Zero Allocation in Java: what it is and why it matters

Zero Allocation — is an approach to writing code in which no unnecessary objects are created in heap memory during runtime.

The main idea: fewer objects → less GC → higher stability and performance.

⚡ 1. Simple explanation

Every time an object is created in Java:

• memory is allocated in the heap
• GC must then clean it up
• pressure arises on memory and CPU

Zero allocation aims to completely avoid this in the hot path.

🔥 2. Example: allocations

❌ Bad:

String[] parts = line.split(",");

• an array is created
• strings are created
• regex works

✔️ Better:

// manual parsing through indexes
int start = 0;
for (int i = 0; i < line.length(); i++) {
    if (line.charAt(i) == ',') {
        // processing without creating objects
        start = i + 1;
    }
}

3. Where hidden allocations appear

1. Collections

List<Integer> list = new ArrayList<>();

• Integer = boxing
• each element = object

✔️ zero allocation option:

int[] arr = new int[1000];

2. Streams

list.stream().map(...).collect(...);

• pipeline objects
• lambda capture
• iterable wrappers

❌ almost never zero allocation

3. String operations

"a" + "b" + "c"

• new String objects are created
• StringBuilder inside

4. Exceptions

• creation of Exception object
• stack trace allocation

⚙️ 4. Where Zero Allocation is Really Important

🔥 Only in the hot path:

• high-load backend
• financial systems
• stream processing
• real-time systems

If the code is called rarely — zero allocation is not critical.

🚀 5. How Zero Allocation is Achieved

• use of primitive types (int, long)
• avoiding boxing (Integer, Long)
• manual work with arrays
• reusing objects (object pooling)
• avoid streams in hot path

🧠 6. Important point about JVM

Zero allocation ≠ magic.

JVM can:

• escape analysis remove object
• inline methods
• optimize temporary objects

But this is not guaranteed.

💣 7. Where Zero Allocation Breaks

• Stream API
• boxing/unboxing
• String operations
• lambda closures
• exceptions

📊 8. Comparison of Approaches

🏁 Result

Zero Allocation is not about "forbidding objects at all".

It’s about controlling the hot path and minimizing GC pressure.

Main rule:

The fewer objects → the less GC → the more stable latency