Apache Mnemonic (Incubating)

Clone this repo:
  1. eb9e7f8 MNEMONIC-455: Simply the Factory proxy construction by Yanhui Zhao · 29 hours ago master
  2. a2abf28 MNEMONIC-463: Should load native library in memory allocator service lazily by taojie · 2 weeks ago
  3. 1dc7126 MNEMONIC-461: Clean generated data files by Wang, Gang(Gary) · 3 weeks ago
  4. c476763 MNEMONIC-460: Add helper code to verify the heap memory usage of data structures by Wang, Gang(Gary) · 3 weeks ago
  5. d998b5a MNEMONIC-440 & MNEMONIC-441: Skip test cases for Spark&Hadoop integration in run-all script if running on JDK9 by Ganhuazhang · 3 weeks ago


Mnemonic Official Website

Build Status

Apache Mnemonic is a non-volatile hybrid memory storage oriented library, it proposed a non-volatile/durable Java object model and durable computing service that bring several advantages to significantly improve the performance of massive real-time data processing/analytics. developers are able to use this library to design their cache-less and SerDe-less high performance applications.


  • In-place data storage on local non-volatile memory
  • Durable Object Model (DOM)
  • Durable Native Computing Model (DNCM)
  • Object graphs lazy loading & sharing
  • Auto-reclaim memory resources and Mnemonic objects
  • Hierarchical cache pool for massive data caching
  • Extensible memory services for new device adoption and allocation optimization
  • Durable data structure collection(WIP)
  • Durable computing service
  • Minimize memory footprint of on-heap
  • Reduce GC Overheads as the following chart shown (collected from Apache Spark experiments)
  • Drop-in Hadoop MapReduce support
  • Drop-in Hadoop Spark support