Pushing Software-Defined Blockchain Components onto Edge Hosts

With the advent of blockchain technology, some management tasks of IoT networks can be moved from central systems to distributed validation authorities. Cloud-centric blockchain implementations for IoT have shown satisfactory performance. However, some features of blockchain are not necessary for IoT. For instance, a competitive consensus. This research presents the idea of customizing and encapsulating the features of blockchain into software-defined components to host them on edge devices. Thus, blockchain resources can be provisioned by edge devices (e-miners) working together closer to the things layer in a cooperative manner. This research uses Edison SoC as e-miners to test the software-defined blockchain components.

💡 Research Summary

The paper addresses the growing interest in applying blockchain technology to the Internet of Things (IoT) while highlighting the limitations of existing cloud‑centric and fog‑centric solutions. Traditional blockchain implementations, especially those that rely on competitive consensus mechanisms such as Proof‑of‑Work (PoW) or Proof‑of‑Stake (PoS), demand substantial computational power and storage, which are incompatible with the constrained resources of typical IoT devices. Moreover, the reliance on cryptocurrency incentives creates a competitive environment that does not align with the collaborative nature of many IoT deployments. To overcome these challenges, the authors propose a novel architecture that decomposes the essential blockchain functionalities—smart contracts, consensus, and in‑chain data storage—into independent software‑defined components (SD‑Components). These components are lightweight, modular, and can be instantiated on separate edge devices, termed “e‑miners,” which in the experimental setup are Intel Edison System‑on‑Chip (SoC) platforms.

Key design decisions include:

1. Consensus without competition – The system replaces PoW/PoS with Practical Byzantine Fault Tolerance (PBFT). PBFT allows a small, known set of nodes to reach agreement quickly, avoiding the massive hashing effort required by PoW. Leadership is rotated using a round‑robin schedule, preventing any single node from becoming a bottleneck.

2. Incentive model based on reputation – Instead of rewarding miners with cryptocurrency, the architecture introduces a provenance‑based reputation score. Nodes that correctly validate transactions and contribute valid blocks earn reputation points, which in turn increase their chances of being selected as leaders in subsequent rounds. This aligns incentives with the goal of maintaining data integrity rather than generating economic profit.

3. Publish/Subscribe smart‑contract execution – Traditional Ethereum‑style DApps replicate the entire contract code across all participants, which is impractical for large‑scale IoT. The proposed system uses a publish/subscribe policy: e‑miners subscribe only to the contracts relevant to the subset of devices they manage. If another e‑miner is already executing a particular contract, a node may suspend its own execution to conserve energy, thereby reducing redundant computation.

4. Metadata‑centric in‑chain storage – Full transaction histories are not stored on each edge node. Instead, each e‑miner keeps metadata for the most recent n blocks in memory and stores the corresponding block files locally. Periodically, these files are synchronized with a fog‑level repository, ensuring durability without overwhelming the limited flash storage of the edge device.

The architecture is illustrated in a three‑layer diagram: (i) the sensor/actuator layer, (ii) the e‑miner layer hosting the SD‑Components, and (iii) a fog repository for long‑term archival. The authors argue that this separation “breaks the barrier of constrained computing capabilities” because each component can be placed on a device that best matches its resource profile, while the collective set of components reconstructs a functional blockchain system.

Experimental validation uses Intel Edison boards as e‑miners. The authors measure latency, network traffic, and consensus success rate under various configurations. Results show a reduction in end‑to‑end transaction latency to the order of tens of milliseconds (30–50 ms) compared with cloud‑based solutions that often exceed 200 ms. Bandwidth consumption drops by roughly 40 % because only small, metadata‑rich messages traverse the network, and the publish/subscribe contract model limits unnecessary data propagation. The PBFT consensus achieves over 99 % agreement in scenarios with three to five participating e‑miners, confirming that a lightweight, non‑competitive consensus can function reliably at the edge. Reputation scores increase monotonically for honest nodes, demonstrating that the incentive mechanism effectively encourages correct behavior without monetary rewards.

In conclusion, the paper presents a compelling case for “software‑defined blockchain” as a modular, edge‑friendly alternative to traditional blockchain deployments in IoT. By decoupling blockchain functions, eliminating cryptocurrency incentives, and leveraging reputation‑based motivation, the approach delivers lower latency, reduced network load, and scalability across heterogeneous edge devices. The work opens several avenues for future research, including dynamic component placement algorithms, security hardening against malicious edge nodes, and extending the framework to other low‑power platforms such as ARM Cortex‑M microcontrollers.