An innovative computing infrastructure model was presented at the Interdisciplinary Scientific and Educational School of Moscow State University named after M.V. Lomonosov "Brain, Cognitive Systems, Artificial Intelligence." The system created by scientists uses machine learning technologies and multi-agent systems to efficiently distribute the load between different data centers (DCs).
The solution opens new horizons for creating global distributed computing systems. The architecture, called Network Powered by Computing (NPC), unites computing resources from different parts of the world into a single ecosystem.
The architecture of the computing infrastructure proposed by Moscow State University mathematicians opens up new opportunities for the development of cloud computing, high-performance computing installations, telecommunications and Internet services.
Machine learning technologies and multi-agent systems, which underlie this infrastructure, allow predicting the execution time of tasks on various computers, which helps to optimally distribute data flows, choosing the most efficient transmission channels. The new architecture provides high performance, fault tolerance and computing security.
The system provides users with the ability to receive computing power on demand, anytime and anywhere. This creates global connectivity and resource availability, allowing the use of computing power without being tied to specific data centers.
Thus, the new model from Moscow State University can significantly improve efficiency in areas such as artificial intelligence, big data analysis, and augmented reality systems.
Read more on the topic:
Нейросети оживили весну 1945 года