Castrol China

I. Business Introduction

Castrol is a globally recognized expert in lubricants. Castrol’s automotive and motorcycle lubricants are trusted by manufacturers, drivers, and riders for their high performance and superior quality. Castrol also manufactures lubricants for various other applications such as maritime, land, and air transportation. BP acquired Castrol in 2002 and has maintained its fine traditions of professionalism, innovation, and collaboration. Today, Castrol products are sold in more than 150 countries/regions, making it the preferred lubricant partner for enterprises including Volkswagen, Audi, BMW, and Komatsu.

II. Challenges Faced

Since entering China in 1989, Castrol has achieved rapid business development. To cope with this high-speed growth, Castrol built an information management system to meet business needs. Over the past 15 years, the information system has played a key supporting role in the business. However, with the development of the online economy, the original traditional IT architecture can no longer meet the growing business demands, especially those from the C-end.

In 2020, Castrol faced business challenges: on one hand, online orders surged due to the global pandemic; on the other hand, the IDC’s operation and maintenance SLA could not be guaranteed during the pandemic. As a leader in cloud technology, Shenzhen Agilewing Cloud Computing Technology Co., Ltd. (hereinafter referred to as “Agilewing”) provided cloud resource services to Castrol. Over more than a year, Agilewing helped Castrol migrate IT resources from the IDC to the cloud service provider in an orderly and smooth manner. During the migration, Agilewing assisted the client in reconstructing some key applications: transforming the original application services into microservices through EKS (Elastic Kubernetes Service); and improving database availability through RDS (Relational Database Service). The migration was officially completed in June 2021.

III. Migration Solutions

Overall Architecture Diagram

Operation and Maintenance Monitoring Architecture Diagram

IV. MSP Operation and Maintenance Hosting Solutions

In 2022, Castrol adopted Agilewing’s MSP (Managed Service Provider) services, completely freeing its IT personnel from underlying IT operation and maintenance work. Agilewing’s MSP helped the client reduce operational costs and improve response efficiency, winning unanimous praise from the client. Currently, Agilewing is stably providing stable, high-quality MSP services to the client, safeguarding its business operations.

Guangzhou Jinxing Computer Technology Co., Ltd.

I. Business Introduction

Guangzhou Jinxing Computer Technology Co., Ltd. is an ERP system provider that offers services to users via the SaaS model.

Its self-built servers lack flexibility and agility in resource allocation, making it difficult to adapt to rapid business changes and meet the evolving needs of the operation team. The month-end reporting peaks impose higher requirements on the computing power and scalability of the machines. The database is growing rapidly and requires high availability. Additional files are stored in shared NAS, posing a risk of single point of failure. Business growth has also brought tremendous pressure to the traditional operation and maintenance team, necessitating a more reasonable operation and maintenance solution.

II. Challenges Faced

Before migrating to the cloud service provider, the company established a data center in Guangzhou. With continuous business growth, the team encountered numerous challenges, and the original IT architecture could no longer meet the needs of rapid business development:

• The self-built servers lack flexibility and agility in resource allocation, failing to adapt to rapid business changes and meet the increasing business requirements of the operation team.
• The month-end reporting peaks place higher demands on machine computing power and scalability.
• The database is expanding rapidly, requiring high availability.
• Additional files are stored in shared NAS, which has a single point of failure risk.
• Business growth has exerted significant pressure on the traditional operation and maintenance team, calling for a more rational operation and maintenance solution.

III. Migration Solutions

The solution implemented by Agilewing is a containerized ERP integrated service system built on the cloud service provider. EKS (Elastic Kubernetes Service) delivers outstanding performance, scalability, and reliability. RDS (Relational Database Service) enables faster construction, deployment, and high availability. S3 provides unlimited storage and retrieval space. Together, these products offer the necessary scalability, performance, and availability required for Jinxing’s large and growing services while resolving the issue of high costs.

IV. MSP Operation and Maintenance Hosting Solutions

• Provide technical support from Agilewing’s TAM (Technical Account Manager) team and architect team, with a maximum response time of 15 minutes and 7×24-hour online support service.
• Daily Operation and Maintenance Services: Offer consultation on the usage, troubleshooting, configuration, and usage limits of cloud services; provide management of the deployment, configuration, and changes of application service environments.
• Monitoring and Alert Services: Monitor the availability and performance of cloud services, help users set reasonable alert strategies and thresholds based on their actual business conditions, and send alert notifications via email.
• Data Backup and Recovery: Provide backup and recovery solutions for business systems.
• Security Management Services: Offer optimization suggestions and implementation support for cloud services and systems based on security best practices.

V. Value of Cloud Migration

Agilewing has received satisfactory ratings from Jinxing for migration and operation and maintenance hosting. With more flexible and powerful scalability support, Jinxing has achieved a perfect balance between cost and efficiency.

• ELB (Elastic Load Balancing) and Autoscaling have helped address peak performance issues.
• By adopting RDS multi-availability zone deployment, Jinxing has improved database availability, reduced downtime risks, and lowered maintenance burdens.
• Replacing shared storage with OSS has resolved the single point of failure problem for Jinxing.
• Provide managed operation and maintenance of cloud services on behalf of users.
• Offer professional-level SLA technical support.
• Conduct regular on-site communications with customers and provide suggestions and configuration guidance for their existing environments.

Guangzhou Hongshi Electronic Technology Co., Ltd.

I. Client Introduction

Founded in 2001, Guangzhou Hongshi Electronic Technology Co., Ltd. is a comprehensive enterprise specializing in smart home security products, integrating R&D, manufacturing, production, distribution, and export. It provides wireless camera solutions for many well-known enterprises at home and abroad. The company’s main products include panoramic cameras, PTZ cameras, low-power cameras, outdoor cameras and other intelligent surveillance cameras. Its flagship product, the V380 solution intelligent surveillance camera, has covered many countries around the world.

II. Project Objectives

With the company’s business development, the new supply chain RPA (Robotic Process Automation) business needs to migrate to the cloud to provide internal process-oriented supply chain product demand and ordering services. The RPA mainly offers process-based services for unified source management and order requirements of various component suppliers, improving the work efficiency of internal personnel and boosting performance.

The current RPA business requires cloud migration, with the following application-level components:

• RPA Commander: Dispatches worker robots to process orders.
• RPA Creator: Generates automated order processes.
• RPA Worker: Continuously processes orders in real time.

The client aims to fully leverage the advantages of cloud resources, requiring support for large bandwidth needs, elastic machine scaling, and convenient storage resources. The planned resource deployment is in Singapore.

Core Requirements

• Achieve unified resource management and optimize order processes through the RPA process platform.
• Simplify workflows and improve work efficiency via the RPA process platform.
• Utilize Oracle Cloud’s resource advantages to achieve redundant bandwidth for machines.
• Leverage Oracle Cloud’s capabilities for elastic host scaling.
• Take advantage of Oracle Cloud’s convenient storage services.

Client Goals

1. Hongshi’s staff logs into the company’s RPA platform for unified source management and querying/ordering of relevant product components.
2. Requests are sent to the RPA Commander, which dispatches Worker robots to process orders according to the processes generated by the RPA Creator.
3. The RPA Creator generates automated order processes and submits them to the Commander for handling.
4. The RPA Worker processes orders continuously in real time as instructed by the Commander and returns the final results to the Commander.

III. Architecture Design

In response to the client’s cloud migration needs, Oracle Cloud hosts are selected to bear business traffic. For elastic scaling and load balancing, Oracle’s Autoscaling and managed Load Balancer are adopted, enabling the client’s business to elastically scale machines corresponding to peak traffic during peak periods. For cloud security, Oracle’s security groups and firewalls are used for protection, and all business-related machines are deployed in Oracle’s private network, greatly enhancing the security of cloud resources.

Architecture Diagram

Architecture Description

• Autoscaling is used to scale machines up or down during business peak and off-peak periods, realizing elastic scaling.
• The Load Balancer uniformly distributes temporary machines, enabling the RPA Commander to call them centrally.
• The bandwidth advantage of Oracle Cloud hosts meets the client’s needs.
• Oracle Object Storage is used for unified shared storage of machines, facilitating access for business machines.
• Oracle’s Firewall and Security Lists services ensure secure cloud access.

Guangzhou Beihui Information Technology Co., Ltd.

I. Business Introduction

Guangzhou Beihui Information Technology Co., Ltd. is an IT company focusing on the development of enterprise sales management systems. Its core independently developed products include:

• Sales Information Management System (SIMS)
• Distributor Order System (DOS)
• Distributor Mall (DOS-Mall)

II. Challenges Faced

All business systems of Beihui Information are developed based on Microsoft technologies and run on the Linux + Mono + .NET + MySQL platform. In recent years, to respond to changes in the business model, Beihui has transformed its product sales model from traditional packaged software sales to the current SaaS (Software as a Service) model. This transformation has brought more business opportunities as well as challenges.

In the early stage of the business transformation to SaaS, Beihui Information deployed its business systems in an IDC in Guangzhou to provide SaaS services nationwide. During the service provision process, the IDC’s SLA (Service Level Agreement) failed to meet the SaaS requirements, seriously affecting user experience.

With the continuous development of the business, the demand for computing power and storage space has been increasing. Beihui Information met the business requirements by adding hardware equipment. In addition, various shopping festivals, month-ends, and quarter-ends are peak periods for user access. To cope with the demand peaks, Beihui Information had to increase hardware investment, which also continuously increased the difficulty of operation and maintenance management.

As hardware increased, commercial software licenses also increased correspondingly. The software licenses for Microsoft operating systems and databases are associated with hardware information, making it difficult for Beihui Information to quickly and legally expand business nodes and databases under the traditional licensing model.

III. Migration Solutions

In 2016, with the help of partner Agilewing, Beihui Information began to replace the traditional IDC with public cloud services. Shenzhen Agilewing Cloud Computing Technology Co., Ltd. (hereinafter referred to as “Agilewing”) followed the best practice methodology for enterprise cloud migration and tailored a phased business system migration and MSP (Managed Service Provider) operation and maintenance hosting solution for Beihui Information.

Phase 1

Reinstall and deploy the business systems on the public cloud, replacing the original SQL Server with RDS (Relational Database Service). The IDC was set as a remote disaster recovery center.

Phase 2

Transform the business systems by separating the session service to make the application servers stateless. Deploy an ALB (Application Load Balancer) in front of the application servers, and place the business servers in an Auto Scaling Group.

Phase 3

Migrate the entire system to the Guangzhou region to obtain the high availability of SQL Server RDS while reducing operating costs. Retain the Beijing region as the disaster recovery center.

IV. MSP Operation and Maintenance Hosting Solutions

• Provide technical support from Agilewing’s TAM (Technical Account Manager) team and architect team, with a maximum response time of 15 minutes and 7×24-hour online support service.
• Monitoring Services: Provide server bandwidth usage monitoring, network performance monitoring, service operation monitoring, system process monitoring, and server CPU/memory resource monitoring.
• System Security Services: Provide system security checks, ARP attack protection, system virus and Trojan scanning/removal, system security patch updates, system security hardening, and system firewall configuration.
• Backup Services: Program backup/recovery, database backup/recovery.

V. Value of Cloud Migration

By migrating business systems and adopting MSP operation and maintenance hosting, Beihui Information has achieved multiple business goals:

1. Improved business SLA: The high-quality public cloud services ensure higher computing reliability in a single region than the original IDC. Moreover, multi-region services provide an optimal infrastructure for off-site disaster recovery.
2. On-demand resource acquisition: After using public cloud services, computing power, storage, and licenses can be quickly obtained according to actual business needs, eliminating the need for large investments in fixed assets. This not only meets the company’s asset-light financial requirements but also efficiently solves the compliance issues related to commercial software licensing.
3. Streamlined operational investment: By utilizing Agilewing’s MSP services, Beihui Information has further optimized operational input and allocated more resources to core businesses.

Nezha Auto (Hozon New Energy Automobile Co., Ltd.)

I. Client Introduction

Hozon New Energy Automobile Co., Ltd. has officially launched active overseas market expansion this year, with a formulated global strategy: “Establish a presence in ASEAN, explore the EU, expand into the Middle East, and penetrate Oceania”. Currently, Nezha Auto has entered 5 overseas countries, including Thailand, Israel, Nepal, Laos, and Myanmar, and developed more than 30 overseas distributors. It will further enter Singapore, Malaysia, Vietnam, India and other countries while continuing to explore the European market. A smart ecological factory has been built in Bangkok, Thailand to carry out whole-vehicle manufacturing business.

II. Project Objectives

Based on this, Hozon New Energy Automobile needs to seek a reliable, secure, and compliant global public cloud vendor partner that can cover various overseas regions and assist Nezha Auto in realizing localized operations in target overseas markets.

To meet the needs of Hozon New Energy’s Thailand Smart Ecological Factory and subsequent continuous overseas business development, as well as accelerated overseas business access, Nezha Auto’s MOM (Manufacturing Operations Management) project involves cloud resources including cloud servers, cloud storage, cloud networks, cloud databases, cloud security, cloud SD-WAN, as well as related services and operation and maintenance support.

III. Architecture Design

In response to the client’s cloud migration needs, Oracle Cloud hosts are used to bear business traffic, with managed load balancers as the entry point. For cloud security, Oracle’s security groups, firewalls, and other protection services are adopted. All business-related servers are deployed in OCI’s private network to ensure the security of cloud resources.

Core Architecture Guarantees

• The basic architecture of WEB and APP is built on high-availability storage and backup/recovery mechanisms.
• The database adopts an active-active (two masters and two slaves) mechanism, ensuring zero data loss for the MES system and a fault switchover time of 0-10 minutes for the MES system infrastructure.
  • Server and storage availability switchover takes minutes (generally 5-10 minutes); if the switchover is caused by an Oracle RAC node failure, there is no impact on applications.
  • Database availability switchover: In case of RAC node downtime, the RAC cluster will switch seamlessly and automatically; if the failure involves shared block storage or the entire RAC cluster, switchover to ADG takes minutes (generally 5-10 minutes).

IAAS Architecture

Network Architecture

Architecture Description

• Services are deployed in different fault domains to ensure service availability.
• Public network traffic is inspected by WAF to detect and block potential threats from the public network, protecting applications from malicious attacks.
• The database adopts a RAC+ADG architecture, featuring high availability and disaster recovery capabilities to ensure business availability and data security.

IV. Value of Cloud Migration

1. Achieve interconnection of business in various regions through one of the technologies such as cloud dedicated lines, physical dedicated lines, or SD-WAN.
2. After using public cloud services, computing power, storage, and licenses can be quickly obtained according to actual business needs, eliminating the need for large investments in fixed assets.
3. Utilizing Agilewing’s MSP (Managed Service Provider) services further streamlines operational input, allowing more resources to be allocated to core businesses.

Guangzhou Zhaohe Cloud Technology Co., Ltd.

I. Client Introduction

Guangzhou Zhaohe Cloud Technology Co., Ltd. (referred to as “MooreCloud”) is a domestic technology startup developing global businesses overseas. Driven by passionate entrepreneurs, it has a sole mission: to democratize access to video games through cloud gaming. Firmly believing that the future of gaming lies in the cloud, its professional team is committed to making games accessible to everyone, anywhere.

The vision behind MooreCloud is: playing your favorite video games on any device, just like streaming music or movies.

MooreCloud is not only a cloud gaming platform but also a gateway to unlimited gaming experiences. It enables players to play any game anywhere without the need for expensive hardware. Thanks to advanced cloud gaming and streaming technologies, MooreCloud allows instant access to video games and brings the gaming experience everywhere, no matter where you live.

II. Project Objectives

With the continuous growth of the client’s user base, the original business architecture struggled to meet the demands of high concurrent access and data processing during peak hours. Game servers frequently faced the risk of downtime, leading to poor player experience. In addition, limited scalability and long deployment cycles for new features and updates affected the speed of game innovation and market competitiveness.

To meet the client’s needs for long-term stable development, the cloud resources involved in the MooreCloud project mainly include cloud servers, cloud storage, cloud networks, cloud databases, cloud security, as well as related services and operation and maintenance support.

III. Architecture Design

To address the above challenges, it was decided to migrate the core game services to Oracle Cloud Infrastructure (OCI), leveraging its high-performance, scalable, and secure cloud infrastructure.

Compute Layer

• Use OCI Compute instances, selecting high-performance virtual machine types such as VM.Standard.E4.Flex to meet the CPU and memory requirements of game servers.
• Utilize OCI Instance Pools to automatically create and manage multiple game server instances, enabling elastic scaling to cope with fluctuations in user traffic.

Network Layer

• Create a private network using OCI Virtual Cloud Network (VCN), isolating resources such as game servers and databases in a secure network environment.
• Configure security lists and security groups to control network traffic and prevent attacks and data leakage.
• Employ OCI Load Balancing service to distribute user traffic across multiple game servers, improving the availability and performance of game servers.

Storage Layer

• Use OCI Block Volumes to provide high-performance block storage for game data storage needs, such as player data and game resource files.
• Utilize OCI Object Storage to store game resource files (e.g., game models, textures, sound effects), offering a highly available and cost-effective storage solution.

Database Layer

• Adopt OCI MySQL Database System (MDS) to provide high-performance, automated data storage and processing capabilities for the game’s backend management system.

Monitoring and Management

• Use OCI Monitoring to track the operational status of game servers, such as CPU, memory, and network metrics, for timely issue detection.
• Utilize OCI Logging to collect game server logs, analyze user behavior, and identify the root causes of problems.

IV. Project Process

Early Stage

• Before migration implementation, Agile Cloud conducted in-depth communication with MooreCloud to understand its game business needs, including user scale, game types, performance requirements, and security requirements.
• Analyzed MooreCloud’s existing IT architecture (including servers, databases, networks, and security) to assess the feasibility of migration to OCI.
• Developed a detailed migration plan based on the assessment results, covering migration steps, timelines, resource requirements, cost budgets, and risk assessments.
• Conducted a cost-benefit analysis to compare the costs after migration to OCI with existing costs and evaluate the benefits brought by the migration.

Mid Stage

• Conducted system testing to ensure that game servers, databases, networks, etc., could operate normally in the OCI environment.
• Formed a joint project team with the client to carry out detailed discovery work, including asset inventory, dependency mapping, and performance baseline testing.
• Based on the assessment results, jointly designed and reviewed the detailed target architecture, migration strategy, and complete rollback plan.

Late Stage

• Trained MooreCloud’s operation and maintenance team on how to use OCI services, helping them independently manage and maintain the OCI environment.
• Assisted MooreCloud’s technical team in selecting OCI services, such as Compute instances, Block Volumes, MySQL Database System, and Load Balancing.

V. Value of Cloud Migration

• High Performance and Low Latency: OCI Compute instances provide high-performance virtual machines, and GPU instances offer stronger computing power, reducing game latency and enhancing user experience.
• Cost-Effectiveness: OCI offers a pay-as-you-go model, charging based on actual usage, reducing operation and maintenance costs by 40%.
• Scalability: OCI provides elastic scaling capabilities to dynamically adjust resources according to user needs, supporting the rapid development of game business.
• Global Coverage: OCI’s global data center layout ensures low-latency access for players worldwide.

VI. Client Achievements

• Improved Player Satisfaction: The stability and response speed of game servers have been significantly enhanced, reducing latency and downtime incidents, and player complaint rates have dropped noticeably.
• Accelerated Innovation Cycle: With automated tools and an agile development environment, the deployment time for new features and game updates has been shortened by more than 50%.
• Expanded Global Market: The global cloud infrastructure has supported the rapid expansion of games in emerging markets, attracting more overseas players.
• Enhanced Brand Image: The high quality and stability of the game have won praise from players and the media, improving the company’s market position and brand influence.

EASY TECH GROUP CO., LIMITED

I. Client Introduction

EASY TECH GROUP CO., LIMITED is a one-stop software service platform dedicated to serving Chinese enterprises going global. Centered on customer needs, the company breaks the limitations of a single software vendor by acting as an agent for multiple categories of software products. From a neutral perspective, it tailors the most suitable software portfolio solutions for enterprises.

Its core service model starts with in-depth business consulting, aiming to fully understand customers’ development goals and challenges, and then provides targeted software selection and efficient, secure implementation services. This consulting-driven and implementation-guaranteed model ensures that software services closely align with enterprises’ actual businesses, ultimately helping overseas-going enterprises reduce costs, improve efficiency, and achieve a win-win situation for both parties.

II. Project Objectives

The client’s core advertising system processes tens of billions of ad requests daily, with a data throughput reaching the terabyte level. With the rapid business development in Europe, America, Southeast Asia and other regions, the original domestic system can no longer meet the needs of overseas customers. Now, the plan is to deploy the system overseas to support overseas business operations.

To ensure the client’s sustained competitiveness in the global digital advertising market, it is necessary to build a more flexible, highly available, high-performance and cost-effective next-generation advertising technology architecture. Oracle Cloud Infrastructure (OCI) and managed services will be used to provide a stable and sound operating environment for the advertising system.

III. Architecture Design

Global Access and Scheduling Layer

• Leverage OCI’s global backbone network and DNS traffic management to intelligently route global user requests to the nearest edge nodes.
• The front end adopts an OCI Load Balancer cluster to intelligently distribute massive user requests from the Internet to the back-end application layer, which is connected to server clusters deployed in multiple Availability Domains (ADs) to ensure that a single computer room failure does not affect services.

High-Performance Computing Layer

• Core servers use OCI high-performance computing instances to ensure high network throughput and low-latency NVMe local storage performance.
• Through OCI Auto Scaling policies, the number of instances is dynamically adjusted based on real-time QPS and CPU utilization.
• All stateless services are containerized and deployed in OCI Kubernetes Engine (OKE) to improve deployment consistency and management efficiency.

Real-Time and Batch Processing Data Layer

• Client logs are collected through API Gateway and written to OCI Streaming service in real time. All raw logs are ultimately backed up to OCI Object Storage.
• Processed structured data is loaded into OCI Autonomous Data Warehouse, which supports all scenarios from real-time queries to complex analysis through its automatic indexing, auto-scaling and built-in machine learning capabilities, improving report generation efficiency.

Security and Management Layer

• At the network level, OCI Virtual Cloud Network (VCN) is used for strict isolation with clear subnet division.
• Implement least-privilege access control through security lists and network security groups.
• Deploy OCI Web Application Firewall (WAF) at the application layer to defend against common web attacks.
• All operation logs and API call logs are collected into OCI Logging service, with key alerts set up.

IV. Project Process

1. In-Depth Assessment and Design

• Form a joint project team with the client to conduct detailed discovery work, including asset inventory, dependency mapping, and performance baseline testing.
• Based on the assessment results, jointly design and review the detailed target architecture, migration strategy, and complete rollback plan.

2. OCI Environment Construction and Pilot Migration

• Build a mirror image of the production environment on OCI, including network topology and Identity and Access Management (IAM) policies.
• Prioritize migrating relatively independent data backup and offline analysis tasks to verify data integrity and performance.

3. Phased Production Migration and Switching

• Migrate applications in multiple batches: First migrate offline or non-core systems such as log processing and data analysis; build and verify new data pipelines on the target cloud, migrate and deploy supporting microservices; adopt a “dual-running” mode to replicate production traffic to the new environment for full business verification, ensuring complete matching of functions and performance.
• Start switching core business traffic: Use load balancer configuration to gray-scale import a small amount of real traffic into the new environment and monitor closely. Then gradually increase the traffic ratio until 100% switching is completed, and the entire process can be rolled back quickly.

4. Optimization, Governance and Knowledge Transfer

• Conduct intensive monitoring and performance tuning of customer resources; use OCI Cost Management tools to analyze expenditures and adjust resource specifications and reservation strategies.
• Deliver a complete set of architecture documents and operation and maintenance manuals, and carry out operation and maintenance training to ensure that the customer’s team can independently manage the new environment.

V. Value of Cloud Migration

OCI has brought comprehensive and measurable advantages to the client:

1. Obtained a high-performance, highly available global technical foundation. The reduction in bidding latency directly enhances platform competitiveness; the modernization of the data platform turns data into real-time assets, empowering business intelligence.
2. Through OCI’s flexible billing model and the automated management of Autonomous Database, the total cost of ownership (TCO) has been significantly reduced. The computing resource utilization rate has increased from less than 30% to over 65%, achieving true cost-effectiveness optimization.
3. The elasticity of the system has enhanced the client’s ability to cope with traffic peaks during large-scale marketing campaigns, ensuring revenue stability. Faster report delivery speed has improved internal operations and customer decision-making efficiency.

VI. Client Achievements

• Performance Indicators: Core business latency has been significantly reduced, and data processing speed has been accelerated.
• Availability and Elasticity: The overall platform availability has been improved to 99.95%. The system can automatically scale to 300% of its original capacity within 5 minutes to cope with sudden traffic.
• Cost-Effectiveness: After three months of optimization, the monthly total infrastructure cost has been reduced by an average of 25% compared with the previous business cost.
• Operational Efficiency: Database management workload has been reduced by approximately 50%, allowing the data team to be freed from tedious operation and maintenance work and focus on data models and business analysis.
• Customer Feedback: “This cooperation with OCI is not only a successful technical migration but also a key upgrade of our platform. The performance and cost advantages of the new architecture have given us a more solid foothold in the fierce international market and opened up new space for our product roadmap.”

Beijing Yunyu Technology Co., Ltd.

I. Client Introduction

Beijing Yunyu Technology Co., Ltd. is committed to providing enterprises with advanced digital transformation solutions. Focusing on cloud computing, big data analysis and artificial intelligence technologies, it customizes efficient and reliable systems and services for clients, helping them maintain a competitive edge in the rapidly changing market.

Its team consists of experienced engineers and industry experts, with profound technical strength and rich project experience. Adhering to a customer-centric philosophy, it provides full-cycle services from consulting and planning, technical implementation to post-maintenance, ensuring clients achieve optimal return on investment (ROI).

Believing that technological innovation can drive social progress, the company will continue to invest in R&D, continuously improve its technical capabilities, and create more value for clients.

II. Project Objectives

With the continuous expansion of the e-commerce platform’s user scale and business volume, the original system architecture faces high concurrent access pressure and data processing bottlenecks during promotion peaks. Risks of slow website response and even service interruption have increased significantly, seriously affecting user shopping experience and order conversion rate. Meanwhile, insufficient system scalability and long cycles for function iteration and version release restrict business agility and market response speed.

To support the platform’s long-term stable development, it is necessary to build a cloud-native architecture with high availability, strong security and easy scalability. The e-commerce platform is rebuilt based on Oracle Cloud Infrastructure (OCI), aiming to achieve dual improvements in system stability and elasticity. Through multi-availability zone deployment and auto-scaling, single points of failure are eliminated and business continuity is guaranteed. Additionally, OCI’s native security services are leveraged to strengthen overall protection.

III. Architecture Design

AI Capabilities

• On the OCI Data Science platform, Natural Language Processing (NLP) models are built and run for intelligent document classification, key information extraction, and automatic summary generation.
• Generative AI is integrated to support large model applications, facilitating better data analysis.
• OCI high-performance computing instances are utilized for data processing.

Architecture Access

• External users access the public load balancer via domain names, and then connect to the application cluster deployed in private subnets (application layer).
• Internal operation and maintenance must be conducted through bastion hosts in public subnets to ensure security.

High-Availability Design

The data layer uses Oracle Cloud’s native database service, MySQL Heatwave, deployed in independent private subnets. This service not only provides a high-availability architecture by default, automatically handling failover to ensure database reliability, but also integrates Online Transaction Processing (OLTP) and Online Analytical Processing (OLAP) capabilities. This unifies transaction and analysis processing, greatly enhancing the real-time performance of data-driven decision-making.原生数据库服务 MySQL Heatwave，部署于独立的专用子网。该服务不仅默认提供高可用架构，自动处理故障切换，确保数据库服务的高可用性。其核心价值更在于深度融合了在线事务处理与分析处理能力。实现了事务与分析的统一处理，极大提升了数据驱动决策的实时性。

The application layer adopts a cluster of multiple virtual machines to form a stateless or horizontally scalable architecture. Combined with the load balancer’s health checks and traffic distribution, seamless failover and auto-scaling are achieved.

IV. Project Process

1. Requirements Analysis and Strategic Planning

• Conduct in-depth analysis of the existing research system’s workflows, data assets, and technical services.
• Collaborate with various research teams to clarify core needs such as intelligent classification, knowledge retrieval, and collaborative analysis.
• Based on OCI’s service capabilities, develop a detailed future architecture blueprint and migration roadmap.
• Select partial functions as proof-of-concept (PoC) scenarios to verify the end-to-end technical path and business benefits from data access to service integration.

2. Infrastructure Setup

• Use Infrastructure as Code (IaC) tools to automatically create and configure basic cloud environments, including VCN, subnets, security policies, and Identity and Access Management (IAM).
• Deploy core network and security components such as bastion hosts and load balancers.
• Securely migrate historical research data to OCI Object Storage. Meanwhile, establish a structured data warehouse in MySQL Heatwave and complete initial data cleaning and labeling.

3. AI Capability Building

• Conduct training, evaluation, and optimization of core AI models on the OCI Data Science platform.
• Integrate OCI Generative AI services to build a research assistant prototype for auxiliary content generation and intelligent Q&A.

4. Testing, Optimization, and Continuous Operations

• Complete full integration of AI services and cloud databases.
• Conduct large-scale concurrent user access testing, data processing pipeline stress testing, and failover drills to ensure system performance, scalability, and stability meet design requirements.
• Formulate detailed release and rollback plans: first guide partial research teams to pilot use, then gradually complete the full migration of all users and data.
• Provide comprehensive technical training for operation and maintenance as well as R&D teams, and hand over automated operation and maintenance scripts, monitoring dashboards, and cost management tools.

V. Value of Cloud Migration

1. OCI’s powerful cloud-based AI platform has improved the automated processing and analysis capabilities of research data by over 50%. Researchers are freed from tedious information screening, allowing them to focus on high-value insight discovery and strategic analysis, shortening the innovation conversion cycle by approximately 40%.
2. Completely eliminated the challenges of hardware procurement cycles and capacity planning. Computing resources required for research are available on-demand in seconds and automatically released after task completion, reducing the Total Cost of Ownership (TCO) of high-performance computing by over 35%. Additionally, MySQL Heatwave’s integrated analysis capabilities eliminate the need for independent data warehouse construction and maintenance costs.
3. Based on OCI’s native high-availability architecture (such as application clusters and Heatwave’s automatic failover) and network defense security model, system availability reaches over 99.95%, providing enterprise-level security protection and compliance assurance for the company’s core knowledge assets.

VI. Client Achievements

• Intelligent Upgrade of Core Research Business: The new-generation intelligent research platform has become a daily standard tool for all the company’s research teams. The number of supported research projects and the quality of outputs have increased by over 50% year-on-year, significantly enhancing the company’s insight and influence in professional fields.
• Agile Innovation: The flexible and open cloud-based architecture enables the company to maintain continuous technical agility and quickly transform the latest technologies into client value.
• Increased Research Output: The depth and breadth of research reports have been highly recognized by clients. The quarterly output of high-quality research reports has increased by 50%, providing stronger support for solution design.
• Team Capability Upgrade: During the project, research and R&D teams have fully mastered the complete set of skills for cloud-native AI development and operation and maintenance, cultivating an interdisciplinary talent team.

Guizhou Jieba Cloud Technology Co., Ltd.

I. Client Introduction

Guizhou Jieba Cloud Technology Co., Ltd. is a full-link solution provider with years of in-depth experience in the cross-border e-commerce sector.

With “data-driven cross-border growth” as its core, the company relies on its independently developed intelligent analysis platform to provide core services for enterprises expanding into North America, Europe and ASEAN markets:

• In-depth market insights based on multi-dimensional data modeling, which enables real-time analysis of consumer behavior, competitive landscape and policy trends in target regions.
• An intelligent product selection system built with machine learning algorithms, realizing seamless connection between dynamic optimization of production plans and market demand forecasting.

Guizhou Jieba Cloud Technology Co., Ltd. is committed to helping clients build a complete closed loop from market research and product positioning to flexible supply, assisting enterprises in reducing costs, improving efficiency, and accurately completing their global layout.

II. Project Objectives

With the continuous expansion of the user scale and business volume of the e-commerce platform, the original system architecture faces high concurrent access pressure and data processing bottlenecks during promotion peaks. The risks of slow website response and even service interruption have increased significantly, which seriously affects the user shopping experience and order conversion rate. At the same time, the insufficient scalability of the system and the long cycle of function iteration and version release restrict business agility and market response speed.

To support the long-term stable development of the platform, it is necessary to build a cloud-native architecture with high availability, strong security and easy scalability. The e-commerce platform is rebuilt based on Oracle Cloud Infrastructure (OCI), aiming to achieve dual improvement in system stability and elasticity. Through multi-availability zone deployment and auto-scaling, single points of failure are eliminated and business continuity is ensured. In addition, OCI’s native security services are used to build a multi-level defense system covering network, application and data, which not only improves system reliability, but also fully protects user data and transaction security, laying a solid technical foundation for the continuous growth of the business.

III. Architecture Design

In terms of architecture, the rich basic resources of Oracle Cloud Infrastructure (OCI) are used to provide high availability, fault tolerance and scalability for the business system.

Architecture Overview

• Load Balancing serves as a unified traffic entry, intelligently distributing requests to multiple high-performance backend servers to ensure high availability and elastic scalability of applications.
• The Internet Gateway enables resources deployed in public subnets to securely access the Internet.

Security and Insight

• The Web Application Firewall (WAF) provides strong protection for Internet-facing application endpoints, effectively resisting malicious traffic and ensuring business security and policy consistency.
• Meanwhile, the Monitoring service is used for proactive observation of cloud resources, and early warnings for key indicators are set up to ensure timely notification and response when anomalies occur.

Data and Connection

• Object Storage is used to securely and reliably store massive amounts of data, which can be retrieved at high speed directly from the Internet, realizing seamless expansion of storage capacity.
• In addition, a remote peering connection is established through the Dynamic Routing Gateway (DRG) to ensure reliable and stable cross-regional data transmission; the Service Gateway enables the Virtual Cloud Network (VCN) to securely and conveniently access other platform services, connecting the network links between services.

IV. Project Process

1. Assessment and Planning

• In-depth research on business and security requirements: Collaborate closely with the client to fully understand the business scale, promotional peak traffic, core transaction processes, user data distribution and mandatory security compliance standards of its e-commerce business.
• Assessment of existing architecture and risks: Conduct a comprehensive analysis of the client’s existing IT architecture (including application servers, databases, caches, network topology and security devices), evaluate its bottlenecks, single points of failure and security vulnerabilities, and clarify the technical feasibility and optimization direction of migration to OCI.
• Design of customized solutions: Design the target architecture based on OCI services, focusing on planning multi-availability zone high-availability deployment, elastic scaling strategies, and a in-depth defense security system covering network boundaries, applications and data layers. Determine the migration method for each module. Formulate an overall security plan covering Identity and Access Management (IAM), data encryption and application protection (WAF).

2. Implementation and Deployment

• Environment setup and selection of core services: Assist the client in planning and creating networks, selecting and configuring appropriate server storage resources, and deploying and migrating to OCI’s high-availability database services and core security services.
• Application migration and integration testing: Assist the client in migration, provide technical support and cooperate in system testing.

3. Post-Delivery and Operation & Maintenance

• Switching online and support: Formulate the final business switching plan to ensure a smooth transition. Provide technical support in the early stage of going online, monitor the system status in real time, and promptly handle any problems that may arise.
• Knowledge transfer and empowerment: Provide comprehensive OCI management training for the client’s operation and maintenance team, covering resource operation, cost management, daily monitoring and incident handling.
• Continuous optimization and support: Hand over detailed architecture documents and operation and maintenance manuals. Provide continuous operation and maintenance support services, and continuously provide suggestions for performance optimization, cost optimization and security hardening based on business monitoring data.

V. Value of Cloud Migration

1. By leveraging OCI’s globally distributed regions and availability zones, multi-active/disaster recovery deployment of applications and databases is realized, eliminating single points of failure. Through native load balancing and auto-scaling, the system can handle instantaneous traffic peaks during events with ease, significantly improving business continuity.
2. With OCI’s high-performance computing and storage, as well as the autonomous capabilities of Autonomous Database, the efficiency of complex queries and report analysis is increased several times, providing powerful computing power for real-time recommendations and precision marketing. At the same time, on-demand elastic resource supply and refined cost management significantly improve resource utilization.
3. Relying on OCI’s integrated in-depth defense system, full-stack security hardening from network boundaries, hosts and applications to data is realized. In particular, native WAF, key management and data encryption services directly protect core payment and user data assets.

VI. Client Achievements

• Extreme Experience: The average page loading time has been reduced by more than 70%, with zero interruption in order processing during promotions, and significant improvements in customer satisfaction and conversion rate.
• Agile Release: The deployment cycle for new functions and activities has been shortened from weeks to days, greatly accelerating the market response speed.
• High-Availability Architecture: Realized active-active deployment in the same city, with a Recovery Time Objective (RTO) of < 30 minutes and a Recovery Point Objective (RPO) of ≈ 0 for core businesses.
• Total Cost of Ownership (TCO) Optimization: Under the premise of ensuring higher performance and security levels, the three-year total cost of ownership of infrastructure is expected to decrease by 20-35%.