PCRAFT: Capacity Planning for Dependable Stateless Services

TL;DR

PCRAFT is a system that combines empirical measurements and probabilistic models to optimize capacity planning for dependable stateless services, balancing availability, performance, and cost across different deployment schemes.

Contribution

It introduces a hybrid approach for capacity planning that integrates empirical data with probabilistic modeling to minimize resource use while satisfying availability constraints.

Findings

Cloud deployments need fewer nodes than on-premises.

Passive failover requires fewer nodes than active route anywhere in on-premises.

Additional integrity mechanisms improve quality but increase resource requirements.

Abstract

Fault-tolerance techniques depend on replication to enhance availability, albeit at the cost of increased infrastructure costs. This results in a fundamental trade-off: Fault-tolerant services must satisfy given availability and performance constraints while minimising the number of replicated resources. These constraints pose capacity planning challenges for the service operators to minimise replication costs without negatively impacting availability. To this end, we present PCRAFT, a system to enable capacity planning of dependable services. PCRAFT's capacity planning is based on a hybrid approach that combines empirical performance measurements with probabilistic modelling of availability based on fault injection. In particular, we integrate traditional service-level availability mechanisms (active route anywhere and passive failover) and deployment schemes (cloud and on-premises) to quantify the number of nodes needed to satisfy the given availability and performance constraints. Our evaluation based on real-world applications shows that cloud deployment requires fewer nodes than on-premises deployments. Additionally, when considering on-premises deployments, we show how passive failover requires fewer nodes than active route anywhere. Furthermore, our evaluation quantify the quality enhancement given by additional integrity mechanisms and how this affects the number of nodes needed.