Tingyun releases 2014 China Network Performance Report

Tingyun releases 2014 China Network Performance Report

Application Performance Management (APM) is booming both at home and abroad and is being accepted by more and more enterprises. Network performance management from the perspective of real user experience is an important part of APM. Tingyun Network monitors 100 websites, mainstream CDN vendors, and five major operators in five industries in China from the perspective of real user experience, and launches the "2014 China Network Performance Report", which aims to standardize the definition of application performance management (APM) indicators in network performance, help DevOps quickly understand the real availability of the business, and improve business-level operation and maintenance capabilities.

The report will elaborate on four aspects: website access quality of websites in different types of industries, key indicators of network performance, the impact of CDN on network performance, and comparison of operator network performance.

2014 China Network Performance Report.pdf

1. Overview of access quality of websites in different industries

The access quality includes three major factors: first screen time, opening time, and opening success rate.

1. First screen time of different types of industry websites

The first screen time refers to the time it takes for the browser to display the first screen home page. If the first screen time is too long, it will reduce the user experience and cause user loss.

Optimization suggestions:

Optimize website page structure

Optimize network performance key indicators (see Part 2 for details)

2. Home page opening time of different types of industry websites

Home page opening time refers to the total time consumed to open a website home page, that is, the time from the start of DNS resolution to the browser returning the completion event. Too slow opening time will cause user loss.

Optimization suggestions:

Reduce page size and number of elements

Optimize network performance key indicators (see Part 2 for details)

3. Website opening success rate

The website opening success rate refers to the ratio of the number of successful web page openings to the total number of visits, reflecting the health of websites in different industries. Access failures will lead to user loss.

II. Overview of Key Network Performance Indicators

Whether the network performance is good or not directly affects the website's first screen time, opening time and opening success rate. The main key indicators that affect network performance are DNS time, connection time, first packet time and download speed.

1. DNS Time

Monitors the time it takes for a client to resolve a specified domain name into an IP address through the Domain Name Resolution Service (DNS).

Optimization suggestions:

Improve the quality of the website's authorized DNS network link

Reasonably set the DNS cache expiration time

Increase monitoring methods to promptly detect DNS resolution anomalies and quickly resolve issues

2. Connection establishment time

The time it takes for the browser to establish a TCP/IP connection with the web server, and the time it takes for the first two handshakes of the three-way TCP/IP connection.

Optimization suggestions:

Enhance server hardware processing performance

Optimize the links of the existing network (such as purchasing CDN services, adding more line rooms, etc.)

3. First package time

The time it takes for the browser to finish sending the HTTP request and receive the first data packet returned by the WEB server.

Optimization suggestions:

Enhance server-side processing performance

Optimizing network links

4. Download speed

Monitors the number of bytes downloaded by the client from network resources every second, in KB/s.

Optimization suggestions:

Increase server-side bandwidth redundancy

Optimize links

3. Impact of CDN on Network Performance

CDN has a significant impact on network access performance, and the acceleration effects of different types of CDN services vary significantly.

1. CDN usage across the industry

[[134512]]

2. Comparison of acceleration effects of different types of CDNs

The impact of CDN on network performance is mainly reflected in three indicators: download speed, connection time, and first screen time:

Download speed: unused (118KB/s), commercial (194KB/s), self-built (213KB/s)

Connection time: unused (85ms), commercial (36ms), self-built (32ms)

First screen time: unused (3.7s), commercial (2.1s), self-built (1.8s)

4. Comparison of Operator Network Performance

1. Operator network response time

The delay when netizens access network applications through various operators’ networks.

China Unicom network response performance in North China and Northeast China***

Central China Henan: China Unicom's network response performance is good

East China: China Telecom's network response performance is good

South China: China Telecom's network response performance is good

Northwest China: China Mobile's network response performance is good

2. Operators’ network response status at different times

Operator response status in different time periods:

From off-peak to peak hours, the operator's network response time is increasing

3. Operators support access to websites in different industries

Due to differences in website files in different industries, each operator has different support for different website file types.

<<:  How to become a legendary full-stack engineer

>>:  A great guide to job hopping for programmers

Recommend

Beginner's guide to UI design: iOS

This article is the first chapter of UI design, m...

What? You pronounced YouTube and App wrongly.

But do you know that foreigners don’t understand ...

Ferrari experts come to Zotye Auto to build Ferrari?

Recently, several manufacturing experts from the ...

Baidu information flow promotion, telephone consultation advertising style!

Telephone consultation style 【Bidding Advertising...

The childhood snack "shredded figs" has nothing to do with figs?

I believe that for those born in the 1980s, the l...

Programmers: How to protect your eyes

[[142325]] In three months, I will be a programme...