DKG V8: Scaling Verifiable Internet for AI to Any Device, for Anyone, on Any Chain
Driving data interconnectivity, interoperability, and integrity, the Decentralized Knowledge Graph (DKG), now in its 6th iteration, delivers significant advancements that have benefited world-class organizations and shaped standards for industrial information exchange. Through partnerships with entities such as the British Standards Institution¹²³, GS1⁴⁵, European Blockchain Sandbox⁶, and various government-funded initiatives, the DKG has also played a crucial role in informing public policies.
DKG uniquely and effectively addresses the challenges of data ownership, AI hallucinations, and bias⁷ with the Decentralized Retrieval-Augmented Generation (dRAG)⁸ framework. dRAG drives a vast advancement of the RAG model initially developed by Meta⁹, by organizing external sources in a DKG while introducing incentives to grow a global, crowdsourced network of knowledge made available for AI models to use.
The DKG V8 has through a prototype demonstrated an unprecedented scale at which the Verifiable Internet for AI can drive value for anyone, on any device, and any chain. Addressing sensitive data concerns, scalability, and AI challenges concurrently has brought encouraging results that importantly shape the expected V8 release timeline.
DKG V8 — for Anyone, on Any Device, on Any Chain at Internet Scale
OriginTrail DKG has been battle-tested in real-world applications increasingly used by an ecosystem of organizations and government-supported initiatives. To date, no decentralized system has scaled in the production environment the way V6 DKG has. However, the current capacity of DKG reached its limits to support the growing usage requirements, prompting a transition to the V8, evolved to tackle the scale at which AI is consumed in any environment.
Data has been growing exponentially for decades, with AI driving further growth acceleration — according to the latest estimates, 402.74 million terabytes of data are created each day¹⁰. This trend is increasingly visible in the rising demands for additional capacity in the DKG, driven by data-intensive industry deployments in aerospace, manufacturing, railways, consumer goods, and construction driving DKG growth.
Version 8 of the DKG has therefore been designed with major scalability improvements at multiple levels, with a prototyped implementation tested in collaboration with partners from the sectors mentioned above.
3 key products of OriginTrail DKG V8
The major advancement that DKG V8 is making is in expanding the OriginTrail ecosystem’s product suite to 3 key products:
- DKG Core Node V8 — highly scalable network nodes forming network core, persisting the public replicated DKG
- DKG Edge Node V8 — user-friendly node applications tailored to edge devices (phones, laptops, etc).
- ChatDKG V8 — the launchpad for creating AI solutions using decentralized Retrieval Augmented Generation (dRAG).
DKG Edge Node — enabling the largest, internet-scale decentralized physical infrastructure network (DePIN)
The newcomer in the product suite is the DKG Edge Node — a new type of DKG node enabling the OriginTrail ecosystem to tackle the global challenges described above. As the name suggests, DKG edge nodes can operate on Internet edge devices. Devices such as personal computers, mobile phones, wearables, IoT devices, but also enterprise and government systems are where we can find huge volumes of very important data activity that DKG edge nodes will enable to enter the AI age in a safe and privacy-preserving way. The DKG edge node will enable such sensitive data to remain protected on the device, giving owners full control over how their data is shared.
Together with being protected on the device, edge-node data becomes a part of the global DKG with precise access management permissions controlled by the data owner. In this way, AI applications that the owner allows data access to will be able to use it together with the public data in the DKG via decentralized Retrieval Augmented Generation (dRAG).
Since such AI applications can equally be run locally on devices directly, this enables fully privacy-preserving AI solutions aimed at the ever-growing number of devices on the network edge that can at the same time use both public and private DKG data. The introduction of the DKG edge node enables the DKG to quickly expand to be the largest, internet-scale decentralized physical infrastructure network (DePIN).
New features of the DKG Edge Node
To unlock these powerful capabilities, DKG edge node will include new features that have previously not been available on DKG nodes but were elements of other proprietary or open-source products.
- To enable a seamless creation of knowledge, DKG nodes will inherit the proven knowledge publishing pipelines from the Network Operating System (nOS).
- The data protection techniques for private and sensitive data will be based on the NGI-funded OpenPKG project outcomes.
- The DKG Node will support all major standards such as GS1 Digital Link, EPCIS, Verifiable Credentials, and Decentralized Identities.
- To support the growing field of knowledge graph implementations globally, it will enable seamless knowledge graph integrations of major knowledge graph providers such as Ontotext, Oracle, Snowflake, Neo4j, Amazon Neptune, and others.
DKG Edge Node V8 Prototype — Oura Ring integration with demonstrated 400 Knowledge Assets published in 10 seconds
DKG V8 Timeline
The V8 DKG launch sequence consists of 4 stages, aligned with the wider OriginTrail ecosystem roadmap, with a forkless upgrade to V8.
Stage 1: V8 multi-chain infrastructure deployment
- Paranet deployment and first IPOs launched
- Base blockchain integration
- Cross-chain knowledge mining support
Stage 2: DKG core internet-scale V8 testnet launch
- Asynchronous backing
- Knowledge assets V2: Batch minting (in prototype)
- DKG Core: Random sampling (in prototype)
Stage 3: DKG edge nodes on V8 testnet
- Edge node beta launch
- Knowledge assets V2: Batch minting & native vector support
- DKG Core: Random sampling deployment
Stage 4: V8 mainnet upgrade deployment (October 2024)
²https://page.bsigroup.com/BSI-Academy-Blockchain-Solution
³https://www.bsigroup.com/globalassets/localfiles/en-th/innovation/blockchain-white-paper-th.pdf
⁴https://www.gs1.org/sites/default/files/bridgingblockchains.pdf
⁵https://www.gs1si.org/novice/novica/origintrail-resuje-izziv-ponarejenega-viskija
⁸https://origintrail.io/blog/decentralized-rag-with-origintrail-dkg-and-nvidia-build-ecosystem
