About IT Square!

After years of managing WordPress websites, we encountered numerous challenges with security and plugin management. While WordPress offers a complete content management system with a low learning curve, PHP's resource consumption and maintenance complexity became increasingly problematic.

WordPress uses the LAMP architecture (Linux, Apache, MySQL or MariaDB, and PHP), which consumes significant server resources. As data volume and traffic increased, server upgrades became necessary. However, the reliance on MySQL for content storage revealed its limitations—complex queries significantly slowed system response times as data grew. Additionally, upgrading and maintaining MySQL and plugins became time-consuming.

Regarding WordPress security vulnerabilities, while updates and patches could address known issues, they couldn't protect against zero-day attacks. This necessitated additional CDN and WAF protection mechanisms, creating a complex, layered architecture that was difficult to extend and maintain.

We began envisioning a solution that combined serverless architecture with AI capabilities. Initially, we considered static websites like Hugo for small content sites, which offered excellent speed but couldn't meet the dynamic requirements of larger websites.

Our new website design needed to be SEO-friendly, capable of handling AI queries, and support automated management. Therefore, we transitioned from WordPress to AWS serverless architecture, using Next.js as the underlying framework and integrating dynamic content capabilities.

We developed an AI team responsible for automated content creation and translation, capable of generating articles and incorporating relevant local technology development information and annotations, along with additional supporting content. This content is stored in Markdown format through S3, creating a headless content storage system. When AI generates content, it automatically updates to S3, and Lambda monitors and triggers website reconstruction. Some content is accessed through metadata from S3.

Through AI-powered website reconstruction, we first adopted a serverless architecture, eliminating the need to manage virtual machines. Second, we implemented complete CI/CD deployment, with source code stored in GitHub and managed through staging and production branches for development versions and the official IT SQUARE website respectively.

The Next.js framework provides numerous practical features, including content recommendations and popular content display. We integrated market information and use AI to generate reader-interesting content. To avoid unnecessary distractions, these generated contents currently require login access.

User content recommendations are based on different user behaviors, filtered through recommendation algorithms to identify relevant content and enhance reader experience. We also select more user-read articles based on user habits.