many different processes Destination port that uniquely identifies the socket A port number is a 16-bit quantity

Using Ports to Identify Services

8

Web server(port 80)

Client host

Server host 128.2.194.242

Echo server(port 7)

Service request for128.2.194.242:80

(i.e., the Web server)

Web server(port 80)

Echo server(port 7)

Service request for128.2.194.242:7

(i.e., the echo server)

OS

OS

Client

Client

Knowing What Port Number To Use

9

Popular applications have well-known ports E.g., port 80 for Web and port 25 for e-mail Well-known ports listed at http://www.iana.org

Well-known vs. ephemeral ports Server has a well-known port (e.g., port 80)

Between 0 and 1023 Client picks an unused ephemeral (i.e., temporary)

port Between 1024 and 65535

Uniquely identifying the traffic between the hosts Two IP addresses and two port numbers Underlying transport protocol (e.g., TCP or UDP)

Delivering the Data: Division of Labor

10

Network Deliver data packet to the destination host Based on the destination IP address

Operating system Deliver data to the destination socket Based on the protocol and destination port

# Application

Read data from the socket Interpret the data (e.g., render a Web page)

Windows Socket API

11

Socket interface Originally provided in Berkeley UNIX Later adopted by all popular operating systems Simplifies porting applications to different OSes

In UNIX, everything is like a file All input is like reading a file All output is like writing a file File is represented by an integer file descriptor

System calls for sockets Client: create, connect, write, read, close Server: create, bind, listen, accept, read, write,

close Winsock Programmer's FAQ:http://tangentsoft.net/wskfaq/newbie.html#interop

Typical Client Program

12

Prepare to communicate Create a socket Determine server address and port number Initiate the connection to the server

Exchange data with the server Write data to the socket Read data from the socket Do stuff with the data (e.g., render a Web

page) Close the socket

Socket programming with UDP

Application Layer13

UDP: no “connection” between client and server

no handshaking sender explicitly attaches IP

address and port of destination to each packet

server must extract IP address, port of sender from received packet

UDP: transmitted data may be received out of order, or lost

application viewpointUDP provides unreliable

transfer of groups of bytes (“datagrams”)

between client and server

Client/server socket interaction: UDP

Application Layer14

closeclientSocket

Server (running on hostid)

read reply fromclientSocket

create socket,clientSocket =

DatagramSocket()

Client

Create, address (hostid, port=x,send datagram request

using clientSocket

create socket,port=x, for

incoming request:serverSocket =

DatagramSocket()

read request fromserverSocket

write reply toserverSocket

specifying clienthost address,port number

Creating a Socket: socket()

15

Operation to create a socket SOCKET WSAAPI socket ( int af, int type, int protocol);

af An address family specification. only format currently supported is PF_INET, which is the ARPA Internet address format.

Type: semantics of the communication SOCK_STREAM: reliable byte stream SOCK_DGRAM: message-oriented service

Protocol: specific protocol 0: unspecified (PF_INET and SOCK_STREAM already implies TCP,

PF_INET and SOCK_DGRAM implies UDP).

Establishing the server’s name and port - history

16

struct sockaddr_in server; server.sin_family = AF_INET; server.sin_addr.s_addr =

inet_addr(IPstring); server.sin_port =

htons(ServerPort);

Sending Data int WSAAPI

sendto ( SOCKET s, const char FAR * buf, int len, int flags, const struct sockaddr FAR * to, int tolen );

s A descriptor identifying a (possibly connected) socket.

buf A buffer containing the data to be transmitted.

len The length of the data in buf. flags Specifies the way in which the call is

made. to An optional pointer to the address of the

target socket. tolen The size of the address in to.

Slides by Jennifer Rexford20

Receiving Data int WSAAPI recvfrom (

SOCKET s, char FAR* buf, int len, int flags,

struct sockaddr FAR* from,int FAR* fromlen

); s A descriptor identifying a bound socket. buf A buffer for the incoming data. len The length of buf. flags Specifies the way in which the call is made. from An optional pointer to a buffer which will hold the source

address upon return. fromlenAn optional pointer to the size of the from buffer.

Slides by Jennifer Rexford21

Byte Ordering: Little and Big Endian

22

Hosts differ in how they store data E.g., four-byte number (byte3, byte2, byte1, byte0)

Little endian (“little end comes first”) Intel PCs!!! Low-order byte stored at the lowest memory location Byte0, byte1, byte2, byte3

Big endian (“big end comes first”) High-order byte stored at lowest memory location Byte3, byte2, byte1, byte 0

IP is big endian (aka “network byte order”) Use htons() and htonl() to convert to network byte

order Use ntohs() and ntohl() to convert to host order

Why Can’t Sockets Hide These Details?

23

Dealing with endian differences is tedious Couldn’t the socket implementation deal with this … by swapping the bytes as needed?

No, swapping depends on the data type Two-byte short int: (byte 1, byte 0) vs. (byte 0, byte

1) Four-byte long int: (byte 3, byte 2, byte 1, byte 0) vs.

(byte 0, byte 1, byte 2, byte 3) String of one-byte charters: (char 0, char 1, char 2,

…) in both cases Socket layer doesn’t know the data types

Sees the data as simply a buffer pointer and a length Doesn’t have enough information to do the swapping

Servers Differ From Clients

24

Passive open Prepare to accept connections … but don’t actually establish one … until hearing from a client

Hearing from multiple clients Allow a backlog of waiting clients ... in case several try to start a connection at once

Create a socket for each client Upon accepting a new client … create a new socket for the communication

Typical Server Program

25

Prepare to communicate Create a socket Associate local address and port with the socket

Wait to hear from a client (passive open) In TCP: Indicate how many clients-in-waiting to

permit Accept an incoming connection from a client

Exchange data with the client over new socket Receive data from the socket Do stuff to handle the request (e.g., get a file) Send data to the socket Close the socket

Repeat with the next connection request

Server Preparing its Socket

26

Bind socket to the local address and port number (Associate a local address with a socket)

#include <winsock2.h> int WSAAPI bind ( SOCKET s, struct sockaddr* name, int

namelen);

Arguments: socket descriptor, server address, address length

 s A descriptor identifying an unbound socket.

name  The address to assign to the socket. Often, you bound your listening socket to the special IP address INADDR_ANY. This allows your program to work without knowing the IP address of the machine it was running on, or, in the case of a machine with multiple network interfaces, it allows your server to receive packets destined to any of the interfaces. When sending, a socket bound with INADDR_ANY binds to the default IP address, which is that of the lowest-numbered interface.

Namelen The length of the name. Returns 0 on success, and -1 if an error occurs

Putting it All Together

27

socket()

bind()

listen()

accept()

read()

write()

Server

block

processrequest

Client

socket()

connect()

write()

establish

connection

send request

read()

send response

Serving One Request at a Time?

28

Serializing requests is inefficient Server can process just one request at a time All other clients must wait until previous one is done

Need to time share the server machine Alternate between servicing different requests

Do a little work on one request, then switch to another Small tasks, like reading HTTP request, locating the

associated file, reading the disk, transmitting parts of the response, etc.

Or, start a new process to handle each request Allow the operating system to share the CPU across

processes

Or, some hybrid of these two approaches

TCP info for later in the course

29

Socket-programming using TCP

Application Layer30

Socket: a door between application process and end-end-transport protocol (UCP or TCP)

TCP service: reliable transfer of bytes from one process to another

process

TCP withbuffers,variables

socket

controlled byapplication

developer

controlled byoperating

system

host orserver

process

TCP withbuffers,variables

socket

controlled byapplicationdeveloper

controlled byoperatingsystem

host orserver

internet

Putting it All Together

31

socket()

bind()

listen()

accept()

read()

write()

Server

block

processrequest

Client

socket()

connect()

write()

establish

connection

send request

read()

send response

TCP info for later in the course

32

Serving One Request at a Time?

34

Serializing requests is inefficient Server can process just one request at a time All other clients must wait until previous one is done

Need to time share the server machine Alternate between servicing different requests

Do a little work on one request, then switch to another Small tasks, like reading HTTP request, locating the associated

file, reading the disk, transmitting parts of the response, etc.

Or, start a new process to handle each request Allow the operating system to share the CPU across processes

Or, some hybrid of these two approaches

Blocking and non blocking

35

"block" is a techie jargon for "sleep“ Lots of functions block. accept() blocks. All the

recv() functions block. Why? Because we programmed them this way:

When you first create the socket with socket(), the kernel sets it to blocking.

Can we set the socket to be non-blocking? What does it mean? Yes. It means you have to poll it for data (check on

it) If there’s no data yet, you get -1 (err) and have to

try again. Isn’t that CPU intensive? Yes! So, you can use select()…

Reading or writing to multiple sockets

36

A server wants to listen for incoming connections as well as keep reading from the connections it has.

Select – a way to monitor several sockets at the same time, and know their situation.

select()—Synchronous I/O Multiplexing Manipulates sets of sockets and lets you know: which ones are ready for reading which are ready for writing How? By creating and handling of sets of sockets,

and obtaining additional information for each.

Manipulating Sets for the select() func.

37

FD_SET(int fd, fd_set *set); Add fd to the set.

FD_CLR(int fd, fd_set *set); Remove fd from the set.

FD_ISSET(int fd, fd_set *set); Return true if fd is in the set.

FD_ZERO(fd_set *set); Clear all entries from the set.

Note: fd in Windows in the Socket structure

Additional select() info

38

What happens if a socket in the read set closes the connection? Well, in that case, select() returns with that socket descriptor set as "ready to read". When you actually do recv() from it, recv() will return 0. That's how you know the client has closed the connection.

One more note of interest about select(): if you have a socket that is listen()ing, you can check to see if there is a new connection by putting that socket's file descriptor in the readfds set.

Wanna See Real Clients and Servers?

39

Apache Web server Open source server first released in 1995 Name derives from “a patchy server” ;-) Software available online at http://www.apache.org

Mozilla Web browser http://www.mozilla.org/developer/

Sendmail http://www.sendmail.org/

BIND Domain Name System Client resolver and DNS server http://www.isc.org/index.pl?/sw/bind/

…

A final note: what is Peer-to-Peer Communication

40

No always-on server at the center of it all Hosts can come and go, and change addresses Hosts may have a different address each time

Example: peer-to-peer file sharing Any host can request files, send files, query to

find where a file is located, respond to queries, and forward queries

Scalability by harnessing millions of peers Each peer acting as both a client and server