Blockchain Infrastructure Landscape: A First‑Principles Framework

Background

A program consists of three essential elements: storage, computation, and communication. Different platforms (mainframes, PCs, mobile devices, cloud services) implement these elements in distinct ways, and specialized components cooperate to achieve desired outcomes.

Basic Blockchain Components

Storage: token storage, databases, file systems / binary large‑object storage.

Computation: stateful business logic, stateless business logic, high‑performance computing (HPC).

Communication: data, value, state.

Blockchain Infrastructure Overview

The following table maps each program element to representative blockchain technologies and projects.

Storage

Computation

Communication

Token storage

Bitcoin, Zcash, .*

Stateful business logic

Ethereum, Lisk, RChain, EOS, Tezos, … client‑side computation (JS, Swift)

Data

TCP/IP, HTTP

File system / binary data

IPFS/FileCoin, Eth Swarm, Storj, Sia, Tieron, LAFS

Stateless business logic

Crypto‑condition based chains (BigchainDB, Ripple, …)

Value

Interledger, Cosmos

Database

BigchainDB + IPDB, IOTA

High‑performance computing

TrueBit, Golem, iEx.ec, Nyriad, VMs, client‑side computation

State

Polkadot, Aeternity

Storage

Token storage: Stores value (assets or securities) such as Bitcoin, airline miles, or digital copyrights. The primary operations are issuance and transfer, with mechanisms to prevent double‑spending.

Bitcoin and Zcash are pure token‑storage systems. Ethereum uses its native token to pursue the “world computer” vision, though most users employ it for token issuance.

Other tokens (e.g., Golem’s GNT, Envoke’s IP‑authorization token) operate on top of existing blockchains without driving the entire network.

Database: Structured data storage (relational tables, JSON documents, key‑value, time‑series, graph). Traditional distributed databases like MongoDB and Cassandra can store terabytes to petabytes with write rates exceeding one million ops/sec.

SQL separates specification from implementation, enabling cross‑industry use.

BigchainDB is a decentralized document store built on MongoDB (or RethinkDB) that adds blockchain features such as immutability and token support. IPDB is a public, governance‑enabled version of BigchainDB.

IOTA can be viewed as a time‑series database.

File system / binary data storage: Decentralized file systems (IPFS, Tahoe‑LAFS) and commercial services (FileCoin, Storj, Sia, Tieron) store large objects. Ethereum’s Swarm, Dat, and Swarm‑JS provide similar capabilities.

Computation

Smart‑contract platforms perform decentralized computation and are divided into two main categories: stateless (combinational) logic and stateful (sequential) logic, plus a third category for high‑performance computing.

Stateless (combinational) logic: No internal state; expressed as truth tables or conditional code. Easy to verify at scale. Implemented via crypto‑conditions (CC) in protocols like Interledger, with libraries in JavaScript, Python, Java, etc.

Stateful (sequential) logic: Maintains memory (e.g., registers in a CPU). Turing‑complete platforms such as Ethereum, Lisk, RChain, DFINITY, EOS, Tezos, Fabric, Sawtooth support this model, enabling complex applications but also increasing verification difficulty.

Formal verification and careful design are needed to avoid costly bugs (e.g., the DAO hack).

Clients can run lightweight decentralized computation using JavaScript or Swift in browsers or mobile devices—often called “fat clients” to distinguish from “fat protocols”. Example stack: JS + IPDB (via js‑bigchaindb‑driver ), IPFS (via ipfs.js ), and Ethereum (via web3.js ).

High‑Performance Computing (HPC): Used for rendering, machine learning, circuit simulation, weather forecasting, protein folding, etc., typically via clusters of CPUs, GPUs, or TPUs. Decentralized HPC implementations include Golem, iEx.ec, Nyriad, TrueBit, and container‑based approaches that sell compute results as token‑gated services.

Communication

Communication is split into three layers: data, value, and state.

Data: The classic TCP/IP stack (originating from ARPANet) connects networks at the packet level.

Value: Protocols such as Interledger Protocol (ILP) enable transfer of assets across networks, preventing double‑spending without relying on centralized exchanges.

State: Mechanisms like Polkadot and Aeternity allow the migration of on‑chain state (e.g., smart‑contract balances) between networks.

Examples

Projects combine multiple components:

Ujo: IPFS/Swarm + IPDB + Ethereum for decentralized music.

Innogy: IPFS + IPDB + IOTA for supply‑chain and IoT.

Related Work

Joel Monegro’s “fat protocol” framework treats every component as a network protocol. Other perspectives include Fred Ehrsam’s DAPP stack and various import‑based library approaches.

Historical references: the original BigchainDB whitepaper (2016) and subsequent updates presented at the 2017 Consensus conference.

Future

Just as AWS evolved from a single S3 storage service to a suite of 50+ services (compute, databases, etc.), the decentralized ecosystem is expected to follow a similar trajectory, with each blockchain component mirroring a cloud service while retaining unique features such as token storage.