DW is a star enterprise in social promotion, which can often get the opportunity of cooperation for promotion with popular influencers. One of the biggest cooperative exposures brought more than 400 times the daily traffic surge to DW’s official website. This is unthinkable and very difficult to handle in companies using traditional rental IDC. However, relying on Amazon Web Services's strong elastic capacity expansion, DW can complete the elastic capacity adjustment of its official website within 1 working day to support every peak traffic visit.

Amazon Web Services's globally unified management structure, uniform API and bilingual support enable the O&M teams in China and Sweden to make use of the 7-hour time difference to operate and maintain each other's cloud platforms, thus greatly reducing labor costs and minimizing night shift and holiday overtime. With Amazon Web Services Cloud and through global collaboration to achieve work-life balance, the DW O&M team is very proud. During every e-commerce promotion season in China, what we are familiar with is the scene of crew working all night. At DW, it is possible to schedule only a small number of O&M staff on duty at night, while the O&M team in another time zone can do the work instead. In case of any system abnormality, go directly to the account number in the relevant area for the first round of inspection, handling and emergency repair operation. If necessary, local personnel will be requested for the second round of troubleshooting. Such a replacement mechanism can enable the O&M personnel to cope with 24-hour non-stop online business with normal working and rest time. In addition, during the service implementation process of DW in China, the O&M team in Shenzhen used Amazon Web Services Tokyo and Singapore regions to support DW business in China, reducing the service implementation time from 1-2 months to 1-2 days.

With only 14 persons in the Asia Pacific O&M support team, Guan Yuhui is managing the IT applications of DW in four major regions including Beijing, Ningxia, Singapore and Japan, involved in the task from architecture, deployment, security to O& M, supporting the business development of the two major regions of DW China and Asia-Pacific with a streamlined but professional team.

Another measure for DW to improve O&M level is to use Amazon ECR and Amazon ECS to transform the traditional Amazon EC2 architecture into a containerized architecture. After containerization is realized, DW uses Amazon EC2 Auto Scaling groups to detect the health degree of each host and each container in a real-time manner. If there is a problem with its health degree or if a component fails, DW will restart the corresponding business process by automated means to achieve smooth operation with no need to get offline.

Through a series of measures, DW China’s IT team took over the official website in China on July 5, 2018, achieved 99.98% availability, increased official website availability by 0.05%, and realized 100% availability records during peak sales hours, almost helping the business department to achieve its business objectives to the greatest extent.

DW heavily uses Lambda and Amazon API Gateway serverless service technology for deployment in interface layer and timing application. Through serverless architecture, the flexibility of interfaces between DW application systems is greatly improved. For example, DW has many order systems such as official websites, Tmall, Jingdong, WeChat applets, etc., which all need to transmit data to ERP system. In the past practice, each order system has to write a program to connect the ERP system. Using Lambda serverless technology, DW can realize the interface between the ERP system and any order system by only one interface application between the order system and the ERP system.

The use of serverless architecture to implement interface application can bring two main benefits: First, the elastic capacity expansion can be realized automatically. Even in the case of a surge in order volume, DW does not have to worry about overpressure on the order system, because Lambda will automatically achieve implement flexible capacity expansion. In case of using the traditional architecture, a stress test must be carried out on the order system before the marketing activities. If the server is found slow, its capacity must be expanded in advance to avoid bottlenecks in the order flow under the circumstances of the "Double 11" promotion. With a serverless architecture, Amazon Web Services can easily handle even 100,000 orders per minute.

Burst traffic is quite common for a fashion brand like DW. Once a fashion blogger created a post on Weibo about DW products, and the traffic of DW’s official website increased 3 times that day. The adoption of serverless architecture makes DW free of stress.

Second, the interface is reusable. The same interface application can connect different ERP and different order systems as long as one interface code is modified. For example, it is possible to interface the ERP of DW in South Korea, Europe and Australia. If this interface application is used simultaneously, it is possible to easily connect official website, Tmall, Taobao and Jingdong.

In fact, DW has 8 E-commerce platforms in China. If the previous model was adopted, it would be developed 8 times to realize the connection of 8 e-commerce platforms with ERP, respectively. Now, the data can be read only by modifying the interface so that it corresponds to the standard development documents of the 8 platforms. And the interface application was copied from the Swedish team. Therefore, the DW’s team in China can achieve the effectiveness of the previous 8 jobs with only one fifth of its energy.