On Blockchain and IoT Integration Platforms: Current Implementation Challenges and Future Perspectives

Table 3

Comparison between the different consensus algorithms and the blockchains that implement them.

Consensus algorithm	Blockchain-based adaptation	Type	Performance	Limitation	Adaptation for IoT
Proof-of-work (PoW)	Bitcoin [21]	Competition consensus	Robust against DDoS and spam attacks; Resistance against Sybil attacks	High-power consumption; Low throughput and scalability; Double spend risk	Access points would serve as miners instead of individual nodes (IoT devices). Thus, taking the computational load of IoT devices
Proof-of-stake (PoS)	Peercoin [99]; Nxt [100]	Competition consensus	Difficult and more costly to attack; Lower power consumption	Unfair as the richest miners would control the chain	All IoT devices can be selected as validators
Proof-of-activity (PoAC)	Decred [101]	Competition consensus	Enhanced network topology; Lower power consumption	Susceptible to double-spend attack	Fog network architecture can be adapted. Each fog layer would have one miner node which creates an empty block header. IoT devices derive pseudorandom stakeholders using the hash of the block header
Practical Byzantine Fault Tolerance (PBFT)	Ripple [26]; Hyperledger Fabric [36]; Stellar [102]	Cooperative consensus	(i) Less power consumption; (ii) Low variance of the reward	(i) High number of communications between nodes, thus increased number of nodes will result in increased messages sent; (ii) Communication overhead increases exponentially when a new node is added	Fog network architecture would be preferred in this case. Communications would take place between fog nodes
Delegated BFT (DBFT)	Neo [103]	Competition consensus	(i) Less power consumption; (ii) Low computational power required	(i) Delegated nodes operate under real identities	(i) A voting system should be implemented among nodes with a randomised decision of which node to delegate