Welcome to the 2015 China Mobile Application Performance Management White Paper, Mobile Application Performance Management

Welcome to the 2015 China Mobile Application Performance Management White Paper, Mobile Application Performance Management

Click the link to download the original report: http://bbs.tingyun.com/forum.php? Mod = viewthread & tid = 136.

The 2015 s were a milestone in the development of the mobile application ecosystem. From a technical point of view, the increasing screen resolution, 64-bit processor, and HTML5 technology supporting all platforms are gradually mature, and the hardware performance is improved, the emergence of new technologies is an important factor affecting the development of mobile applications. In terms of types, the subscription profit model for online videos, online music, and dating applications has been greatly successful. downloads and usage of games, carpools, and mobile commerce applications have also continued to grow.

However, rapid development does not mean the improvement of Application quality and user experience. According to statistics, more than 74% of users choose to silence, endure, or leave in the face of application performance problems, and 10 seconds after mobile applications suffer from performance problems, nearly 5% of real users will abandon the application. In addition, compared with the loss of users, the performance of mobile applications may cause more losses to users, such as key business interruptions and reduced income caused by application crashes or errors; for example, slow end user experience, decreased user retention rate, and slow page element Loading Caused by slow application interaction, layout disorder caused by choppy or incomplete la S.

As a developer, to make users love your products in tens of thousands of applications, in addition to meeting users' needs, it is also necessary to ensure the ultimate performance and perfect user experience of mobile applications in the process of rapid iteration. In the 2015 China Mobile Application Performance Management White Paper, listen to the cloud iDaaS data center to check the performance of mobile applications on iOS and Android platforms in 2015, the performance network quality of each operator, and the average performance indicators of various industries, IT helps developers better understand the real situation of the mobile industry, effectively improve user experience, end user loss, and further improve sustainable R & D iteration and reduce it o & M costs.

2015 China Mobile Application Performance Management White Paper

 

Crash

 

 

First, let's review the crash of mobile apps in 2015.

Overall, the crash rate of iOS apps is much higher than that of Android apps, basically seven times the average crash rate of Android apps. From the data, the crash rate is particularly prominent in the 3-6 months and 8-9 months of 2015, or related to the release of the new version.

According to analysis, the cause of the low overall crash rate of the Android system is:

• The stability of Android4.X is significantly higher than that of previous versions;
• In terms of the update policy, iOS updates are pushed for a long period of time, and Android will push updates in real time;
• Because of the particularity of the language/system architecture, the OC needs to directly face the underlying API, and the error probability is high. However, due to the OS version, hardware differences have a greater impact than Java;
• IOS systems are more limited, such as memory, background, and API restrictions.

IOS & AndroidTwo platforms: Top 10 crash rates for operating versions

From the data point of view, Android 2.3.x has the worst performance in all Android versions, with an average crash rate of 1.39%. IOS7.x. x is the most popular version in all versions of iOS, with an average crash rate of 1.895. This indirectly causes the overall crash rate of iOS to be much higher than that of Android.

 

Crash status of popular models

 

Interactive Performance

 

In addition to the crash, interaction is also extremely important for apps. This directly reflects the time-consuming experience of interactions between user and mobile APP interface elements and content, it is displayed by the first packet time and HTTP response time.

 

HTTPResponse time

As you can see, more than 81.17% of the HTTP request packets are smaller than 50 KB, and 10.17% of the HTTP request packets are in the [50,100] KB range. The response time is calculated based on these intervals. As you can see, as the number of data packets increases, the response time will inevitably get longer. Based on the response time of each packet interval, you can give corresponding APP developers a reference to the corresponding average response time.

Performance Comparison of mobile network operators

By comparing and analyzing the data of mobile operators, the performance of mobile applications is obviously 4G better than 3G, and 3G better than 2G. In terms of time indicators of the first package, the level of the three carriers in the 2G and 4G networks is close, but China Unicom and China Telecom are far better than China Mobile in the 3G network.

Performance Comparison of major Wi-Fi network operators

• In terms of WiFi, founder broadband, gehua cable, and radio/TV broadband have the best performance.
• The performance of the three major mobile operators in this indicator is close, but China Unicom and China Telecom are still superior to that of China Mobile.
• The longest time for China Telecom's first packet is already lower than the average 4G first packet time of mobile network operators by 701 ms.

 

Show the first packet time in each region

• From the perspective of the region, the first packet in Xinjiang, Tibet, and Gansu took the longest time.
• There is a distinction between the West, the East and the West.
• The entire east region is under ms, and the entire west region is above Ms.

 

Error

 

APP response failures are caused by multiple reasons, including network errors and HTTP errors.

Network Error Rate

 

Network Error cause distribution

By comparing the network errors of Android and iOS, we can clearly see that there are significant differences between the error types of the two systems, proving that network errors are also related to the system. In terms of connection timeout, client protocol errors, illegal response content, and SSL Certificate Errors, Android is significantly higher than iOS. In the case of unknown host errors, iOS is significantly higher than Android.

 

Network Error operator comparison

Compared with network access methods, the three major mobile network operators are 4G, which is significantly better than 2G and 3G. However, there are many obvious problems with China Mobile's 3G network, which is higher than 2G network errors.

Comparison of the three major network operators with the same standard found that:

• In 2g, China Telecom is better than China Unicom and China Mobile.
• In 3G, China Unicom is better than China Telecom and China Mobile.
• In terms of 4G, China Telecom is better than China Unicom and China Mobile.

 

Regional distribution of network error rates

• According to the regional distribution, the overall network error rate is about 1.20%.
• In northwest China, Xinjiang has the highest network error rate, reaching 1.64%, and Tibet has 1.42%.
• In the central and eastern regions, Anhui experienced the worst performance, reaching 1.27%.

 

HTTPError cause distribution-Android

For Android systems, the cause of HTTP errors is mainly distributed in 404 (the server cannot find the requested page), 503 (this service is not provided), and 414 (the requested URI is too long, the server cannot process), 401 (unauthorized), and 403 (the request is rejected by the server). The above reasons account for 58.34% of the total.

 

HTTPError cause distribution-iOS

For iOS systems, the cause of HTTP errors is mainly distributed in 502 (error gateway), 404 (the server cannot find the requested webpage), and 403 (the server rejects the request), 500 (the server encounters an error and cannot complete the request), 401 (authentication is required for unauthorized requests), and 67.69% of the above reasons.

 

Two platforms, three major performance indicators, and fourteen industry data presentations

• In terms of crash rate, apart from the great differences between social networking, real estate, hotel accommodation, living services, and video apps iOS and Android, the overall performance of the two products is close to that of the two.
• Android system outstanding performance top five industries: Application download platform, ordering, hotel accommodation, Life service, music
• Top five iOS systems: Application download platforms, reading, games, ordering, and news media

 

• In terms of network error rate, in addition to ordering meals, hotel accommodation, application download platforms, and mobile office data, Android systems in other industries are generally higher than iOS systems.
• Top five industries with outstanding Android system performance: Food ordering, aviation, hotel accommodation, real estate, and lifestyle services
• Top five iOS systems: Reading, aviation, real estate, Life service, and video