TerseJSON Performance Benchmarks

1Memory Efficiency: The Primary Advantage

Why Memory Matters More Than Wire Size

Network compression (gzip/Brotli) only helps data in transit. Once the payload arrives, it's fully decompressed and parsed into memory. TerseJSON's Proxy solves a different problem: it prevents unused fields from ever allocating in the first place.

Memory Benchmarks: 1,000 Records x 21 Fields

2The Proxy Does Less Work, Not More

A common misconception: "Adding a Proxy layer must add overhead." In reality, the Proxy does FEWER operations than standard JSON.parse because it parses a smaller string and defers expansion.

CPU Benchmark Results

3TerseJSON vs Binary Formats

Protobuf and MessagePack are often suggested as alternatives. But they have a fundamental limitation: full deserialization is required. Every field must be parsed and allocated, even if you only need 3 fields from a 21-field object.

4Real-World Use Cases

TerseJSON shines when you fetch objects with many fields but only render a subset. This is extremely common:

// CMS fetches 1000 articles with 21 fields each
const articles = await terseFetch('/api/articles');

// But list view only needs 3 fields
articles.map(a => ({
  title: a.title,
  slug: a.slug,
  excerpt: a.excerpt
}));

// Result: Only 3 keys translated per object
// The other 18 fields stay compressed in memory
// Memory saved: ~100%

5Network Bonus: Smaller Payloads Too

Memory efficiency is the primary benefit, but you also get significant network savings. TerseJSON compresses repetitive JSON keys, reducing payload size by 30-80%.

Bandwidth Savings by Endpoint Type

Stacking with Gzip/Brotli

6Mobile Performance Impact

Mobile devices are memory-constrained and often on slower networks. TerseJSON helps on both fronts: less memory allocation AND smaller payloads.

7Enterprise Cost Savings

At enterprise scale, memory + bandwidth savings compound. Less memory per request = more requests per server. Smaller payloads = lower egress costs.

*At $0.09/GB (AWS CloudFront pricing). Memory savings translate to reduced server instances.

8Integration: 5 Minutes to Memory Savings

Server Setup (2 lines)

import { terse } from 'tersejson/express';
app.use(terse());

Client Setup (1 line change)

// Before
const data = await fetch('/api/users').then(r => r.json());

// After
import { createFetch } from 'tersejson/client';
const terseFetch = createFetch();
const data = await terseFetch('/api/users').then(r => r.json());
// data works exactly the same — Proxy handles expansion transparently

9Run the Benchmarks Yourself

Don't take our word for it. Run the benchmarks on your own machine:

# Clone the repo
git clone https://github.com/timclausendev-web/tersejson
cd tersejson/demo

# Memory benchmark (requires --expose-gc)
node --expose-gc memory-analysis.js

# CPU benchmark
node cpu-benchmark.js

10Summary: Memory-First JSON Processing