Shulou(Shulou.com)11/24 Report--

As the saying goes, "laymen watch the hustle and bustle, experts watch the way". Being able to surf the Internet and knowing how to surf the Internet are two different things.

You must be dissatisfied with this, and the archivist will ask you the most basic question:

What supports us to browse on Douyin, chase dramas and buy in the studio?

Is it a wallet? Is it thick hair? Is it "the Internet"?

That's too easy for you to think!

And the server ~

In our daily online news, online chat, online video playback, we have to go through the server, by the server to unify the organization and response to our needs.

You may still have a lot of question marks:

What is a server?

How does the server work?

What are the types of servers?

How is server performance measured?

You are so smart, after reading this article, you must be able to find answers to these questions.

01. What is a server? The English name of the server is "Server". As the name implies, it refers to the network equipment that provides various services. As the node of the network, the server is responsible for storing and processing 80% of the data and information on the network, so it is also called the soul of the network.

The server has a similar function to an ordinary computer. But compared with the ordinary computer, the server requires higher stability, security, professional performance and so on.

In terms of functionality, the main differences between a server and an ordinary computer include:

Support one-to-many communication mode: the server can support multiple terminal connections, such as: PC, tablet, mobile phone and other fixed or mobile network terminals. For example, our Singles' Day this year, tens of millions of people participated in the 0: 00 second kill, behind it can not be separated from the powerful server to provide services.

Resources can be shared through the network: by listening for service requests submitted by other terminals on the network, the server provides the connected hard disk, printer, Modem and various special communication equipment to the customer sites on the network under the control of the network operating system, and can also provide network users with services such as centralized computing, information publication and data management.

The hardware performance is more powerful: the high performance of the server is mainly reflected in the high-speed computing power, long-term reliable operation, strong external data throughput and so on.

02. How does the server work? The document gives you an example from life to show you how the server works:

There are always more than 300 days of shopping festival in a year. I guess you must have been to the express station to pick up the delivery.

So what is the process of picking up express delivery?

1) know where the express station is, and then walk there. -- establish a connection

2) report the pick-up code to the courier. -- raise a demand

3) the delivery boy finds the express delivery according to the pick-up code. -- providing services

4) the courier will find the courier and give it to you. Get the result

So what does the working principle of the server have to do with picking up and delivering?

Let's review the daily routine of surfing the Internet on mobile phones every day, assuming that you want to look up discount information on June 18 this year.

1) when we open App, it is equivalent to arriving at the express station, that is, establishing a connection with the server.

2) Click the "discount information" button, which is tantamount to reporting the pick-up code to the courier, that is, putting forward the demand.

3) the query system running on the server is like a courier. According to your request, finding the "discount information" you want to see in the App database is like querying the location of the express delivery according to the pick-up code, that is, providing services.

4) the server needs to send the queried information back to your mobile phone, and the transmission program starts to work. This process is like a courier giving you the express delivery, that is, getting the result.

In this process, the part that provides the service is the server, the store full of parcels is the file storage of the server, the pickup code record book of the courier is the database of the server, and the courier is the interface of the server, referred to as API.

03. What are the performance indicators of the server? After understanding the basic working principle and architecture of the server, how to evaluate the performance of a server? The document will introduce several common server performance indicators RT, Throughput, TPS, QPS.

RT (response time): the processing time of a transaction. Usually refers to the time interval, in milliseconds, from a request sent to the server to return after processing, and then to the completion of receiving the reply data.

Throughput (throughput): throughput refers to the number of requests processed by the system in a unit time. TPS and QPS are commonly used quantitative indicators of throughput.

TPS (Transactions Per Second, number of transactions transferred per second): the number of transactions processed by the server per second. It includes three processes: the user requests the server, the server processes internally, and the server returns to the user. If you can complete N of these three processes per second, TPS is N.

QPS (Queries Per Second, query rate per second): the corresponding number of queries per second by a server, which is a measure of the traffic handled by a specific query server within a specified period of time, that is, the number of response requests per second, that is, the maximum throughput capacity.

QPS is basically similar to TPS, but the difference is that one visit to a page, that is, counting one TPS; but one page request, may result in multiple requests to the server, which are counted in the QPS.

04. What are the categories of servers?

According to the product appearance, the mainstream servers in the market are: tower server, rack server, blade server.

Tower server is the most common at present, its shape is similar to the host computer, the tower server motherboard expansibility, many slots, can be very convenient for redundant expansion.

However, there is no unified standard for external dimensions, and when multiple servers work together, management is a headache.

Rack servers have a unified appearance, height in U units (1U=44.45 mm), divided into 1U, 2U, 3U, 4U, 5U, 7U, generally 1U, 2U is the most common. Use uniformly with the cabinet, save space and facilitate management.

However, due to the limitation of internal space, the expansibility and heat dissipation are poor.

Blade server, as the mainstream product of high-performance computing cluster, means that multiple card server units (system motherboard) can be inserted in the chassis of standard height, hence the name of "blade" server.

Unlike other servers, blade servers can directly integrate these system motherboards into a server cluster without going through external devices. Therefore, in the cluster mode, all motherboards can be connected to provide more convenient and efficient network services.

Tower server

Rack server

Blade appearance

Advantages

Strong redundancy

Easy to expand

Low cost of equipment

The heat dissipation problem is small

Work independently

Appearance standardization

High space utilization

Economic and efficient

Strong redundancy

High space utilization

Convenient for centralized management

Easy to expand

Easy to cluster

Shortcoming

Take up a lot of space

There is no uniform standard for appearance

Complex management

High upgrade cost

Weak expansibility

The problem of heat dissipation is prominent

The cost of equipment is high

The problem of heat dissipation is prominent

Application scenario

Small business

Large enterprises

Large data centers or large-scale computing areas, such as banking, telecommunications, finance and Internet data centers.

According to the instruction set, the server can be divided into CISC (Complex Instruction Set Computing) server, RISC (Reduced Instruction Set Computing) server and EPIC (Explicitly Parallel Instruction Computing) server.

So, what is the instruction set?

We need to know what instructions are first.

Instructions are the instructions and commands that direct the work of the machine, the program is a series of instructions arranged in a certain order, and the process of executing the program is the working process of the computer.

Instruction set is a set of instructions used to calculate and control computer systems in CPU, and each new type of CPU is designed to specify a series of instruction systems that cooperate with other hardware circuits.

Whether the instruction set is advanced or not is also related to the performance of CPU, and it is also an important symbol of CPU performance.

Compared with EPIC server, CISC server and RISC server are more common, and the two instruction sets have their own distinct characteristics, so the competition between CISC and RISC has never stopped.

Mr. documentation summarizes the characteristics of CISC and RISC for everyone.

CISC

RISC

Characteristics

There are many instructions, and one instruction performs multiple functions.

There are few instructions, and complex tasks are completed by a combination of multiple reduced instructions.

Advantages

The designated functions are performed efficiently.

Such as: multimedia processing.

The common functions have high execution efficiency and low power consumption.

Shortcoming

The system design is complicated.

Complex tasks are inefficient.

Such as: multimedia processing.

Typical architecture

X86

ARM 、 Power 、 Alpha 、 MIPS

According to the number of processors, the server can be divided into one-way server, two-way server, four-way server, eight-way server and so on.

Among them, "path" refers to the number of CPU inside a server, for example, the number of CPU inside a single server is 1, the number of two-way servers is 2, and so on. At present, the mainstream server is a two-way server.

The multiplex server uses symmetric multiprocessing technology (SymmetricalMulti-Processing, referred to as SMP). On one server, multiple CPU share memory subsystem and bus structure. When the server is running, multiple CPU are running at the same time, the system distributes the task queue symmetrically on each CPU, and all CPU can access memory, I / O and external interrupts equally, which can greatly improve the data processing capacity of the whole system.

See here, I believe you must have some understanding of the server!

Welfare time

Do you want to know what it's like in the server?

Long press to identify the QR code in the following picture, go and experience the fun of disconnecting the machine.

This article comes from the official account of Wechat: ZTE documents (ID:ztedoc)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.