Silentium! Run-Analyse-Eradicate the Noise out of the DB/OS Stack

TL;DR

This paper demonstrates how to achieve near-deterministic query latencies in multi-tenant database systems by systematically removing noise from the software stack, approaching bare-metal performance without harming other tenants.

Contribution

It introduces a controlled experimental approach to eliminate noise sources within the system software stack to improve latency determinism in multi-tenant databases.

Findings

Query latencies close to bare-metal performance achieved.

No significant impact on other tenants' workloads.

Existing operating systems can be sufficient with proper configuration.

Abstract

When multiple tenants compete for resources, database performance tends to suffer. Yet there are scenarios where guaranteed sub-millisecond latencies are crucial, such as in real-time data processing, IoT devices, or when operating in safety-critical environments. In this paper, we study how to make query latencies deterministic in the face of noise (whether caused by other tenants or unrelated operating system tasks). We perform controlled experiments with an in-memory database engine in a multi-tenant setting, where we successively eradicate noisy interference from within the system software stack, to the point where the engine runs close to bare-metal on the underlying hardware. We show that we can achieve query latencies comparable to the database engine running as the sole tenant, but without noticeably impacting the workload of competing tenants. We discuss these results in the context of ongoing efforts to build custom operating systems for database workloads, and point out that for certain use cases, the margin for improvement is rather narrow. In fact, for scenarios like ours, existing operating systems might just be good enough, provided that they are expertly configured. We then critically discuss these findings in the light of a broader family of database systems (e.g., including disk-based), and how to extend the approach of this paper accordingly.