Technology

MMC Protocol is a blockchain and peer-to-peer network-based edge computing protocol that includes features such as smart contracts, decentralized storage, and consensus mechanisms for transaction verification. It aims to connect various servers, personal computers, and other devices to form a decentralized computing network. It can serve as the infrastructure for connecting decentralized computing power and dApps, providing users with more flexible, reliable, and efficient computing resources, enabling them to share computing resources and process data together.

The advantage of MMC Protocol is that it can integrate computing power and applications from different sources, allowing them to exchange and collaborate on a unified platform. It provides a computing power marketplace where users can access MMC Network's computing power socket to find suitable computing power or applications. This market is decentralized, and users can achieve more secure and fast payments and settlements through smart contracts and cryptocurrencies. At the same time, MMC Protocol also provides some basic infrastructure, such as data storage and identity verification, making it more convenient for users to use the platform.

MMC Protocol also provides some innovative features, such as intelligent routing and data caching, to optimize the use of computing resources and network efficiency. Its security is also highly valued, adopting some advanced encryption technologies and security mechanisms to ensure the protection of users' data and transaction information.

MMC Protocol Topology (Computing power scheduling)

State machine

Validating nodes

A state machine in the MMC network refers to the consensus requests initiated by the validating nodes, which involve reaching consensus on the entire network's state and writing it into the MMC blockchain. The state machine uses the blockchain format to represent the computing power status of the entire MMC network.

In the MMC network, a group of validators participate in verifying the network's computing power status and packing data submissions to the state machine by locking a certain amount of MMC tokens. These validators are known as validating nodes, and they obtain the qualification to become validating nodes through elections and selections. These validating nodes periodically submit blocks to the blockchain network to verify and confirm the state machine, thereby earning rewards.

Unlike traditional blockchains, the consensus requests initiated by the MMC network are not transactions but a set of datasets (including computing power proof, computing power task scheduling, and computing power transactions) that record the computing power status. In the MMC network, this dataset is called the "state."

When any node receives a computing power call, it verifies the source of the computing power call request and puts it into a queue for processing. The node then calls the specified computing power according to the computing power call request, verifies the source of the provided computing power result, and broadcasts it to the entire network. The validating nodes then verify both parties of the computing power call transaction, initiate consensus, and wait for confirmation from other validating nodes. Other validating nodes verify the transaction generated by the initiating validating node to ensure that the validating node did not forge the transaction. When 2/3 of the validating nodes confirm the transaction, it is confirmed and written into the state machine. The entire process is asynchronous, and all real-time computing power calls are processed immediately without waiting for consensus to complete, after which the state is recorded into the state machine.

Smart router nodes

Smart router nodes are an important component of the MMC network, and they have the following functions:

Connect all validator nodes and edge computing nodes: Smart router nodes connect all validator nodes and edge computing nodes in the MMC network, enabling them to communicate and interact with each other.

Synchronize and record the full-network state machine: The full-network state machine in the MMC network needs to be recorded and synchronized by all validator nodes. As a key intermediary node, Smart router nodes can synchronize all state machine information and ensure consistency throughout the network.

Route requests to edge computing nodes based on shortest path and node load capacity, and return results: Smart router nodes can route requests to the best edge computing nodes based on shortest path and node load capacity, and return the results to the requesting party. This can improve the performance and stability of the MMC network.

Real-time communication broadcasting: Smart router nodes are also used to provide real-time communication broadcasting to the entire network. They can broadcast messages to all nodes in the network, ensuring real-time communication and interaction between nodes.

Smart router nodes earn rewards by providing computing resources and network bandwidth.

Computing nodes

Computing nodes are another important component of the MMC network, which provide computational power to the entire network. Computing nodes can be personal computers, servers, smartphones, or any other computing devices running MMC's software client. They act as target nodes for intelligent routing nodes, receive computation requests from intelligent routing nodes, complete corresponding computation tasks, and return results to the intelligent routing nodes, thereby improving the computational efficiency and performance of the MMC network.

Computing nodes are rewarded in two ways: through proof of compute and through compute tasks. Proof of compute is obtained by joining the compute node and undergoing state bit verification, while compute tasks are obtained through compute trading in the compute market.

Computing pool

To serve large-scale computation tasks, a Computing pool is a way to organize multiple Computing nodes to work together through resource sharing and task scheduling, thereby improving the computational efficiency and performance of the entire pool. Typically, these Computing nodes are equipped with high-performance GPUs and have computational power that can be used to perform various computationally intensive tasks, including AI model training and AI application inference. In a Computing pool, nodes can share resources such as GPUs, memory, etc., to improve the computational efficiency and performance of the entire pool. At the same time, Computing pools also provide many AI frameworks, such as TensorFlow, PyTorch, etc., to provide users with more convenient, flexible, and cost-effective computing services. Compared to a single compute node, a Computing pool has greater computational power and higher reliability, and can meet larger-scale computational task requirements.

Consensus

MMC adopts a proof-of-stake (PoS) consensus algorithm, and all validation nodes achieve consensus through the IBFT strategy after 2/3 confirmation.

This consensus algorithm has advantages such as high efficiency, low energy consumption, and low transaction costs, which give the MMC network a great advantage in fast application interaction and low-cost transactions.

Staking

The existence of validator nodes is necessary to ensure the security and reliability of a network. Validator nodes require a certain amount of MMC tokens as collateral to ensure that they do not behave maliciously, otherwise their collateral will be forfeited. In PoS consensus algorithms, the reward for validator nodes is proportional to the amount of tokens they have staked, meaning that validator nodes with more staked tokens have greater rewards and are more likely to become consensus validators. Therefore, they are incentivized to remain honest and contribute to the entire network.

Intelligent routing nodes are responsible for routing calculation requests and scheduling computational tasks. They also require a certain amount of MMC tokens as collateral to ensure that they do not behave maliciously, otherwise their collateral will be forfeited. In PoS consensus algorithms, the stake of intelligent routing nodes is proportional to the number of tokens they hold, meaning that nodes with more tokens have greater power and are more likely to participate in and receive computational tasks.

The number of intelligent routing nodes directly affects the performance and stability of the entire MMC network. If the number of intelligent routing nodes is insufficient or unevenly distributed, it can lead to network congestion and delays. To ensure the stability and efficiency of the MMC network, it is necessary to plan and configure the number and distribution of intelligent routing nodes appropriately. Intelligent routing nodes need to stake a certain amount of MMC tokens to maintain the stability and reliability of the network. Once the nodes are selected and able to complete their tasks successfully, they can receive corresponding rewards. In PoS consensus algorithms, the reward for intelligent routing nodes is proportional to the amount of tokens they have staked. Overall, the contribution of intelligent routing nodes in the blockchain network is widely recognized and rewarded.

For Computing nodes, PoS mechanisms can ensure that the stake of the node matches their contribution to the network, thereby promoting active participation and continued contribution. The tasks completed by Computing nodes in the network and the rewards they receive are related to the amount of stake they hold. Therefore, Computing nodes can prove their identity and willingness to participate in the network by staking a certain amount of MMC. At the same time, PoS mechanisms can prevent malicious behavior by Computing nodes, such as attacking the network, double spending, etc., thus ensuring the security and stability of the network.

Computing nodes can choose not to stake and undertake some non-continuous tasks to receive rewards. This allows them to flexibly utilize their computational resources. For example, real-time communication, AIGC applications, etc., these tasks generally do not require the computational resources of nodes for a long time and can coexist well with other computing tasks. For the MMC network, these non-continuous tasks can further improve the computing efficiency and performance of the network, while providing more revenue opportunities for small Computing nodes, promoting the development of the MMC ecosystem.

Computing socket

aassA Computing socket is a technology that converts computing resources into digital assets, where each Computing socket is assigned a unique identifier that can be thought of as an NFT. End-users can purchase these NFTs to access the corresponding computing resources, and can also sell or exchange Computing sockets to access the computing resources they need more conveniently. This is a blockchain-based distributed computing platform, where smart contracts can automate the management and execution of Computing sockets' use and transactions. Computing sockets can be configured to support specific AI applications, for example, one Computing socket may be configured to support image generation tasks based on Stable Diffusion, while another may be configured to support natural language processing tasks based on PyTorch. When a computing node uses its Computing socket to execute a task, the smart contract ensures that the node receives appropriate rewards, and pays the rewards to the node after the task is completed. This configuration allows computing nodes to optimize according to their computing capabilities and the application's requirements, providing more efficient and reliable computing resources for AI applications.

Computing power unit

It should be noted that verifying computing power is a complex issue, as many factors can affect the accuracy of the measurement, such as the hardware type used, software frameworks, and network conditions. To verify a node's overall computing power, MMC adopts a method of measuring the time required for the node to complete a standard task. This standard task can be a computational problem that requires a certain level of computing power to solve, such as a complex mathematical equation or an encryption puzzle. The time required for the node to complete this task can be used as a measure of its computing power, denoted by the symbol E.

Computing marketplace

The Computing marketplace is a blockchain-based market that provides a transparent, secure, and efficient way to trade computing resources. In the Computing marketplace, computing nodes can sell their idle computing power, while users who need computing resources can purchase these computing power units to meet their specific computing needs.

The Computing marketplace is implemented based on smart contract technology, where smart contracts automate the computing power trading process, ensuring transparency, fairness, and trustworthiness of the transactions. Each computing power transaction in the market is recorded on the blockchain to ensure that the transaction records are tamper-proof and traceable.

The Computing marketplace brings many benefits to computing nodes and users. Computing nodes can use idle computing power to earn additional income while managing their computing resources more effectively. Users can more conveniently obtain the computing resources they need without spending a lot of time and effort searching for suitable computing resources.

With the continuous development of artificial intelligence and blockchain technology, the prospects for the Computing marketplace are becoming increasingly broad. It can not only provide individuals and enterprises with more computing resources but also promote the optimization and utilization of computing resources, thereby driving the development of artificial intelligence technology.

Last quoted from https://bitcoin.org/en/ （MMC is open-source; its design is public, nobody owns or controls MMC and everyone can take part.）