Growth Laws (Moore, Koomey, Bell, Zimmerman)

Context: FIT1043_MOC · the exponential trends behind Big Data’s growth · four named laws · why “big” is a moving target

Quick Revision

  • 🎯 Objective: explain the exponential growth of computing/data ➔ four laws (Moore, Koomey, Bell, Zimmerman).
  • 📦 Core Components: Moore (transistors) | Koomey (energy efficiency) | Bell (computer classes) | Zimmerman (surveillance).
  • ⚡ Key Constraint: Koomey and Bell are corollaries of Moore’s Law — efficiency and new device classes follow from doubling transistor counts.

📝 How It Works

1. Moore’s Law (Gordon Moore, Intel, 1965)

  • Claim ➔ transistors per chip double every ~2 years (from 1975) ➔ more/bigger/faster memory and CPUs (smaller = faster).
  • Nature ➔ microprocessor growth is exponential; the pace is currently slowing.

2. Koomey’s Law (Jonathan Koomey, Stanford, 2010)

  • Claim ➔ computations per joule of energy double every ~1.57 years — energy efficiency doubles.
  • Consequence ➔ battery needed falls ~100× per decade ➔ ubiquitous computing (a corollary of Moore’s Law).

3. Bell’s Law (Gordon Bell, DEC, 1972)

  • Claim ➔ roughly every decade a new, lower-priced computer class forms on a new platform/network/interface ➔ new usage + a new industry.
  • Confirmed ➔ PCs, mobile computing, cloud, IoT.

4. Zimmerman’s Law (Phil Zimmermann, 2013, PGP creator)

  • Claimsurveillance constantly increasing, privacy constantly decreasing.

⚖️ Core Decision Matrix

LawWho / yearWhat growsRate / effect
MooreMoore, 1965transistors per chip×2 every ~2 years
KoomeyKoomey, 2010computations per joule×2 every ~1.57 years
BellBell, 1972new computer classesone per ~decade
ZimmermanZimmermann, 2013surveillance (privacy ↓)continuous

When It Flips: Moore drives the others — doubling transistors (Moore) implies more compute per joule (Koomey) and cheaper new device classes (Bell); those together fuel the data explosion that makes "big" a moving target.

🧠 Active Recall