Demystifying the black technology behind the double 11: distributed architecture + Feitian super calculation to solve the traffic flow peak

"With VR glasses shopping, it's amazing! Follow a small blue dot to travel around the world to shop!" On November 11th, at the Tmall "Double 11" main venue, the reporter who came to the scene reported Xiao Fu experienced VR When I was shopping for the BUY+ channel, I couldn’t help but sigh. Shopping has never been as cool as it is today.

Xiao Fu is not the only one who is emotional. From putting on VR glasses global shopping to shopping pages to achieve "thousands of thousands of faces", and then using drones to send express delivery to rural buyers, this year's "Double 11" is more than just a "hands-on party" carnival, Become a feast of new technology.

"From next year, Alibaba will no longer mention 'e-commerce' because e-commerce is just a ferry boat. In the future, new retail, new manufacturing, new finance, new technologies and new energy will profoundly affect China, The world and everyone." Ma Yun said at the 2016 Hangzhou Yunqi Conference. In fact, new technologies such as artificial intelligence, big data, and VR are becoming new kinetics in the economics of manufacturing, consumption, and retail.

These new technologies are gradually entering the lives of ordinary people. This year's "Double 11" is a concentrated presentation of the new technology.

This year, Tmall "Double 11" reached 120.7 billion yuan in 24 hours. In the first hour of the opening, it created a new record of 175,000 transactions per second. The record of paying a peak of 120,000 per second was also last year. 1.4 times, compared with the first "double 11" eight years ago, it has increased hundreds of times.

The data has grown substantially to test the technical strength. In the flood of hundreds of thousands of transactions per second, how to ensure the smooth operation of the system, to ensure that hundreds of thousands of users across the country can not trade at the same time? Ant Financial Technology Chief Technology Architect Hu Xi Their solution is to create a new financial Database and architecture systems.

The traditional financial payment architecture is mostly a centralized and powerful single-point architecture. It is like a central cashier with high cost and high efficiency in financial payment. It is stable and reliable, but it has limited scalability and flexibility. Sudden business activities such as “Double 11”. If you want to receive more guests, you will have to spend more money to upgrade this "Central Cashier".

"The problems encountered by Alipay are not solved by any existing technology. We must find a solution." Hu Xi said that the traditional financial payment structure is difficult to meet the rapid development of e-commerce, from 2009 Since then, Alipay has started to independently research and develop the database OceanBase. They have tried a new kind of countermeasure - the architecture of distributed "live and live". Hu Xi described this architecture as the removal of the “Central Cashier” and replaced it with countless small calculators scattered throughout the mall. Although each calculator is not as good as the “Central Cashier”, it can be booked. And the calculators scattered around have an interrelated logical relationship, so that they complement each other, back up each other, and the failure of any single calculator will not affect the entire transaction.

"It is the core cloud computing capability of this architecture, which guarantees reliable operation from the whole." Hu Xi said that this kind of attempt proved to be useful. More critically, this new database and architecture system greatly reduces the cost of use, and now Alipay's single transaction costs can be controlled at two or three cents.

The improvement of the payment system is not enough. Alibaba's technical team also independently developed the ultra-large-scale general-purpose computing operating system "Flying." This operating system can connect millions of servers all over the world into a supercomputer, providing computing power to the society in the form of online public services.

At present, "Flying" has become a universal ultra-large-scale general-purpose computing operating system, and is providing computing services to more than 2.3 million users worldwide. Developers, businesses, or organizations can gain flexible computing power through networking.

In addition, on the eve of this year's "Double 11", Alibaba's technicians also launched a full-link stress test to evaluate the zero-point peak flow and test the website's pressure capacity. Since the most massive data and traffic of “Double 11” each year appears in the tens of minutes after the zero point on November 11th, this test is also called “Zero Traffic Defence War” or “Double 11” Simulation Test. Alibaba's entire technical system must participate in this actual combat training.

After paying the “Double 11” order, Mr. Zhong, who lives in Jiangbei District of Chongqing City, is still a bit puzzled: Why is Taobao and Tmall in the mobile phone more like an exclusive shopping guide, and more and more understand their preferences? “I just thought about it at first. I bought a pair of leather shoes. After I bought it, I also recommended a few shoes for me. The styles and styles are very similar. It is also good to match with the pants I bought. If it is not a budget, I will buy them all."

Like Mr. Zhong, there are many people who are magical. This year's "Double 11" users may see different shopping pages, achieving a shopping experience of "Thousands of People". Alibaba's personalized recommendation feature allows consumers to experience the exclusive "Double 11", with its own unique main venue, unique information, unique content recommendations and scene recognition.

Behind these effects are big data and artificial intelligence technologies and product recommendations, search, news sources, community user generated content, wish lists and other information, automatically generated content. According to reports, this year's "Double 11" period produced more than 3.5 billion personalized pages, and even after every second refresh, users see new page effects. These 3.5 billion pages are designed by Alibaba's artificial intelligence "e-commerce brain." By understanding the consumer psychology, the e-commerce brain tries to become a user-friendly, quiet shopping guide.

“We are one of the few large companies that have developed their own computing platforms. On this platform, we have carried out a lot of machine learning and artificial intelligence research and development, and are deeply applied in the business of e-commerce,” said Zhang Jianfeng, CTO of Alibaba Group. .

In addition to the experience of “Thousands of People” when shopping online, another technology experience of this year’s “Double 11” comes from express delivery.

According to the monitoring data released by the State Post Bureau on the 12th, on the 11th, all postal and express delivery enterprises handled a total of 251 million express shipments, a year-on-year increase of 52%. The country's major e-commerce companies generated 350 million express logistics orders, a year-on-year increase of 59%. The data provided by the rookie network shows that the transmission speed of “Double 11” has been increasing by sending 100 million parcels. In 2013, it took two days. In 2014, it only took 24 hours, and in 2015, it accelerated to 16 Hours.

The number of express delivery has increased by nearly 60%, but the speed of delivery has increased, thanks to the wide application of new technology in express delivery. In the face of the "double 11" logistics problem, Jingdong gave its own answer with "black technology". As early as half a month before the “Double 11”, the Jingdong X Business Intelligence Logistics Lab demonstrated a fully automated warehousing scenario created by robots, artificial intelligence algorithms and data-aware networks, some of which are already in the “Double 11” application. In the Jingdong logistics system.

Xiao Jun, president of Jingdong X Business Department, said that Jingdong “unmanned warehouse” a large number of intelligent logistics robots cooperate with each other, behind which is Jingdong’s self-developed customized and systematic integrated logistics solution. It has made technological breakthroughs and innovations in control algorithms, industrial design, mechanical structure, electrical design, application scenarios, etc., and has accumulated more than 100 patent applications.

As an important logistics support for “Double 11”, Alibaba's rookie network is also based on algorithms for global logistics planning and design, from climate, business, consumers, transportation and other aspects of overall planning. With the large-scale use of electronic orders, smart distribution is also fully applied this year's "Double 11". According to the rookie network, 70% of the packages will be distributed by intelligent algorithms this year.

Landing to the courier level, the rookie parcels in the rookie network get the courier's coverage data from the courier history collection, and then integrate the real-time location of the courier to recommend the message of the user to the package. The courier, the average collection time is only 30 minutes, greatly improving the efficiency of collection.

In addition, this year's rookie network launched the "single straight hair" data products, based on big data forecasting and cloud computing, you can centrally generate orders in the same area in the merchant side in advance, directly sorting and loading by destination area, directly Delivery to the destination improves delivery efficiency.

If the technical vocabulary such as artificial intelligence, big data, cloud computing and so on is still strange and cold in the past, then this year's "Double 11", these new technologies are beginning to be at your fingertips, and every ordinary user can feel the new technology. Convenience and inclusiveness.

Take the Jingdong Express with fast delivery as an example. During this year's "Double 11" period, Jingdong launched the delivery of drones and unmanned vehicles, and focused on rural areas with low population density, complex road conditions and few orders to complete the township distribution station to the village courier. "Last mile" delivery. At present, Jingdong UAVs in Suqian area have surveyed and tested a number of routes, and carried out distribution tasks for five of them.

In another customer service field with more common consumer contact, the artificial intelligence service robot “Ali Xiaomi” with powerful voice recognition, image recognition and deep learning ability has become the main force of Tmall “Double 11” service.

According to the real-time data provided by Alibaba, as of 4 pm on the 11th, “Ali Xiaomi” took over 94% of the services, the resolution rate exceeded 94%, and the response service request exceeded 4.2 million. As early as March, "Ali Xiaomi" has been on the line, receiving 4 million people per day in half a month, and the service capacity is comparable to 33,000 servings.

What is even more exciting is that these new technologies of artificial intelligence, big data and cloud computing are gradually opening up to merchants. The combination of more advanced new technologies and specific industries and merchants has given birth to greater innovation vitality.

As a hot-selling merchant of Tmall Home Textiles, the number of packages for Jinxi Road “Double 11” has been more than 300,000 in the past two years. How to safely and quickly deliver the goods to consumers has always been the operation of Jinxi Road Home Textiles. Logistics Director Huang Hua worries about the issue. This year, Jinxilu has fully integrated the rookie network electronic face and smart selection logistics products, and became the first batch of direct sales. "Through these intelligent platforms, it is easy for us to understand the delivery schedule of each courier company and the status of each step of the package, which has played a significant role in the timely adjustment and improvement of our actual operations." Huang Hua said.

New technologies such as artificial intelligence have broken through the boundaries of online and offline, not only improving the shopping experience in online transactions, but also being applied to offline shopping operations and urban management. For example, Hangzhou West Lake Yintai City uses artificial intelligence technology to analyze how many people around the mall go to work, shopping, and their characteristics of residence, gender, spending power, shopping preferences, etc., and analyze the children around the guests are mostly around 6 years old, thus strengthening Confidence in the layout of baby goods.

Some sections of Xiaoshan District in Hangzhou have also tried a somewhat “crazy” plan to help urban transport learn to think and make decisions through big data analysis of the “Flying” system. This kind of attempt has already begun to bear fruit – in the initial test, the traffic lights were intelligently adjusted and the vehicle speed increased by up to 11%.

Couplings And Accessories

Quick coupling is a kind of connector can be connected or disconnected without tools. Quick connectors can be divided into: quick connectors for air, quick connectors for oxygen fuel gas, quick connectors for common gas and liquid, quick connectors for hydraulic pressure, quick connectors for inert gas, quick connectors for cooling oil temperature and semiconductors.

The material inclulding stainless steel, brass, aluminium or required demand.

Couplings And Accessories,Aluminium Couplings,Quick Couplings,Universal Couplings

HEBEI ZIFENG NEW ENERGY TECHNOLOGY CO.,LTD. , https://www.zifengpipeline.com