Capacity of blockchain based Internet-of-Things: Testbed and analysis
Abstract: An integration of Internet-of-Things (IoT) and blockchain becomes increasingly important to secure IoT data in an anti-tampering manner. Challenges arise from the immense scale of IoT and the resultant impact of network partitioning on blockchain. We design a new testbed to evaluate the impact, where resource-limited IoT devices, acting as light nodes, are an integral part of a blockchain. Our testbed is built on the Ethereum platform with non-trivial modifications on key modules. The partitioning of IoT is emulated by probabilistically dropping blocks travelling among the miners. We also propose a new discrete-time Markov chain model to validate our testbed and analyze the impact of block mining rates and network conditions on the capacity of public blockchains. The model is first formed to be non-ergodic with an infinite state space. By exploiting the eventual consistency property of blockchain, the model is collapsed to be ergodic and approximated with a finite state space and significantly improved tractability. Both the testbed and analysis reveal the blockchain capacity can be improved by accelerating the block mining rates which, however, increases stale blocks. Our analysis provides an asymptotic upper bound for the blockchain capacity.
Loading