棒谷

Hundred Ways’ Successful Case of Assisting Banggu

Company Profile

Banggood Technology Group (referred to as Banggood Group) is a group company specializing in B2C cross-border e-commerce. Headquartered in Guangzhou, China, it has subsidiaries in Hong Kong, Shenzhen, Dongguan, Zhaoqing, Changsha, Hangzhou, Ningbo, Ganzhou and other places. It now has a young elite team of thousands of people. Initially, Banggood Group was deeply engaged in third-party platforms such as eBay, Amazon, AliExpress, Wish, and Newegg, and won the praise of third-party platforms and a large number of customers with high-quality products and attentive services.

industry

Cross-border e-commerce

Case background and challenges

With the rapid growth of sales performance, the continuous expansion of the enterprise, and strong support from government policies, Banggood Group began to mainly operate the comprehensive mall website (www.banggood.com) in 2009. It is committed to the sales of products on the cross – border self – operated platform, opening up overseas markets for domestic developers. Guided by customer needs, it independently develops new products and provides a personalized shopping experience. Up to now, Banggood Group has become a large cross – border e – commerce complex in the Guangzhou area and one of the leading enterprises in the industry. Banggood Group always takes “bringing high – quality Chinese products and high – quality supply chains to the world” as its corporate mission and “becoming a company that satisfies top – level global customers” as its vision, focusing on B2C cross – border export e – commerce.

With the continuous expansion of overseas business, in order to improve marketing management, enhance the service quality of sales, and promote good agent management, Banggood needs to deploy a marketing management system on the public cloud to improve the integration among goods, logistics, and agents. The IDC system deployed previously no longer meets the requirements. There are numerous imperfections in terms of network latency, availability zones, performance, reliability, and business architecture. If migrated to GCP, it can better interact with the applications of other departments running on GCP, and can also directly utilize the Direct Connect line that directly connects GCP to the local data center, providing lower latency, less jitter, and data interaction. Different optimization methods such as using CloudCDN to enhance the access experience of agents, containerizing the business to enable rapid business iteration, and deploying across multiple availability zones to improve availability can help Banggood Technology improve overall stability and access experience. This will lay a solid foundation for future SaaS services and enable rapid service to global agents.

Description of the solution architecture

For the architecture of the Banggood site on the PC side before the migration, adjustments were made to the framework on GCP, as described below:

1) Replace the four proxy layers, ha1.banggood.com, ha2.banggood.com, wapha2.banggood.com, and habg1.banggood.com, with an HTTPS Load Balancer. Correspondingly, write the URL Map Config Host rules to match different web sites, Nginx + PHP, and Solr search nodes.

For example, for an external HTTP(S) load balancer, you can use a single URL map to route requests to different destinations based on the rules configured in the URL map:

• Requests for https://example.com/video go to one backend service.
• Requests for https://example.com/audio go to another backend service.
• Requests for https://example.com/images are redirected to the Cloud Storage backend storage partition.
• Requests for any other combination of host and path are redirected to the default backend service.

2）Google Cloud HTTP(S) Load Balancing is a global proxy-based Layer 7 load balancer that enables you to run and scale services globally behind a single external IP address. External HTTP(S) Load Balancing is implemented on Google Front Ends (GFEs). These GFEs are distributed globally and work together through Google’s global network and control plane.

In the Premium tier, GFE can provide cross – regional load balancing, thereby directing traffic to the nearest backend that is in good health and has sufficient capacity, and terminating HTTP(S) traffic as close to the user as possible. For the Standard tier, the system will handle load balancing within the region.

3）The read and write of the Redis cache/Session cluster service corresponding to Nginx+PHP in the backend is connected to Cerberus. The final business access is achieved through DNS domain name resolution, thus connecting with Nginx+PHP. At the same time, the backend is connected to the GFS file sharing service for data transmission.

4）For Solr search nodes, the backend of solr*.banggood.com is connected to TiDB built on Compute Engine to perform database synchronization and transmission.

5）Considering the minimal changes to the entire migration service, for the entire web site service, the log collection and analysis of the entire architecture can be carried out through the self – built ELK cluster.

Use the product

• Migrate for Compute Engine
• Cloud Storage
• Cloud CDN
• Cloud Compute engine
• Cloud Load balancing
• Cloud SQL

Features of the plan

Compared with other cloud vendors, Google has numerous regions. In terms of the public cloud market, Compute Engine provides an ideal solution regarding throughput, stability, pricing, backup, and security.

The cross – region network input/output of Compute Engine is faster than that of AWS. Google’s global network infrastructure – the backbone of Compute Engine – is superior to AWS which uses the public Internet. As of the first quarter of 2020, Google’s Compute Engine infrastructure had 22 regions and 61 zones. In addition, Google is investing billions of dollars to strengthen its footprint in the cloud computing field.

Compared with AWS and Azure, this provides 100% uptime through transparent maintenance. Therefore, you can set up multiple cloud scenarios, including synchronous database replication between regions.

Compute Engine’s persistent disks can support up to 257 TB of storage space. This is more than 10 times the capacity that Amazon Elastic Block Storage (EBS) can hold (currently, the maximum is 16 TB). Therefore, Compute Engine is most suitable for organizations that need more scalable storage options.

Compared with AWS, Compute Engine provides a more stable service because of its ability to provide live migration of VMs between hosts. This means that organizations can operate 24 hours a day, 7 days a week, and 365 days a year without any downtime or other performance obstacles.

In the GCP ecosystem, you only pay for the computing time you consume. Compute Engine uses a per-second billing plan, unlike AWS which is based on per hour. For long-running workloads on Compute Engine, you are also entitled to attractive discounts. Although Azure also offers discounted prices, you can only get a 5% discount for annual prepayment, while Compute Engine offers a 30% discount for one-month prepayment.

GCP has a powerful built – in redundant backup system. Compute Engine uses this system for its flagship products, such as search engines and Gmail.

Google has been around for over 20 years. When you choose GCP, you get the security advantages that Google has developed over the years to protect its powerful products like search engines and Gmail.