Skip to main content

Virtual Machines and Execution Environments

In the HyMatrix network, virtual machines (VMs) are the isolated environments where computation tasks are executed. VMs provide standardized, reproducible, and verifiable execution contexts. They support diverse runtimes and hardware acceleration, enabling both Web3 smart contracts and high-performance Web2 computing such as AI and big data.

Multi-VM Support​

HyMatrix features an open, multi-VM architecture that supports a variety of runtime environments, including:

  • Docker containers: Ideal for complex Web2 applications, AI model inference, big data processing, and backend services, offering full OS and language environment support.
  • WASM (WebAssembly): A lightweight, fast, and cross-platform sandbox, well-suited for portable and restricted logic.
  • EVM (Ethereum Virtual Machine): Fully compatible with the existing smart contract ecosystem, allowing developers to migrate and deploy DeFi protocols, NFT projects, and other on-chain assets seamlessly.

This multi-VM design allows HyMatrix to accommodate both Web3’s deterministic contracts and Web2’s non-deterministic, data-intensive workloads.

Open and Extensible VM Ecosystem​

HyMatrix’s VM system is modular and extensible, not locked into built-in VM types:

  • Node operators can deploy and register new VM types locally.
  • As long as the VM conforms to the system’s standard interface, it can join the network.

Developers can build their own VMs or mount existing custom VMs onto running nodes, flexibly expanding the supported execution environments.

The network can continuously evolve to support new languages, frameworks, and hardware accelerators—such as Triton, JAX, ONNX, CUDA, and more.

This flexible architecture encourages a community-driven ecosystem, making HyMatrix a truly diverse computing network.

Reproducible and Verifiable Execution​

A core design principle of HyMatrix is log-driven consensus. In this model, the VM itself does not need to generate execution logs. Instead, HyMatrix nodes automatically record all required data when invoking a VM.

The VM focuses solely on business logic computation, without needing to implement logging, replay, or verification mechanisms. The HyMatrix system will:

  • Capture and log all input parameters, call sequences, and contextual metadata.
  • Timestamp and order VM invocations to ensure traceability.
  • Upload these execution logs to immutable, permanent storage like Arweave.

Any validator can download these standardized logs and replay the execution in the same VM environment to confirm that the same input and process yield the same output.

This means that even if a task runs locally on one node, the entire network can still independently verify its correctness—achieving true trustlessness. This design significantly reduces VM complexity while preserving replayability and verifiability.

Inter-VM Communication​

HyMatrix VMs are not isolated sandboxes—they can communicate across VMs and across nodes via the network.

  • Each VM has a unique address.
  • All VM addresses and their hosting nodes are registered in the global Registry.
  • When one VM needs to call another, the node routes the message over the network to the destination VM.

HyMatrix guarantees that VM-to-VM messages are 100% deliverable. Nodes require an acknowledgment receipt from the receiving VM to confirm successful delivery.

The system also guarantees that messages will not be replayed: each message is executed exactly once by the target VM. This prevents issues like double-spending and ensures execution consistency similar to blockchain transaction ordering.

This mechanism enables HyMatrix to function like a distributed microservice platform, where VMs across the globe can collaborate securely, reliably, and verifiably.

Supporting Both Deterministic and Non-Deterministic Workloads​

HyMatrix VMs are designed to support both Web3’s deterministic contract requirements and Web2’s non-deterministic tasks:

  • Deterministic VMs: Ensure the same input and logic always produce the same output. Ideal for use cases like DeFi where consensus is critical.
  • Non-deterministic VMs: May include randomness or probabilistic results (e.g., AI inference). While output may vary, the full execution process is still verifiable and auditable.

Developers can choose the appropriate VM type based on their needs—or even mix them within the same network.

In summary, the HyMatrix VM system is an open, flexible, and standardized computing layer. It breaks away from traditional blockchain execution models and allows seamless interaction between Web2 and Web3 applications within the same decentralized network.