Roadmap

1. Improve read buffer management

• Benchmark different approaches
• Cache implementation

2. Dynamically grow read buffers

• See Java 8's Striped64 for example heuristic
• Reduces clean up penalty in low contention usages
• Reduces memory usage

3. Code generate policy configurations

• Reduces the memory overhead for caches by only retaining the required fields

4. Use relaxed reads and writes

• Reduces memory usage by replacing atomics with volatile fields
• Reduces CPU stalls due to memory barriers

5. Improve write buffer algorithm

• Use a multiple producer / single consumer unbounded queue (replacing ConcurrentLinkedQueue)
• Add combining arena to existing implementation

6. Tracing

• Instrument the caches to capture tracing data

7. Profile and code audit

• Request external reviews
• Implement atomic merge operation
• Signal that incomplete async computations cannot expire
• Extra care needed to review weak/soft value race conditions

8. Implement JCache (jsr107) adapters

• May be delayed until after initial release

9. Release 1.0
10. Advanced eviction policies

• Simulator implementations (2Q, LIRS, TinyLFU, etc)
• Evaluate and adopt