Traditional blockchains are built on a principle of radical redundancy: every participant stores every transaction that has ever occurred. While this provides security in a world of powerful servers, it creates a "storage wall" for the Internet of Things (IoT). A Raspberry Pi or a mobile sensor cannot store a ledger that grows by gigabytes every month.
To bring the security of decentralization to the edge, researchers are developing lightweight distributed ledgers. These systems move away from the "all-nodes-store-everything" model, instead using hierarchical structures, periodic compaction, and information theory to keep the ledger small enough for hardware with strictly limited memory.
The Storage Bottleneck
In a standard blockchain, the ledger is monotonically increasing. As Huawei Huang and Song Guo note in their survey of the state-of-the-art, this overhead acts as a centralizing force. If only massive data centers can afford the storage to be "full nodes," the network is no longer truly decentralized.
For IoT devices, this isn't just a theoretical concern—it's a hardware limit. Ali Dorri and Salil S. Kanhere argue in Blockchain in Internet of Things: Challenges and Solutions that we must strip away the "heavy" components of traditional blockchains—specifically the requirement to maintain a full historical ledger—to make them viable for smart environments.
Hierarchical and Hybrid Architectures
One way to solve the bloat is to change who does the work. Instead of a flat network where every node is equal, researchers propose hierarchical models.
In the Microchain framework, Ronghua Xu and Yu Chen propose a hybrid consensus mechanism. By offloading heavy computations to a "fog" layer—intermediate nodes with more power than a sensor but closer than the cloud—edge devices can participate in security without bearing the full weight of the chain. This approach is particularly effective in trusted avionics systems, where Erik Blasch and colleagues use microservices to manage access control on resource-constrained flight hardware.
Forgetting the Past: Pruning and Compaction
For many IoT applications, the entire history of a device is less important than its current state. If a sensor is tracking food temperature, we care deeply about the current safety status, but perhaps less about a specific reading from three weeks ago.
Anurag Jain and Pan Hui address this in their work on Mneme, a ledger for mobile device-to-device ecosystems. Mneme uses periodic ledger compaction. By summarizing old transactions into a "compressed" state, the system allows mobile devices to maintain a verifiable history without the storage requirements of a full chain. Similarly, Youzhi Zhang proposes Quick Searching Blockchains for food traceability, specifically designed to allow mobile phones to verify data without hosting massive datasets.
Coded Ledgers: Efficiency through Information Theory
The most radical reduction in storage comes from changing how data is distributed across the network. Instead of every node storing a full copy (100% of the data), what if every node stored a tiny, mathematically unique "shard" of the data?
Changlin Yang and Zibin Zheng introduced a rateless coded blockchain that uses Raptor codes—a type of error-correcting code. In this system:
- The ledger is encoded into many small fragments.
- Each IoT node stores only a few fragments.
- As long as a sufficient number of nodes are online, the full ledger can be reconstructed.
This method achieves a 99.6% reduction in storage and communication overhead. It turns the "redundancy" of blockchain from a storage burden into a distributed puzzle, where the network as a whole remembers everything, but no single device has to.