Que sont les réseaux informatiques et comment les comprendre réellement

Que vous soyez nouveau dans le monde du développement, ou que vous construisiez des choses depuis longtemps - ou même si vous êtes une personne qui aime juste les ordinateurs et utilise Internet quotidiennement - vous devez connaître les bases du réseautage et plus particulièrement Réseaux informatiques.

Si vous aimez approfondir vos connaissances sur les serveurs, leur sécurité et comment vous vous connectez à vos serveurs à partir d'un client distant, tout cela nécessite une certaine connaissance des réseaux informatiques et de leurs composants. J'ai essayé de couvrir la plupart des sujets concernant les réseaux informatiques dans cet article.

De plus, à partir de là, je ferai référence aux «réseaux informatiques» simplement comme aux «réseaux».

Regardons d'abord ma définition de travail des réseaux informatiques:

Les réseaux informatiques peuvent être définis comme l'échange de paquets réseau entre des machines informatiques à travers le monde à l'aide de lignes de données telles que des câbles métalliques, des fibres optiques, etc.

L' Internet est une sorte de réseau informatique. Sorta.

Nous examinerons certains termes et composants couramment utilisés et leur fonctionnement dans un réseau informatique, dont certains sont indiqués dans le diagramme ci-dessus.

Termes couramment utilisés dans les réseaux informatiques

Nœuds

Les nœuds dans les réseaux informatiques désignent tout appareil informatique tel que des ordinateurs, des téléphones mobiles, des tablettes, etc. qui essaient d'envoyer et de recevoir des paquets réseau à travers le réseau vers un autre appareil similaire.

Paquets réseau

Les paquets réseau ne sont rien d'autre que les informations ou unités de données qu'un nœud source souhaite envoyer / recevoir vers / depuis le nœud de destination. Dans cet article, les paquets réseau / paquets de données ont tous la même signification.

Protocole Internet (IP)

Considérez que vous souhaitez envoyer un cadeau d'anniversaire à votre ami le jour de son anniversaire, où allez-vous l'envoyer? À leur adresse, non?

Il en va de même ici. Les premiers informaticiens voulaient identifier les ordinateurs sur Internet avec un numéro unique, quelque chose comme un numéro de téléphone aujourd'hui. Ainsi, ils ont proposé le concept de TCP / IP.

L'adresse IP d'un périphérique informatique est l'adresse de ce périphérique dans un réseau informatique. Techniquement, c'est un nombre de 32 bits utilisé qui identifie les périphériques dans un réseau. Toutes les communications aller et retour de l'appareil dans ce réseau seront effectuées en fonction de son adresse IP.

Considérez que vous téléchargez un fichier sur n'importe quel site ou dites à Google Drive.

Parlant au niveau le plus bas de la communication réseau, votre fichier est converti en paquets et chaque paquet a l'adresse du nœud de destination avec lui qui n'est rien d'autre que l'adresse IP.

À un niveau supérieur, les adresses IP sont classées en deux types:

  • IPv4 : les adresses IPv4 sont de 32 bits (quatre octets) comme expliqué dans la définition. Un exemple d'adresse IPv4 serait 104.244.42.129 qui est l'adresse IPv4 de twitter.com .Ils sont stables à utiliser et sont donc utilisés aujourd'hui pour identifier les machines dans le monde.
  • IPv6 : Les adresses IPv6 sont assez nouvelles dans le monde et sont essentiellement huit nombres hexadécimaux séparés par «:». Un exemple d'adresse IPv6 serait 2001: 0cb8: 85a3: 0000: 0000: 8a2e: 0370: 7334 . Ils sont instables et ne sont donc pas encore largement utilisés. Le Web utilise toujours IPv4 en raison de sa stabilité et il n'y a pas d'estimation du moment où nous commencerons à utiliser IPv6 car il n'est pas stable pour le moment.

IPv4 est classé en cinq classes nommées Classe A, B, C, D, E.

Classe A : comme indiqué dans la troisième colonne de l'image ci-dessus, pour une adresse IP de classe A, le premier bit du premier octet de l'adresse IP est constant et vaut «0».

La deuxième colonne indique les bits de réseau et les bits d'hôte de la classe d'adresse IP correspondante. Considérons dans le cas d'une adresse IP de classe A, nous avons la formule suivante:

Nombre de réseaux / sous-réseaux = 2^(# of network bits).

Nombre d'hôtes valides dans chaque sous-réseau = 2^(# of host bits) — 2.

Le nombre de bits de réseau et de bits d'hôte est déterminé par le masque de sous-réseau par défaut de la classe d'adresse IP.

Le masque de sous-réseau par défaut pour une adresse IP de classe A est 255.0.0.0, c'est-à-dire 11111111.00000000.0000000.00000000`. Ainsi, pour la classe A:

Bits réseau = 8 et bits hôte = 24.

Puisque les bits de réseau = 8 , les bits d'hôte = 24 , leur somme doit être de 32, car les adresses IPv4 sont de 32 bits. Mais, puisque nous utilisons le bit (premier bit du premier octet) pour identifier la classe:

Nombre de bits réseau utilisables = Nombre de bits réseau - Nombre de bits constants = 8–1 = 7

Ainsi, le nombre de réseaux possibles en classe A =2^7 — 2 = 126 et,

Nombre d'hôtes possibles (c'est-à-dire des périphériques qui peuvent être connectés au réseau) par réseau en classe A = 2^24-2 = 16277214.

Maintenant, ici, pour la classe A, vous vous demandez peut-être pourquoi j'ai soustrait 2 supplémentaires du nombre de réseaux possibles. C'est parce que, pour la classe A, 127.xyz était réservé. Pour les autres classes, la formule habituelle est utilisée.

Ainsi, les adresses IP de la classe A vont de 1.x.x.xà 126.x.x.x.

Classe B: le cas est similaire à la classe B.La seule différence est que 2 bits du premier octet sont constants (10) et ils identifient la classe d'adresse IP qui est la classe B.Tous les autres calculs sont les mêmes, et je ne mentionne pas les ici car ils sont faciles à saisir dans le tableau ci-dessus. Ils vont de 128.0.x.xà 191.255.x.x.

Classe C : 3 bits du premier octet sont constants (110) et identifient la classe comme classe C. Ils vont de 192.0.0.xà 223.255.255.x.

Classe D et Classe E : Les classes D et E sont utilisées à des fins expérimentales.

Les adresses IPv4 sont principalement de deux types:

  • Statique : ces adresses IP sont celles qui restent constantes pour un appareil au fil du temps. Des exemples de ceux-ci sont les serveurs distants que nous utilisons pour héberger nos applications, sites Web, etc. où nous utilisons le client ssh pour ssh sur notre serveur.
  • Dynamic: Generally, these are the IP addresses that a common computer in an Internet network is assigned. Try switching your router off and you will see a change in the IP address of your computer! (But only after reading this article ?). Now, you may be thinking who allocates these IP addresses? It is the DHCP (Dynamic Host Configuration Protocol) server which is explained briefly further in this article.

Note: A device can have multiple IP addresses at the same time. Consider a device connected to two networks, wifi as well as any LAN network — it will have two IP addresses. This implies that the IP addresses are assigned to the interfaces and not directly to the computer.

Okay, so far so good. Let’s continue.

Routers

As its name suggests, a Router is a hardware component that takes care of routing packets. It determines which node the packet came from and which destination node the sender node want to send it to. No computer knows where other computers are located, and packets are not sent to every computer. A Router identifies the destination node address to which a network packet has to be sent and it forwards it to the desired address.

Routers have a specific “Routing Protocol” which defines the format in which they exchange data with another router or networking nodes. In other words, routing protocol defines how routers communicate with each other.

Routers build up a “Routing Table” which identifies the most optimized paths to be taken in the network while sending packets.

Technically, a routing table is just a table with the list of “routes” from one router to other. Each route consists of the address of the other routers/nodes in the network and how to reach them.

Routing table:
Destination Gateway Genmask Flags Metric Refs Ifacedefault 192.168.0.1 0.0.0.0 UG 1024 233 eth0192.168.0.0 * 255.255.255.0 UC 0 0 wlan0192.168.0.0 * 255.255.255.0 UH 0 2 eth0

Above is an example of a routing table. The key points to take a note of here are:

  • Destination: This is the IP address of the destination node. It indicates where the network data packet should end up.
  • Gateway: Gateway is the component which connects two networks. Consider that you have a router connected to another router. Each of the routers has devices connected to it. So, the address of the last router (say R1 here) after which the network packet enters the other network (say R2’s network) is called the gateway. Usually, the gateways are nothing but the routers. Let me give one more example: say that your room is one network and your sibling’s room next to yours is another network, then the “door” between the two rooms can be considered the gateway. People sometimes refer to the “routers” as the gateway, because, that’s what they are, “a gateway to another network”.
  • Genmask/Subnet mask: It is nothing but the net/subnet mask. A subnet mask is a number which when combined with an IP address allows you to divide the IP space into smaller and smaller chunks for use in both physical and logical networks. The explanation of how subnet mask calculations happen is beyond the scope of this article.
  • Flags: Different flags have a different meaning. For example, in the first route, “U” in “UG” means the route is UP, whereas “G” in “UG” means GATEWAY. Since the route signifies a GATEWAY, it is a door to the other network. Whenever we send any data through this route, it gets sent to another network.
  • Iface (Network interface): Network interface refers to the network that the route defined in the routing table is having the destination computer in. That is if you are connected to Wifi, then it would be “wlan” and when you are connected to a LAN, then it would be “eth”.

So this is the way a router works, with the help of Routing Protocol and Routing Table.

All good up to now. But, you must be thinking —

“Okay! But hey, we are learning about components here. I need to stitch them together and get to know how the internet works.”

Cool! Some more terms and you will have a proper understanding of how everything goes.

Network Address Translation (NAT)

Network address translation is a technique used by routers to provide internet service to more devices with less usage of public IPs. Thus, a router is assigned a single IP address by the ISP and it assigns the private IPs to all the devices connected to it. NAT helps the ISPs provide internet access to more consumers.

Thus, if you are connected to the router of your house, your public IP will be visible to the world, but the private one will not. Whatever network packets are communicated will be addressed by your public IP (that is the public IP assigned to the router).

Consider the above figure. Let’s say that in your home network, you are trying to access medium.com (remote static IP: 72.14.204.147), from your computer (private IP: 192.168.1.100).

So, for your computer, the connection looks like:

192.168.1.100:3764172.14.204.147:80 .

“37641” is the random port number assigned by NAT router to your device/computer. (When there is network communication between daemons running on different ports on a computer, the respective port is used by NAT). Each outbound connection gets an assigned port by the NAT router.

The connection is established in NAT like:

Private IP |PrivatePort |PublicIP |PublicPort |Remote |RemotePort
------------- ------------ --------- ----------- ------- -----------
192.168.1.100 | 37641 | 104.244.42.129 | 59273 | 72.14.204.147 | 80

But, since the outside world of the network doesn’t know about your private address, the connection looks like the following to medium.com:

104.244.42.129:5927372.14.204.147:80 .

That way, we achieve assigning a higher number of IP addresses without wasting many public IPs.

Now, when medium.com sends the response back to 104.244.42.129:59273 , it travels all the way to your home router which then looks up for the respective private IP and private port and redirects the packet to your device/computer.

Note: NAT is a generalized concept. NAT can be achieved as 1:1, 1:N where 1, N are the number of IP addresses in the network. A technique called as “IP Masquerading” is a 1:N NAT.

Dynamic Host Configuration Protocol (DHCP)

Dynamic Host Configuration Protocol or DHCP is responsible for assigning dynamic IP addresses to the hosts. The DHCP server is maintained by the ISP or previous router if there is a chain of routers to reach the host.

Thus, allocation of IP addresses is carried out by the DHCP server. Generally, ISP maintains a DHCP server and the routers in our houses get assigned a public IP from the DHCP server.

Note: Whenever a router or say a DHCP server maintained by an ISP or router restarts, the IP address allocation starts again and devices are allocated IPs which are different than the previous ones.

Domain Name System/Server

We have already discussed that any machine is identified by the IP address.

Okay, so you are running a web server on your localhost on your machine. If you have dug around in the hosts on any Linux machine, you would have encountered something like this:

127.0.0.1 localhost255.255.255.255 broadcasthost::1 localhost

which means that even if you type 127.0.0.1 in your browser’s URL bar, it would mean the same as localhost .

Similar to the above, the websites you use daily are web servers running on some remote instance/node having a static IP address. So, typing that IP address in your browser’s URL bar will take you to the website?

Yes, surely it will. But, are you a superhuman to remember the IP addresses of thousands of sites?

NO.

Thus, there come the domains that we use, say medium.com, twitter.com, behance.net, codementor.io, etc.

A Domain Name Server is a server having huge records of domain name mapping IP addresses which searches for the domain input and returns the respective IP address of the machine on which the website you want to access is hosted.

How does DNS work actually?

  1. DNS is managed by your ISP (internet service provider).
  2. When we type an URL in the address bar, the data packets travel through your router, maybe multiple routers to your ISP where your DNS server is present.
  3. DNS server present at the ISP looks up for the domain in its database. If an entry is found, then it returns it.
  4. If any entry is not found in its primary database that it maintains, the DNS server will travel through the internet to another DNS server maintained by another ISP and check if the entry is available in that another DNS server’s database. Along with returning the IP address taken from another DNS, it will update the primary database with this new entry also.
  5. Thus, sometimes (very rarely) a DNS server may have to traverse to multiple DNS servers to get a matching entry.
  6. If after traversing a lot of DNS servers across the internet, it doesn’t get a matching entry, then the DNS server throws an error indicating that the “domain name is invalid or doesn’t exist”.

Note:

The Internet Corporation for Assigned Names and Numbers (ICANN) is a consortium (a non-profit corporation) that manages the assignment of domain names and IP address ranges on behalf of the community.

A domain is divided into three parts as shown in the following figure.

  1. Protocol: The protocol used to access the website, for example, HTTP, HTTPS, etc.
  2. Domain name: The main domain name in our domain. This can be anything that is available as per the ICANN registry.
  3. Domain extension: This is one which is considered important while buying a domain. Generally, it is classified into two types:
  • Generic Top-level Domains (gTLDs): This includes most popular domain extensions like .com, .org, .net, .edu, .co, etc.
  • Country Code Top-level Domains(ccTLDs): These indicate that the domain is related to the country code specified in the domain extension. For example, “.in” indicates that the website is originated from India. Also, some of the ccTLDs require that the person purchasing the domain should be from the same country. Most of the small country code extensions are not searchable from outside that country.

Internet Service Providers (ISPs)

Internet Service Providers are the companies that provide everyone Internet. The article you are reading now is because of the internet that your ISP provides you.

ISPs provide internet, handle routing your requests to the correct destination, resolve domain names with the help of DNS cache that they maintain, and handle all this network infrastructure which enables us to use the internet.

ISP is a hierarchical thing working across the internet. There are certain types of ISPs namely Tier 1, Tier 2, Tier 3 ISPs.

  • Tier 1 ISPs are the ones which connect major networks on the internet. Consider them as the major highways of the internet. They are connected to almost every network on the internet. Also, they provide internet access to the Tier 2 ISPs. ex. CERFNet, UUNet, PSINet. They are also called Network Service Providers. These ISPs are connected to each other by means of large cables going beneath the sea.
  • The Tier 2 (Regional) ISPs are the ones who primarily provide Internet services to organizations, consumers (that is “us”) or the Tier 3 ISPs. The internet connection you are using is from a Tier 2 ISP. However, organizations can also get Internet access from Tier 1 ISPs.
  • Tier 3 (Local) ISPs are just like Tier 2. It’s just one more level of hierarchy out there that purchases bandwidth from Tier 2 ISP and sells it to consumers.

The traffic that goes through your router also goes through Tier 3 (if present), Tier 2, and ultimately through Tier 1 ISPs all the way to another network.

Woot Woot! I am happy that you are still with me. We will put all the things together now.

Putting all of the above things together

Up until now, we have learned about all the components needed to make everything work. Now, we will glue them together.

Let’s summarize all the things we’ve learned:

  • When a computer/device comes online, it gets a private IP assigned by the router. The router gets a public IP from the ISP.
  • Other devices in the network are allocated unique private IPs.
  • ISPs are the ones who are present across the world and are connected to each other. They sell Internet services to the regional and local ISPs, from whom we, the consumers, purchase Internet.
  • Thus, when a device tries to establish a network connection with some other device on some other network, it does it with the identity of its gateway (the router). The router then maps the private IP and private port number with the public IP and random high integer public port number.
  • The router then sends the packets to the desired destination where some other router or gateway does the same thing as the previous router and analyses which computer/device that packet came from.
  • The remote computer/device responds by sending the destination as the public IP and public port of the router.
  • The router then again checks for the private IP and private port and forwards the network packets.

So, this is how the Internet aka A kind of Computer Network using TCP/IP protocol works.

Thanks for reading the article. If you have any questions, please feel free to ask them in the comments below.

If you liked this article, please give me some ? and share it with others. See you in the next one. You are awesome!

Also, if you’d like to support me ?

Buy Sumedh Nimkarde a Coffee - BuyMeACoffee.com

Hello, I am Sumedh and my work is to build, break and rebuild things.www.buymeacoffee.com

Thanks a lot again for reading! Feel free to connect with me on Twitter, GitHub.