WEBVTT 00:04.820 --> 00:06.350 So welcome back, guys. 00:06.350 --> 00:10.670 In this lecture, we will try to understand the concept of sockets. 00:10.670 --> 00:17.180 We will see how sockets can be used to carry out inter-process communication between the processes. 00:17.180 --> 00:23.300 The communicating processes can be processes which are running on the same machine, or they can be 00:23.300 --> 00:29.030 processes which are actually running on different physical machines and communicating over the network. 00:30.320 --> 00:34.700 So we will start with our first IPC technique that is sockets. 00:35.090 --> 00:44.150 So Unix or Linux like operating system provide an interface called socket interface to carry out communication 00:44.150 --> 00:46.670 between various types of entities. 00:47.270 --> 00:52.760 Now here entities means the processes which wants to communicate with each other. 00:55.340 --> 01:00.230 The socket interface are a bunch of socket programming related APIs. 01:00.320 --> 01:08.210 So we will discuss the bunch of socket programming APIs or functions, which is provided by the Linux 01:08.210 --> 01:09.290 operating system. 01:09.290 --> 01:18.500 And the developers uses these APIs in order to implement socket based communication in their applications. 01:19.980 --> 01:25.170 Two socket interface is nothing, but it is a collection of APIs or functions. 01:26.010 --> 01:30.900 We shall be using these APIs to implement sockets of various types. 01:31.230 --> 01:31.890 Right. 01:31.890 --> 01:39.120 So sockets are actually a generic mechanism to implement a communication between processes. 01:39.900 --> 01:44.580 So in this course we shall learn how to implement two types of sockets. 01:44.820 --> 01:47.250 The first is the Unix domain sockets. 01:47.280 --> 01:54.150 These type of sockets are used to implement or carry out inter-process communication between processes 01:54.150 --> 01:58.350 which are actually running on the same system or same machine. 01:58.470 --> 02:02.100 So such type of sockets are called Unix domain sockets. 02:02.490 --> 02:05.460 The other type of sockets are network sockets. 02:05.490 --> 02:11.910 These sockets are used to carry out communication between processes, but these processes are actually 02:11.910 --> 02:18.270 running on different physical machines and these different physical machines are deployed over the network. 02:18.840 --> 02:25.210 In fact, the socket interface can also be used to implement various other type of communications, 02:25.210 --> 02:29.050 but those are out of the scope of our course. 02:29.590 --> 02:35.560 In this module we will discuss the implementation of Unix domain sockets and network sockets. 02:36.270 --> 02:42.630 So remember, socket interface are the bunch of APIs, and these APIs can be used to implement Unix 02:42.630 --> 02:46.500 domain sockets or network sockets or any other type of sockets. 02:46.800 --> 02:55.260 The only difference is that how we actually use these APIs to implement a specific type of socket communication. 02:56.220 --> 03:00.030 So let us first cover how sockets work in general. 03:00.060 --> 03:06.420 So in general means we will going to discuss that how actually sockets work on a Linux system. 03:07.560 --> 03:12.690 It is important that you know that how sockets work internally. 03:12.690 --> 03:19.860 And then we will see how socket APIs that is socket interface can be used to implement a specific type 03:19.860 --> 03:20.960 of communication. 03:20.970 --> 03:27.000 Now where a specific type of communication means either Unix, domain socket based communication or 03:27.000 --> 03:28.890 network socket based communication. 03:29.400 --> 03:33.900 So let us first build some background on sockets. 03:35.860 --> 03:43.150 So going forward, you can see that here are the steps listed and we need to follow these steps in the 03:43.150 --> 03:48.940 same sequence as they are listed in order to implement a socket based communication. 03:51.040 --> 03:57.670 So we will discuss these steps one by one in detail, and we will learn how to implement a socket based 03:57.670 --> 03:58.780 communication. 04:00.240 --> 04:06.330 You can see that there are some function calls or APIs which are highlighted in a color. 04:07.110 --> 04:12.000 These APIs together are called socket system calls. 04:14.080 --> 04:18.730 Or socket interface, right. 04:18.940 --> 04:26.440 So using these APIs, we will going to implement different types of socket based communication that 04:26.440 --> 04:30.640 is Unix domain sockets as well as network sockets. 04:32.000 --> 04:35.300 For each of these API, we will discuss that. 04:35.300 --> 04:42.050 What is the purpose of each of these APIs and when to use and how to use while implementing socket based 04:42.050 --> 04:43.010 communication? 04:44.280 --> 04:51.780 So before diving into these steps, it is important that we study how socket based communication state 04:51.780 --> 04:57.030 machine works and various socket APIs provided by Linux operating system. 04:58.730 --> 05:04.790 So let us have some high level understanding regarding how socket based communication actually works 05:04.790 --> 05:06.470 on the Linux operating system. 05:06.770 --> 05:13.520 That is, we will discuss the state machines of socket based communication, the related APIs, the 05:13.520 --> 05:19.760 necessary system calls that is provided by Linux operating system, which facilitates us implementing 05:19.760 --> 05:21.320 socket based communication. 05:22.700 --> 05:28.490 And once we get familiar with the concepts of socket based communication, then we will walk through 05:28.490 --> 05:35.450 the example of implementing a client and server process which communicate with each other using socket 05:35.450 --> 05:36.620 system calls. 05:37.450 --> 05:39.040 Our socket interface. 05:40.540 --> 05:48.850 So next, let us try to understand that, how these socket APIs actually fits in the computer architecture 05:48.850 --> 05:49.540 diagram. 05:54.130 --> 05:59.230 So here you can see that we had a discussion on computer layer architecture in the beginning of this 05:59.230 --> 06:05.500 course, that our computer architecture is actually roughly divided into three layers. 06:05.650 --> 06:11.530 The hardware layer and the top of the hardware layer is the operating system or kernel, and the topmost 06:11.530 --> 06:13.240 layer is the application layer. 06:15.680 --> 06:16.790 Right now. 06:16.790 --> 06:20.510 You can see that in between application layer and the operating system. 06:20.630 --> 06:23.720 Sandwich is the system call interface. 06:25.030 --> 06:33.250 So this is actually broadly speaking, it is actually called a socket layer, right? 06:33.970 --> 06:35.980 Or system call interface. 06:37.130 --> 06:44.120 So you can see in this system call interface layer, there are a bunch of APIs like socket, accept, 06:44.120 --> 06:48.110 connect, send to receive from close, select and etcetera. 06:48.140 --> 06:51.890 These APIs are actually called system calls. 06:52.850 --> 06:55.310 And what is the purpose of these system calls? 06:55.340 --> 06:58.730 These system calls are invoked by the application. 06:59.090 --> 07:06.890 So by invoking these system calls, the applications actually ask the operating system for its service. 07:06.920 --> 07:14.270 For example, the application may invoke a system call and ask the operating system that, hey, I want 07:14.300 --> 07:21.710 X, Y, Z service operating system in response to the system call invocation, provide a service to 07:21.710 --> 07:22.670 the application. 07:22.670 --> 07:31.160 So so you can see here, application is like a master and the operating system acts like a slave applications 07:31.160 --> 07:39.480 orders the operating system to provide the application some service and this order is done using system 07:39.480 --> 07:41.490 call interface APIs. 07:43.220 --> 07:50.000 So Linux provides a set of APIs called System Calls, which application can invoke to interact with 07:50.000 --> 07:51.650 the underlying operating system. 07:52.940 --> 07:59.130 Socket APIs are the interface between application and operating system using these APIs. 07:59.150 --> 08:03.620 Application instructs the operating system to provide its service. 08:03.740 --> 08:09.110 So let us take a very simple example which you must have used already. 08:09.110 --> 08:12.470 Even if you don't know that you are using actually system call. 08:12.920 --> 08:15.760 So familiar example is the malloc and free. 08:15.770 --> 08:22.100 If you have ever written a single C program then you must have used malloc or free function. 08:22.100 --> 08:27.920 So these are actually system calls and you make use of the system calls in your program. 08:27.920 --> 08:35.660 So the moment your program calls a malloc, it means your program is asking the operating system to 08:35.660 --> 08:38.180 give you a program, a chunk of memory. 08:38.450 --> 08:45.140 So this is the example where your application is requesting a service from the operating system using 08:45.140 --> 08:47.270 system call called malloc. 08:48.900 --> 08:49.500 Right. 08:49.500 --> 08:56.610 And similarly, another system call is a free system call, using which your application asks the operating 08:56.610 --> 08:59.610 system to reclaim back the memory. 09:01.330 --> 09:08.590 So from this example, we clearly understand that system calls are actually a set of APIs which is invoked 09:08.590 --> 09:15.340 by the application to demand a service from the operating system or kernel. 09:17.360 --> 09:17.900 Right.