How Gleasy Tackles Distributed Cloud Challenges with Open‑Source Middleware

Gleasy, a cloud technology service provider, addresses two key challenges: application integration and technical framework simplification, aiming to reduce development effort on distributed, high‑concurrency, massive‑storage problems.

Application Integration

Gleasy uses an online inter‑process communication technology to integrate independent applications (e.g., Meitu XiuXiu, OneDisk) within the platform, handling launching, resource loading, management, monitoring, and invocation.

Application Process: An application runs as a URL or a JavaScript class/DOM. After installation, a registration entry is created; the kernel’s ProcessManager reads the entry, loads resources, and executes the MAIN function to create a process instance.

Inter‑process Communication: Applications declare services via the ProcessManager API. Callers invoke these services, prompting the manager to start target applications if needed and deliver events through message queues, enabling seamless communication.

Through this mechanism, Gleasy offers services such as single sign‑on, organization selection, file open/save, email, workflow, calendar, and chat integration.

Technical Framework

To cope with massive storage and high concurrency, Gleasy built a suite of distributed middleware based on open‑source projects and custom development.

Distributed Message Queue (CloudMQ): Inspired by Kafka and Meta, it uses Java, Redis for storage, and Zookeeper for coordination, providing lightweight high‑throughput messaging.

Distributed File System (CloudFS): Combines concepts from HDFS and FastDFS, adding auto‑deduplication, rapid replication, and resumable uploads, implemented with Java NIO, Redis indexing, and CloudMQ for real‑time sync.

Distributed Instant Messaging (CloudIM): Extends Openfire with Redis storage and Zookeeper coordination, exposing HTTP APIs for backend integration.

Distributed Search Platform (CloudIndex): Built on Lucene/Solr, it replaces Solr’s replication with CloudMQ‑driven synchronization to improve real‑time indexing and query performance under large data loads.

Distributed Job Scheduler (CloudJob): Uses Redis hashsets and Zookeeper coordination to replace Quartz in multi‑node environments, delivering reliable timed tasks such as calendar and birthday reminders.

Gleasy’s four‑year journey highlights the importance of selecting technologies that fit specific needs rather than chasing the “best” solution.