Computers provide value to people as platforms for application development. The raw computation, memory, and I/O can be applied to specific problems. A computer provides resources and services to applications, and these applications ultimately provide utility to people, or to other applications.

The decentralised infrastructure of Web3 will be valuable in a similar way. The value of a Web3 system lies in its potential as a platform for application development. This platform is, of course, different to that of a personal computer, or a data centre, or a mobile device. Each of these is a platform for applications, providing similar basic resources of computation but with differences in architecture that give rise to different capabilities, e.g. around proximity to the user, scale, internal redundancy etc. These differences make each platform more or less suitable for certain applications, but each still ultimately provides value by supporting those applications.

The Web3 community is building a world computer. Or rather, many world computers, and many applications atop them. It’s enlightening to consider an analogy between the emerging components of Web3 infrastructure and the physical components of a personal computer (or data centre etc). The analogy clarifies the basic resources that form a strong platform, and points to future developments.

• Blockchain VM ⟷ CPU. A blockchain’s VM provides compute services, the execution of code. The term blockchain here includes side-chains, rollups, and other layer-two execution services that evaluate code. The dominant metric is instruction throughput.
• Chain state ⟷ RAM. Blockchain state represents the random-access memory of our world computers: primary storage. It’s fast to access, synchronous, but limited in size and expensive. Both code and data are stored and accessed here. (The RAM of the machines executing a blockchain is more like CPU cache in the decentralised analogy). More RAM is better.
• Messages ⟷ I/O and IPC. Messages from the outside world are how people interact with applications executing on the CPU. Blockchain messages are a bit richer than key presses or mouse clicks, but perform the same basic function. Messages sent between applications resemble inter-process communication. More throughput is better.

Early computers didn’t have much more than a CPU, RAM and I/O. But the next most important resources to be developed was secondary storage.

• Filecoin ⟷ secondary storage. Storage networks are the tape, disk, and solid state persistent storage for our world computers. Large and cheap, but slower to access. Important metrics are capacity, and latency and throughput of access.

Secondary storage supports data-intensive applications: rich media like images and audio, files, and databases. Most blockchains, like most early computers, don’t have secondary storage built-in. Filecoin’s opportunity is to be the storage network for all world computers, driving network effects to provide larger, cheaper, faster, and more secure storage to all.

This analogy of the Web3 platform as a PC also points to a missing, or at least underdeveloped, piece of infrastructure.

• Bridges ⟷ network. Bridges are how applications running on different world computers communicate with each other. Important metrics are latency and throughput.

Secure bridging is a hard problem. Today, I think we lack a convincing implementation of a secure, low-trust, general-purpose inter-blockchain bridge. But it’s an active area of research and development, and I expect we’ll see good bridge technology emerge in the next few years. And that’s something to strive for, because the eventual development of high-throughput, general purpose message bridging may be as transformative to the Web3 platform as the Internet was to the PC platform.