Pronto

Abstract

Non-Volatile Main Memories (NVMMs) promise an opportunity for fast, persistent data structures. However, building these data structures is hard because their data must be consistent in the wake of a failure. Existing methods for building persistent data structures require either in-depth code changes to an existing data structure using an NVMM-aware library or rewriting the data structure from scratch. Unfortunately, both of these methods are labor-intensive and error-prone. Pronto is a new NVMM library that reduces the programming effort required to add persistence to volatile data structures using asynchronous semantic logging (ASL). ASL is generic enough to allow programmers to add persistence to the existing volatile data structure (e.g., C++ Standard Template Library containers) with very little programming effort. Furthermore, ASL moves most durability code off the critical path, and our evaluation shows Pronto data structures outperform highly-optimized NVMM data structures written with other libraries by a large margin. CCS Concepts. • Hardware → Emerging technologies; • Software and its engineering → Software libraries and repositories; • Information systems → Data structures; • Computer systems organization → Processors and memory architectures.

TL4x

“…Many software-based techniques for NVMM programming use a redo-log [17,18,24,32,34,36,53,56], or an undo-log [6,7,30,56,57]. TL4x does not use logging.…”

Converting Concurrent Range Index Structure to Range Index Structure for Disaggregated Memory

“…However, legacy storage systems still depend on volatile index structures. Moreover, newly developed persistent indexes are less stable than mature volatile indexes [26][27][28][29][30]. Hence, the researchers focus on leveraging the existing concurrent indexes for persistent memory.…”

“…Pronto [28] is a framework to convert a legacy volatile index structure to a persistent index structure. In Pronto, the volatile index structures are stored in both volatile DRAM and persistent memory.…”

Efficient algorithms for persistent transactional memory

“…For Trinity there is no bound, and Quadra is bounded by the capacity of the count variables that hold the number of words modified during the transaction. Some existing software-based techniques in the literature use redo-log [22,25,27,40,43,47,60,61], while others use undo-log [8,9,35,42,61,62]. PMDK [53] uses an hybrid undo-redo log approach, with undo for the transactional changes and redo for the allocator.…”

“…Workload A executes 10 6 operations, with half being value overwrites and the other half key lookups, while workload B executes 5% overwrites and 95% lookups. In addition to the previous KV stores, for the YCSB benchmark we compared with Pronto [47], a recent technique that uses a persistent semantic redo-log approach and a volatile replica of the data structure. Pronto does not support transactions and creates a new checkpoint file at user request, which must be done before the semantic log fills up.…”