JPDHeap: A JVM Heap Design for PM-DRAM Memories

Real-world e-commerce systems need large cache capacities. Persistent memory (PM) can be employed to enlarge JVMs’ cache capacities, meanwhile they incur heavy write slowdowns and garbage collection overheads. This paper proposes JPDheap, a JVM heap design for PM-DRAM memories. A JPDheap is composed of a standard Java heap on DRAM and another heap on PM. The core insight is to separate heap objects and store them on DRAM or PM, allowing objects to be accessed much more efficiently. Our evaluation shows that JPDheap outperforms state-of-the-art heap designs by up to 115.96% in increasing applications’ throughput and by up to 87.03% in decreasing the average latency.

I. Introduction

Real-world e-commerce systems need large cache capacities. For example, the transaction peak reached 540,000 transactions per second in the 2019 Alibaba’s shopping festival [18]. The performance of the e-commerce system needs to be guaranteed for such an event, particularly during the transaction peak period. The rapidly increasing data in recent years requires JVMs to provide large cache capacities to applications, aiming at achieving high cache hit rates for heavy workloads. Conventional JVMs use DRAM as cache, which suffers from its limited capacity, high power consumption, and high cost. Meanwhile, DRAM is not scalable enough: a JVM’s heap memory is constrained by the main memory, whose size is eventually constrained by the memory DIMMs.