IRISnet is named after Iris, the Greek goddess of the rainbow, who was a faithful messenger between earth and heaven. IRISnet is a self-evolving BPoS cross-chain service hub. Its mission is to become a trusted "bridge" connecting digital and physical economies, providing next-generation public chain infrastructure for complex distributed business applications. Developed in collaboration with the Tendermint team, IRISnet will support seamless integration between public chains, consortium chains, and traditional business systems, enabling data and complex computations to interoperate across heterogeneous networks and facilitating cross-chain service invocations.
Project Vision
We aim to evolve current blockchain technology so that thousands of small and medium-sized businesses (SMBs), and even individual freelancers, can provide and benefit from services on an open network. To achieve this, we have identified several challenges and the corresponding technological innovation opportunities:
Not all computations can or should be performed on-chain as smart contracts.
Ethereum offers a Turing-complete virtual machine for executing smart contracts, fueling hopes for decentralized applications. However, smart contracts can only handle deterministic logic, ensuring each node reaches the same state after processing the same transactions and blocks. Much of existing business logic is non-deterministic and changes over time and under different environmental parameters. Particularly now, business systems increasingly rely on algorithms for decision optimization, including natural language processing (NLP), machine learning, and operations research. We often deliberately introduce randomness into these algorithms to avoid merely local optimal states while striving for better suboptimal results.
On the other hand, some real-world business logic should run off-chain, not as repeatable computations like smart contracts. Integrating and coordinating off-chain services and resources using distributed ledgers is key to further advancing blockchain technology in more practical scenarios.
How to leverage existing blockchain resources, including public and consortium chains.
Using a single public chain to address all use cases is impractical. New blockchains launch daily, each focusing on solving one aspect of a problem, such as distributed storage, asset ownership, or market prediction. Over 1,000 cryptocurrencies are currently active on various trading platforms according to coinmarketcap.com.
When building business applications involving storage and diverse data sources, a motivation is how to reuse some existing work, such as specialized blockchains for storage (IPFS, SIA, Storj.io, etc.), data feeds (Augur, Gnosis, Oraclize, etc.), and IoT (IOTA, etc.), rather than "reinventing the wheel."
Many (near) real-time business transactions do require closer consortium/permissioned/private chains to address performance, security, and business governance requirements. Our vision for distributed business infrastructure is to enable interoperability across multiple heterogeneous chains, including public/consortium/permissioned/private chains.
Cross-chain technology naturally addresses this need, but existing solutions primarily focus on token value transfer between existing blockchains. The question remains on how to utilize resources provided by different blockchains.
Comparing existing cross-chain technologies like Cosmos and Polkadot, we find that Cosmos provides a more mature foundation for interoperability and scalability. Specifically, Cosmoss design of "many hubs and many zones," where each zone is an independent blockchain with its own governance model, offers a suitable architecture for modeling the complexity of the real world using SOC (Separation of Concern).
To best reuse the existing framework, we propose the IRIS Network, a decentralized cross-chain network consisting of a hub and numerous zones, implemented based on Cosmos/Tendermint with more refined token utilization.
Given that the IRIS Network is designed based on Cosmos/Tendermint, we will first discuss Cosmos/Tendermint, summarizing the features we inherit and our unique innovations.
Use Cases
Distributed Artificial Intelligence for Data Analysis with Privacy Protection
The service infrastructure for this use case has been prototyped by Boundary Intelligence, a tech startup in Shanghai, and applied to their consortium chain product BEAN (BlockchainEdge Analytics Network) to address the challenge of obtaining data for running analytical models. While homomorphic encryption is a key method for performing computations on encrypted data, it is practically inadequate for real-world machine learning problems due to poor performance. Therefore, BEAN provides another solution—leveraging model parallelism and SOA design patterns from traditional distributed AI research to develop distributed analysis services as an additional layer on top of the blockchain.
To protect data access, (partial) models running at the data end need to be open to clients and specified in the service definition. Since only part of the model is exposed to clients, model developers do not need to worry about their ideas being stolen; similarly, data owners never need to fear losing control over the use of their data, as it does not leave their data sources.
Potential benefits may include:
- Exchanging only a small amount of parameter data on-chain, which can help improve performance.
- A more practical approach to data usage auditing, commonly used in healthcare.
Healthcare data is highly sensitive and involves numerous security requirements, presenting challenges for using healthcare data for cross-organizational collaboration purposes (e.g., searching for consultation records across hospitals for diagnostic assistance, patient identification for new drug clinical trials, and automated health insurance claims). The MVP service layer is built on Ethermint, aiming to connect multiple hospitals, insurers, and analytics service providers to offer privacy-protected healthcare data analysis capabilities.
Smart contracts supporting service registration and invocation on-chain have been implemented. An example of off-chain data processing is the grouping analysis service for Diagnosis Related Groups (DRGs). Specifically, when a hospital user invokes the DRG service, the original medical records are processed off-chain, using client-side NLP code stubs (implemented in SQL and Python) provided by the service provider to extract structured data passed via the blockchain without transmitting highly confidential original medical records.
The BEAN scenario illustrates a more complex service use case, including implementing distributed analysis, connecting service providers and consumers, leveraging blockchain to provide an auditable transaction platform, and a trusted distributed computing foundation.
Data and Analytics E-Marketplace
Through researching several AI + blockchain projects, it seems most aim to provide data exchange markets and analytics API markets. In the proposed IRIS infrastructure, these networks can be easily constructed by publishing data as data services and packaging analytics APIs as analysis services using the IRIS Service Provider SDK.
Distributed E-commerce
Integrating the proposed IRIS infrastructure with traditional systems (such as ERP) to obtain inventory information or perform inter-chain queries on trusted data sources for traffic and weather data is similar to methods familiar to many enterprise application developers. By integrating these services to support distributed e-commerce applications, it is possible to provide a user experience comparable to centralized systems like Amazon or Alibaba.
Combining Public and Consortium Chains
For many business scenarios, adopting a hybrid architecture that combines the best characteristics of public and consortium chains can yield beneficial outcomes, particularly regarding performance, security, and economic incentives.
For example, hospitals and insurers can form a consortium chain to support high-performance health insurance transactions while accessing statistics about global services for certain diseases from other public chains. Tokens received from public chains can be returned to information providers on the consortium chain, incentivizing system participants to improve and enhance service quality. Leveraging the infrastructure provided by IRIS, large-scale spontaneous collaborations can be achieved while meeting strict performance and security requirements.
The IRIS service infrastructure can support many use cases, such as more efficient asset-backed security systems, distributed regulatory technologies (like strict evaluations, mutual aid markets, etc.). One plan for the IRIS project includes closely collaborating with application project teams to support and enable them to have the required blockchain infrastructure, allowing them to focus on delivering expected business value more efficiently.
Related Links:
https://www.qukuaiwang.com.cn/szhb/3238.html###