The release of the AO public testnet marks a significant step forward for the decentralized storage project Arweave. Two weeks after announcing the launch of the super-parallel computer AO, it was officially unveiled on the early morning of February 28, 2024, Beijing time.
The AO computer is an actor-oriented machine designed for participants, with its core data protocol nodes running on the Arweave network. It is a singular, unified computing environment hosted across a set of heterogeneous nodes in a distributed network. AO aims to provide an environment where any number of parallel processes can reside and be coordinated through an open messaging layer. Unlike existing decentralized computing systems, AO can support computational operations without protocol-enforced restrictions on size and form, while maintaining the verifiability of the network itself, achieving minimal trust. The core objective of AO is to enable computation services without trust or coordination, without any practical scale limitations, providing a previously impossible design space for applications. Its scalability allows developers to generate their own command-line processes within the network and begin issuing commands. From the perspective of end-users or developers, AO is essentially a shared computer where any number of processes can run. These processes are not hosted on any specific server and are not controlled by any individual or group. Instead, once initiated, these processes can be delegated cryptographically and serve permanently in a provably neutral manner.
Core Features of AO
Compared to existing decentralized and distributed computing systems, the AO protocol possesses the following characteristics:
Parallel execution of any number of processes ("futures"): In AO, applications are built from any number of communicating processes. Inspired by the original Actor model (Carl Hewitt, 1973) and Erlang, AO does not allow shared memory between processes but permits them to coordinate through native message-passing standards. Each of these processes can then run at full speed with available computational resources without interfering with each other. By focusing on message passing, AO can achieve a scalability mechanism more similar to traditional Web2/distributed system environments than traditional smart futures platforms.
Infinite resource utilization within processes: Built upon the delayed-evaluation architecture of SmartWeave and the original version of LazyLedger (later renamed Celestia), nodes in the AO network do not need to perform any computation to reach consensus transitions in program states. The state is implied by the process message logs hosted by Arweave, known as "holograms." The computational cost is then delegated to users who can compute their own state or request it to be computed by nodes of their choice.
Access to native infinite disk on Arweave: AO processes can seamlessly load any size of data directly into memory, execute it, and write it back to the network. This setup eliminates typical resource constraints and achieves fully parallel execution, greatly expanding the possibilities for application development beyond the limitations of traditional smart futures platforms. Thus, it opens doors for complex applications requiring extensive data processing and computational resources, such as machine learning tasks and high-computational autonomous agents.
Automatic activation of futures: In traditional smart futures environments (such as Ethereum, Solana, Polygon, etc.), futures are "awakened" to perform computations based on user transaction requests. This creates an environment where programs are not "real-time" unless interacted with by users, thus narrowing the scope of applications that can be built on them. AO eliminates this limitation by allowing futures to schedule "cron" interactions, which automatically awaken them and perform computations at set intervals. Any user, or indeed the process itself, can pay nodes to "subscribe" to processes to trigger computations at appropriate frequencies.
Support for extensible modular architecture: The core architecture of AO is an open data protocol that anyone can build implementations upon. Everything—from sorters, and message relay units, to the system's virtual machine—can be freely exchanged and extended. This flexibility will allow existing smart futures systems within the Arweave ecosystem (such as Warp, Ever, Mem, etc.) to be plugged into the unified AO network and be able to send and receive messages from the unified network. This will also allow all these smart futures systems to share some of the same infrastructure and tools, thereby providing a more cohesive computing experience on Arweave.
Basic Architecture of AO
Processes: The computing units of the network. Processes are represented at a consensus level by interaction message logs stored on Arweave and initialization data items. Processes define their required computational environment (their virtual machine, scheduler, memory requirements, and necessary extensions) upon initialization. While processes are represented in this manner at a consensus level, they also imply that their states can be computed and chosen for execution by computational units that meet the requirements. In addition to receiving messages from user wallets, processes also forward messages from other processes via message unit relays. Process developers are free to determine the trustworthiness of these messages.
Messages: Each interaction with a process is represented by a message. The core of a message is a data item that complies with the ANS-104 standard. Users and processes (through their outbox and message units) can send messages to other processes on the network via scheduler units. The semantics of AO messages lie between UDP and TCP packets: they guarantee delivery only once, but if a message is never forwarded by a message unit or actually processed by a recipient, its delivery will not occur.
Scheduler Units (SUs): Responsible for assigning slot numbers to messages sent to processes and ensuring data uploads to Arweave. Scheduler units assign atomic incrementing slot numbers singly to messages sent to processes. Once assigned, schedulers need to ensure data uploads to Arweave, making it permanently accessible to others. Processes can freely choose their preferred sorter, which can be achieved in multiple ways: decentralized, centralized, or even user-hosted.
Compute Units (CUs): Compute units are nodes that users and message units can use to calculate the state of processes in the AO. Although SUs are obliged to order messages received for processes, CUs are not required to compute the state of processes. This creates a peer-to-peer computing market where CUs provide services to resolve process states and compete with each other—balancing prices, process computational requirements, and other parameters. Once the state calculation is completed, CUs return a signed proof of the specific message resolution (logs, outboxes, and requests generated for other processes) to the caller. CUs can also generate and publish signed state proofs that other nodes can load—opting to pay UDL-specified fees.
Messenger Units (MUs): Nodes that crank-based processes in the AO network to pass messages, deliver them to compute units, and coordinate to compute output results. Essentially, when MUs send messages in the system, they send them to the appropriate SU for processing, then coordinate with CUs to calculate output interactions, and then recursively repeat the process for any generated outbox messages. This process continues until no more messages need processing.
The release of AO will bring new possibilities for future computing environments. AO combines the advantages of smart futures applications and traditional computing environments, bringing new development opportunities to the entire industry. Under the framework of AO, various smart futures can be automatically executed, achieving decentralized organizational management and decision-making, greatly improving efficiency and transparency. At the same time, AO can also ensure the security and immutability of data through blockchain technology, providing enterprises with a more reliable operating environment.
For more analysis, please follow Aibit's media account for real-time updates! This article is for reference only, does not represent any position, and is not intended as investment advice. Investment is risky, caution should be exercised.
Facebook: https://www.facebook.com/aibitcom
X: https://twitter.com/aibitcom
Telegram (CN): https://t.me/aibitcom_cn
Telegram (EN): https://t.me/aibitcom
Telegram (Announcements): https://t.me/aibitcom_announcements
Discord: https://discord.com/invite/aibitcom
Medium: https://medium.com/@aibitcom
Youtube: https://www.youtube.com/@aibitcom