Application of the 9P protocol in organizing a shared speed memory buffer for intermediate computations (RAMPL)

In the context of the development of the modern economy, the scope of application of Big Data (databases, performance computing, high-speed network channels) and automated methods of work in this area (design of data systems, management and monitoring) is rapidly expanding. Artificial Intelligence systems are becoming especially popular as methods for automating all of the above points. In our country, serious problems are being solved in these areas of computer science. Despite a number of systemic problems, in Russia a distributed network of Supercomputer Centers for Collective Use is being formed, which is designed to solve the problem of rational use of supercomputer systems. There are also advances in the work to create technologies for writing and translating parallel Software. The article provides a brief overview of the software solutions used in the Plan 9 Operating System. In particular, the 9P file system network protocol is reviewed and its brief comparison with NFS (at the moment) is given. Based on this protocol, the author proposed a software architecture for the Rapid Access for Multi Processing and Learning for intermediate calculations, which is necessary in the optimization of numerical methods and machine learning within distributed computing systems. Also, the author proposes to use the protocol’s capabilities to organize monitoring of remote hosts, as well as manage processes and devices on it. The article outlines the tasks for improving the 9P network protocol. Some comments are made about the features of using the protocol in its modern form. It is concluded that there is great potential for using technologies based on it.

9P, Plan 9, NFS, 9pnet, Linux, Neural networks, Numerical analysis, SAPFOR, DVMH, RAMPL