Decentralized Hosting

Recognizing the need for a hosting infrastructure capable of supporting high streaming throughput and intensive model training, the Virtual Protocol implements a cost-effective and efficient solution.

Resource Sharing for dApps:

• GPU Resource Challenges: GPU resources, essential for heavy model training, can be costly to scale.

• Dynamic Resource Utilization: Since dApps might not require these resources continuously, a shared resource model is adopted. VIRTUAL hosts the necessary infrastructure, allowing dApp owners to share resources. This collaboration reduces costs while maintaining operational efficiency.

Distributed Model Hosting Approach:

• Empowering Token Holders with Computing Resources: Token holders with robust computing resources are invited to host and serve the Protocol's models on their equipment.

• 'Edge' Server Configuration: Each hosting unit operates as an 'edge', furnished with a dedicated network connection to the Protocol server. Usage requests are intelligently routed to the nearest and most readily available server, optimizing response times.

• Current Access Restrictions: Presently, this feature is limited to trusted parties to preserve the integrity and quality of the Protocol. Plans are underway to expand this decentralized hosting model, making it publicly accessible for wider participation in the ecosystem.

Role and Requirements of an 'Edge':

• Essential Connectivity and Security Standards: Edges must maintain robust internet connectivity and adhere to stringent security measures to serve users effectively.

• Specialized Software Package: Edges are equipped with a specialized software package installed on their servers. This package facilitates the minting of the Protocol and manages traffic distribution based on the efficiency of the edge.

• Consistency and Session Dissemination: To guarantee uniformity and current information across all edges, every session initiated is distributed through the installed software package, ensuring a seamless and consistent user experience.