Regular Sequential Serializability and Regular Sequential Consistency

Abstract

Strictly serializable (linearizable) services appear to execute transactions (operations) sequentially, in an order consistent with real time. This restricts a transaction's (operation's) possible return values and in turn, simplifies application programming. In exchange, strictly serializable (linearizable) services perform worse than those with weaker consistency. Switching to such services, however, can break applications.This work introduces two new consistency models to ease this trade-off: regular sequential serializability (RSS) and regular sequential consistency (RSC). They are just as "strong" for applications; we prove any application invariant that holds when using a strictly serializable (linearizable) service also holds when using an RSS (RSC) service. Yet they are "weaker" for services; they allow new, better-performing designs. To demonstrate this, we design, implement, and evaluate variants of two systems, Spanner and Gryff, weakening their consistency to RSS and RSC, respectively. The new variants achieve better read-only transaction and read tail latency than their counterparts. *CCS Concepts: • Information systems → Parallel and distributed DBMSs; Distributed database transactions.

NOC-NOC: Towards Performance-optimal Distributed Transactions

“…Benchmarking with Realistic Workloads. Our evaluations only consider synthetic YCSB-like benchmarks that are widely used by the database community [10,22,26,36,37,53]. Although we have experimented with a variety of workload parameters, standard benchmarks such as TPC-C would provide more insights into our proposed optimizations.…”

“…A spectrum of isolation levels. 𝐴 → 𝐵 means that 𝐴 is strictly weaker than 𝐵. RC: read committed [6]; RA: read atomicity [5]; RA+: RA with read-your-writes [5,48]; TCC: transactional causal consistency [34]; TCCv: TCC with convergence [2,31]; SI: snapshot isolation [6], including its weaker variant parallel SI [45]; SER: serializability [42]; POS: process-ordered serializability [13,32]; RSS: regular sequential serializability [22]; SSER: strict serializability [42].…”

“…Scylla-PORT [34] is NOCS-optimal (satisfying NOC for reads) with process-ordered serializability, but it does not support transactional writes. Spanner-RSS [22], providing regular sequential serializability, improves latency for Spanner's reads by reducing their blocking chances in the presence of conflicting writes. The strictly serializable system NCC [33] exhibits optimal performance for its read-only transactions (i.e., NOC for reads) in the best case; however, it still aborts/retries transactions in general, thereby not satisfying O 𝑅 .…”