Architecture Overview

Data Lifecycle

This section describes how data moves through Irys from the moment it is written to the network to the point where applications can rely on it.

At the highest level, data on Irys follows a simple path:

• it is written to the network
• it becomes verifiable
• it remains available and usable over time

Each stage is enforced by the protocol and does not require application-level coordination.

• Writing data to the network
  Data enters Irys through a data transaction. When data is written, the protocol assigns responsibility for storing the data and ensuring its availability across the network. Storage commitments are explicit, and pricing is determined upfront based on the amount of data and the duration it is expected to persist. At this stage, data is accepted by the network but not yet treated as a durable, shared resource.
• Verification and availability
  Once data is written, the protocol enforces guarantees about its availability and integrity. Through consensus, cryptographic proofs, and economic incentives, the network ensures that: the data exists in the form it was written, it remains available to be retrieved, and a specific set of miners is accountable for maintaining it. Which miners are assigned to a given piece of data, and whether they are active, is public and can be queried. This verification process is handled entirely at the protocol level. Applications do not need to introduce their own mechanisms to confirm that stored data is real, intact, or accessible. When verification completes, the data becomes part of the network's long-lived state.
• Persistent access
  Verified data remains accessible to the network over time. Applications can reference stored data repeatedly without re-uploading it, duplicating it, or re-validating it. Applications can retrieve and reuse stored data across different workflows while retaining the same guarantees. Because storage and execution are separated, increases in stored data volume do not degrade execution performance.
• Planned interaction with execution
  Programmable Data would let contracts use stored bytes as inputs during execution, applying logic to shared data within the same network. See Programmable Data status.

The sections that follow describe how execution works with stored data in more detail, and how programmable data is implemented at the protocol level.