← Back to Home
05 / RESEARCH UPDATES — PROF. HAMMOUD

When One Computer Is Not Enough

The Fascinating World of Distributed Systems

Q1. What happens when the problem you want to solve becomes too big for any one computer?

When a certain problem becomes too big for a single computer, the concept of distributed computing comes in. The same computing task is now split across multiple machines that work together to reach a solution.

Main Problem → Smaller, bite-sized pieces → Each piece sent to a node → Each computer finishes its assigned task → Final answer/solution sent to the central hub

Q2. Suppose 1,000 computers work together. Do you now have one computer that is 1,000 times more powerful? Why or why not?

Connecting 1,000 computers does not automatically equate to having a single computer that is just as powerful — a machine doesn't just become more powerful overnight. A hard task requiring the input of all the computers cannot always be completed more efficiently just because there are more machines. Splitting a task up like that won't always work as efficiently as we think it will.

There are limitations, such as:

  • Delays in communication
  • Software limitations
  • Error handling
  • Individual speed
  • Setup and management
  • Handling simultaneous tasks

Q3. Can 1,000 computers agree on something if some of them fail or even lie?

Yes! Even if some of the 1,000 machines malfunction, lose their network connection, or send false information, they can still reach a consensus. Consensus algorithms are used to address the Byzantine Generals Problem, a foundational problem in computer science.

However, there is a strict mathematical limit to how many failures or "liars" the system can handle while still reaching a correct agreement. Crash faults and Byzantine faults are certain limitations tied to this problem — though there are protocols designed to refine and overcome these issues.

Q4. When you use ChatGPT, Google, Instagram, or an online game, where is the computation actually happening?

Most of the computation that occurs when you use these applications takes place in distant data centres that are either owned or leased by the businesses providing the service. The heavy lifting — such as generating AI text, searching billions of web pages, or running complex game logic — is done on powerful servers thousands of miles away, while your local device just manages what you see and interact with.

  • ChatGPT → OpenAI's cloud servers
  • Google Search → Google's global data centres
  • Instagram → Meta's massive data centres

Q5. If you could make millions of computers behave like one dependable machine, what could humanity build that we cannot build today?

There are certain tasks humanity hasn't yet achieved that relate to the astronomical sector — real-time modelling of Earth and truly accurate forecasting haven't yet been achieved by our networks. Advanced scientific discoveries, like the instant molecular design of a protein or complex material physics, are other fields humanity hasn't been able to fully solve.

Things we could build:

  • Real-time Earth modeling
  • Accurate forecasting
  • Instant molecular design
  • Complex material physics
  • Unified cognition
Reference Links