The EMC ecosystem is divided into four layers from bottom to top: protocol layer, network layer, application layer, and asset layer. The EMC hub is the third layer architecture (application layer), providing developers with more user-friendly development tools and an application marketplace.
The design of the EMC hub integrates AI model repository and computing power onto one platform. Model developers package their AI models into Docker and upload them to the platform, providing examples and parameter descriptions to receive developer rewards. This greatly simplifies subsequent work. Users who want to use the models only need to subscribe to computing nodes and deploy the model's Docker with one click. They can run a complete AI instance with the model automatically configured with its API. This eliminates the cumbersome tasks of downloading models, purchasing cloud services, and setting up the deployment environment in traditional modes. It allows developers to focus more on AI development itself.
Through the EMC hub platform, the workload of building a commercial AI service similar to Midjourney is reduced to within a few hours, and the subscription costs are only 30% of traditional cloud services.