warning

The Fallacies of
Distributed Systems

Design decisions made on false assumptions are the root of most system failures.

The 8 Fallacies

Originally coined by Peter Deutsch in 1994, these eight assumptions are still the hidden traps in modern cloud architecture.

1. The network is reliable.
2. Latency is zero.
3. Bandwidth is infinite.
4. The network is secure.
5. Topology doesn't change.
6. There is one administrator.
7. Transport cost is zero.
8. The network is homogeneous.

1. The Network is Reliable

The Reality: Packets Drop

Switches fail, cables are cut, and congestion causes loss. You cannot assume your request will arrive.

Architectural Fix

Retries, Timeouts, and Idempotency keys.

2. Latency is Zero

The Reality: Light is Slow*

Even at the speed of light, a round trip across the globe takes ~200ms. In microservices, every hop adds up.

Architectural Fix

Caching, CDNs, and Batching requests.

3. Bandwidth is Infinite

The Reality: Congestion is Real

Moving massive JSON blobs or raw logs can saturate the network quickly, leading to slow responses for everyone.

Architectural Fix

Compression (Protobuf/gRPC) and filtering data at the source.

Interview Guidance

The "Skeptical Architect"

To show seniority in an interview, be a skeptic. When a plan depends on a synchronous network call, ask: "What happens when this times out?"

Mentioning Fallacies

Explicitly name-dropping the "Fallacies of Distributed Computing" shows you have a formal background in system design theory, not just ad-hoc experience.