Infoscience

Report

Log-Free Concurrent Data Structures

Non-volatile RAM (NVRAM) makes it possible for data structures to tolerate transient failures, assuming however programmers have designed these structures in a way to preserve their consistency upon recovery. Previous approaches, typically transactional, inherently used logging, resulting in implementations that are significantly slower than their DRAM counterparts. In this paper, we introduce a set of techniques that, in the common case, remove the need for logging (and costly durable store instructions) both in the data structure algorithm as well as in the associated memory management scheme. Together, these generic techniques enable us to design what we call log-free concurrent data structures, which, as we illustrate on linked lists, hash tables, skip lists, and BSTs, can provide several-fold performance improvements over previous, transaction-based implementations, with overheads of the order of milliseconds for recovery after a failure. We also highlight how our techniques can be integrated into practical systems, by introducing a durable version of Memcached that is able to maintain the performance of its volatile counterpart.

Related material