Basic of computer Network 2: Network Architecture (Virtual Connection ) （网络架构）

The computer network is composed of computer systems, communication links, and network nodes. It is the field that combines computer technology and communication technology, which undertake data communication and data processing these two kinds of work. Logically, the internet can be divided into resource subnet and communication subnet. Resource subnet provides the ability to access the internet and data process, it consists of hardware (DTE: Data terminal equipment) and software (like Browser); the Communication subnet provides the function of communicating, it consists of DCE (Data Communication Device) and protocol. The structure of communication subnets can affect network structure directly. The technology used in the communication subnet can be divided into Peer to peer, Client/Server (Tiered), Fat/Thin Client.
Overall, It is a blueprint that provides an architectural and technical foundation for designing, building, and managing a communications network. Network architecture defines every aspect of a data network communication system, including, but not limited to, the type of interface used by the user, the network protocol used, and the type of network cabling that may be used.

Peer-to-Peer (P2P)

Peer-to-peer is also known as a workgroup network. Unlike enterprise professional networks, which are controlled by domains, there are no “domains” in P2P, only “workgroups”. Obviously, the concept of “workgroup” is not as broad as the concept of “domain”, so the number of users that P2P can accommodate is also very limited. P2P network is relatively simple, and the number of computers usually does not exceed 20. A peer-to-peer network is a common networking method for small LANs.

In a peer-to-peer network, all computers have the same function, and there is no master or slave, which can be described in one word “Equality”. Any computer on the network not only can serve as a network server and provide resources for other computers; it also can act as a workstation to share resources from other servers. Any computer can serve as both a server and a workstation, or just one. Because peer to peer networks do not require dedicated servers for network support or other components to improve network performance, that means p2p relatively inexpensive. Overall, Peer to Peer is a kind of practical small network system with less investment, it has a quick effect and high-cost performance, if there is no special need, it is suitable for family or small office.

Advantage

1. Low network cost
2. Easy to set up and use
3. Simple network configuration
4. Easy to maintenance
5. Fewer single points of failure (topology
   dependent)

Disadvantage

1. Low network performance;
2. Because any computer can be a server which means the user must remember which computers in the workgroup have shared resources or information they want to access and log into each of them to access the data;
3. When files backup, it needs to backup separately on each computer;
4. Like most of the network system, unsafe unsigned code victims may allow remote access to the file on the computer, and even endanger the whole network;
5. User need to set up anti-virus software on each machine

Client/Server (C/S)

The Client/Server structure usually has two layers. The server is responsible for managing the data, and the client is responsible for interacting with the user. For example, when a user uses Google to search an article, the user’s machine and web browser are the clients. The client connects to the server through the LAN, receives the user’s request, and requests the server through the network, and operates on the database. At the same time, the computers, databases, and applications that makeup Google act as servers. The server accepts the client’s request and submits the data to the client. Then, the client the result presents to the user. The server also has to provide security protection and data integrity processing operations and allow multiple clients to access the server at the same time, which puts forward a high demand on the server’s hardware data processing capabilities.

In the Client/Server structure, the application is divided into two parts: the server part and the client part. The server part is the information and functions Shared by multiple users, performing background services, such as controlling the operation of the Shared database; The client part is exclusive to the user and is responsible for performing the foreground function. It has powerful functions in error prompt, online help, and other aspects, and can switch freely between subroutines.

Client/Server structure has been very mature in technology, its main characteristics are strong interactivity, with secure access mode, fast response speed, to process a large number of data. However, the Client/Server structure is lack of universality, system maintenance, and upgrade need to be redesigned and developed, which increases the difficulty of maintenance and management, and further data expansion is more difficult, so the C/S structure is only limited to small LAN.

Advantage:

1. It facilitates the realization of resource sharing. Resources in Client/Server model are also distributed, the users can not only access the resources on the server and local workstation, but also enjoy the resources on other workstations, realizing resource sharing;
2. It is more secure than Peer to Peer;
3. It is easy to maintain;
4. It can process information quickly. Because Client/Server is a point-to-point operating environment, when a task requests processing, the load of that task can be evenly distributed across all possible servers. In this way, requests made on the client side can be processed by multiple servers in parallel, providing extremely fast response time and high transaction throughput for each request;
5. It can work with multiple different clients of different specifications.

Disadvantage:

1. Lack of effective security. Because the client is connected to the server directly, when the client accesses some sensitive data, the user can directly access the central database, which may cause the modification or loss of sensitive data.
2. Server-side work low efficiency. Because each client has to connect to the server to access data resources, this makes the server consume a lot of server resources which are already very tight due to the connection established by the client access, resulting in the low efficiency of the server;
3. Easy to cause network congestion. Every time when the client visits the server, which makes the network traffic increase sharply and causes the network block easily.
4. Compare to P2P, in Client/Server model, if one server fails, the request form client cannot be served; while in P2P, if one node failed, the user can still be served by other nodes.

Thin Client

The thin client is to use professional embedded processors, small local flash memory, a simplified operating system based on PC industry-standard design of small special commercial PC. Generally, there is no hard disk, CD-ROM, and an expansion slot. Most resources are acquired through the network, and the processing capacity and application loading are dependent on the server. From the point of view of these structures, thin clients are small and portable.

Moreover, the user cannot remove the input/output options for flash memory which use to store OS (Operating System), local system memory, network adapters, and so on. There are no removable parts that provide a more secure and reliable environment than ordinary PCs.

In the education area, it is suitable for library, student’s dormitory which would help student for their study to get more useful information form browser, and allows them to communicate by using e-mail; in the public area, it is suitable for airport, train station, hospital, etc. it allows user to look-up the timetable, allows staff uses FTP to do the update for any time change caused by some reason, allows the doctor to store the patients’ record in the system and so on.

Instead of putting a lot of logical processing on the client-side, the thin client focuses on processing the UI data presentation, so that the client only needs simple document data validation logic and form data submission.

A large amount of business logic will be processed through the form data after receiving the relevant request on the server-side, and then the result will be packaged into HTML and returned to the client for parsing.

In this way, the client-side will contain less logical code and the server-side will have to withstand the pressure.

Advantage:

1. Because Thin client doesn’t have hard drive or floppy drive, it can protect against unauthorized software use and virus;
2. During the transmission, the only changes are the information from screen, mouse and keyboard that network bandwidth requirements are low;
3. It is cheaper than original PCs;
4. Because all applications are run and stored on the server, it doesn’t need to be upgraded which reduce the maintenance costs

Disadvantage:

1. It can’t work offline; It can’t process its own data or programs locally;
2. It requires high-quality servers to support them work;
3. If the server does down, all clients that connect to it will be affected;
4. It is not suitable for graphic designers and others who often use multimedia content or graphically intensive software;
5. It is not suitable for massively multiple-player online game.

Fat Client

A Fat Client sometimes called a Thick client is a network computer with rich resources installed locally, rather than spreading them around the network as a thin client would. For example, many PCS (personal computers) are fat clients because they have their own hard drives, CD/DVD drives, software applications, etc.

The fat client is relative to the thin client, which divides the processing of the application into two parts: some tasks are processed on the Client-side, and the rest of the tasks are processed on the Server Side.

The application on the client-side of the Fat client is responsible for most of the business logic processing in addition to presenting the program’s UI interface to the user for interaction. This type of application needs the client part to have the ability to execute the task, the request to the client machine is relatively high, but can reduce the server to send a part of the pressure, reduce the request to the server performance.

Typical fat client contain one or more on the server and client application running on a PC, client installed on the PC function rich interactive user interface, the user can through the interface to view and manipulate data, such as business affairs, to share the server part or all of the business logic processing (so fat client would seem to be very “heavy”, so called fat client, also called the rich client).

The server portion typically handles access to the data center while also handling some or all of the business logic.

Overall, the difference between a fat client and a thin client is whether a large amount of business logic needs to be placed on the client.

In C/S structure, the main logical program is placed in the client, most of the time the server refers to the database server, so it is a fat client

In the B/S (Browser/Server) structure, the business logic is all on the server, the client only has a browser, and the server sends to the client only strings that follow the HTML rules, which is the thin client. Of course, if you’re using an applet or something like that you’re putting some business logic on the client-side.

Reference:
Deepanshi Verma, What are the disadvantages and advantages of a peer-to-peer (P2P) network? 26 OCT 2018, Available from: https://www.quora.com/What-are-the-disadvantages-and-advantages-of-a-peer-to-peer-P2P-network [25/02/2020]