Node operators process transactions and receive token incentives. These independently owned high-spec node machines around the world are connected via Internet Computer Protocol (ICP), creating a self-sovereign network not reliant on any cloud services. Advanced cryptography allows any application or service to be run on the network in a decentralized manner via smart contracts.

The Internet Computer is a decentralized cloud blockchain, where apps can be built fully onchain, using AI, or by coding. Try caffeine.ai & icp.ninja.

ICP tokenholders own and govern the protocol. They currently take (burn) all of the total transaction fees paid by users (revenue). ICP Holders can stake their ICP on the Network Nervous System (NNS), an advanced fully on-chain DAO that facilitates liquid democracy. Staking ICP allows users to submit proposals, vote on proposals, and earn rewards for their participation in governance. Internet Computer's investors include Polychain, Scalar Capital, Amino Capital, etc.

github

twitter

discord

website

Blockchains (L1)

Internet Computer

# Internet Computer Total canister state Methodology

## 1. Overview

Total canister state measures the total memory usage across all canisters (smart contracts) on the Internet Computer network, expressed in standardized units. This metric tracks the aggregate storage utilization of the Internet Computer blockchain, providing insight into network capacity usage and growth. The calculation processes daily canister memory statistics from Internet Computer network APIs and normalizes the values using ICP token decimal precision.

## 2. Data sources

- **Internet Computer network statistics**: Daily blockchain statistics obtained from external APIs that track Internet Computer network metrics
- **Canister memory usage data**: JSON-formatted daily statistics containing canister memory usage measurements in bytes
- **Scraper insertion timestamps**: Metadata tracking when each data point was collected and inserted

## 3. Step-by-step calculation

### Component A: Raw data extraction
- **Purpose**: Extract daily canister memory usage from Internet Computer network statistics
- **Logic**: Parse JSON-formatted data from Internet Computer scraper sources. Extract nested daily statistics arrays and retrieve the `canister_memory_usage_bytes` field for each day. Convert day timestamps to proper timestamp format and assign daily grain granularity. Tag all records with Internet Computer chain identifier.
- **Key fields**: `day` timestamp from JSON, `canister_memory_usage_bytes` as raw memory usage, `inserted_at` for deduplication

### Component B: Data deduplication
- **Purpose**: Handle multiple data points for the same day by selecting the most recent
- **Logic**: Apply row numbering partitioned by timestamp and ordered by insertion timestamp in descending order. Select only records with row number 1, ensuring each day has exactly one data point representing the latest available measurement for that date.
- **Key fields**: `row_number` for deduplication, `inserted_at` for recency ordering

### Final aggregation
- **Combining components**: Take the deduplicated daily canister memory usage values and apply decimal normalization
- **Currency conversion**: Apply ICP token decimal precision by dividing raw byte values by 10^8 to create standardized canister state units
- **Time aggregation**: Use forward-fill strategy to handle missing dates, ensuring continuous daily time series data
- **Final calculation**: The final metric represents total canister memory usage in normalized units, calculated as raw canister memory usage bytes divided by 10^8, with gaps filled forward from the last available value

*This methodology was automatically generated by Token Terminal's methodology analyzer.*

Total canister state stored on the network, in bytes.

Internet ComputerICP

Total canister state

Total canister state

Competitive landscape