A Blockchain Architecture for the Internet of Things

Success in an Internet of Things application typically requires the connected, concerted operation and management of a large number of distributed and loosely coupled smart devices that identify and trust each other. While the IoT promises that this integration should ideally map to a decentralized hardware and software platform, current solutions are mostly based on centralized infrastructures. The disadvantages of a centralized infrastructure are, among others: high maintenance costs; low interoperability due to restricted data aggregation with other centralized infrastructures; and single points of failure (SPOF) against security threats.

Decentralization of an IoT infrastructure brings advantages including reduction of the amount of data transferred to the cloud for processing and analysis, improvements in security and privacy of the managed data, and more concerted and autonomous operations. For example, in smart home environments, IoT devices autonomously exchange and process data; assure data security, operations accountability, device identification; and collectively and autonomously execute smart homes operations. Ensuring the veracity of these operations means achieving distributed consensus across IoT devices.

In the recent IEEE International Blockchain Conference paper “Hybrid-IoT: Hybrid Blockchain Architecture for Internet of Things – PoW Sub-blockchains”, Gokhan Sagirlar, Barbara Carminati, Elena Ferrari, John D. Sheehan, Emanuele Ragnoli explain the first phase of development for an end-to-end, decentralized framework to autonomously manage the orchestration of geographically distributed IoT networks. It achieves this through the combination of analytics, well-established blockchain community-based open source modules and a novel code base written by the team. In essence, they adapted the Bitcoin Proof of Work- PoW consensus protocol and coded analytics that optimally cluster the IoT devices into separate blockchains, or sub-blockchains.

One of the biggest challenges in the integration of blockchain into IoT is scalability. Due to the massive number of devices and resource constraints, deploying blockchain in IoT is particularly challenging. The optimal blockchain architecture must scale to many IoT devices (they become the peers on the blockchain network), and it must process a high throughput of transactions. Hybrid-IoT, the platform designed in this work, exploits both PoW blockchains and Byzantine Fault Tolerant (BFT) protocols to achieve scalability. First, PoW blockchains achieve distributed consensus among many IoT devices, the peers on the blockchain. Then, Hybrid-IoT leverages a BFT inter-connector framework to achieve interoperability among the sub blockchains.

To measure and qualify the performance of their approach, they first defined a set of blockchain-IoT integration metrics and tested our design with a simulation framework. They studied the sensitivity of blockchain parameters including blockchain block sizes and block generation intervals, device locations, and number of peers. The data gathered from the tens of thousands of simulations with hundreds of thousands of simulated devices was done on an IBM POWER8 supercomputer and helped us frame this problem from the perspective of multi objective optimization. That is, how to optimize the clustering and selection of the sub blockchains subject to scalability, security, peer’s roles and other integration metrics. Examination of these results led to the definition of a set of “sweet spot” guidelines and a set of analytics that cluster geographically distributed IoT devices into sub-blockchains, according to preferences made by the user (e.g., one could give more importance to security or to scalability, and consequently apply different weights to the relevant cost function).

They tested this in a geographically distributed framework, in which virtualized IoT devices became peers on Hybrid-IoT with different roles within the different PoW sub blockchain. Thousands of experiments were performed to validate the scalability of the system first, and to check its security on a second step. For instance, they tested it for DDOS and other types of attacks. The performance evaluation proves the validity of the PoW sub blockchain design under the “sweet spot” guidelines. Furthermore, they demonstrate that the “sweet spot” guidelines also prevent security vulnerabilities.

The plan is to continue this work with additional testing through both simulated and physical IoT networks.