Anna is a low-latency, autoscaling key-value store.

The core design goal for Anna is to avoid expensive locking and lock-free atomic instructions, which have recently been shown to be extremely inefficient. Anna instead employs a wait-free, shared-nothing architecture, where each thread in the system is given a private memory buffer and is allowed to process requests unencumbered by coordination. To resolve potentially conflicting updates, Anna encapsulates all user data in lattice data structures, which have associative, commutative, and idempotent merge functions. As a result, for workloads that can tolerate slightly stale data, Anna provides best-in-class performance. A more detailed description of the system design and the coordination-free consistency mechanisms, as well as an evaluation and comparison against other state-of-the-art systems can be found in our ICDE 2018 paper.

Anna also is designed to be a cloud-native, autoscaling system. When deployed in a cluster, Anna comes with a monitoring subsystem that tracks workload volume, and responds with three key policy decisions: (1) horizontal elasticity to add or remove resources from the cluster; (2) selective replication of hot keys; and (3) data movement across two storage tiers (memory- and disk-based) for cost efficiency. This enables Anna to maintain its extremely low latencies while also being orders of magnitude more cost efficient than systems like AWS DynamoDB. A more detailed description of the cloud-native design of the system can be found in our VLDB 2019 paper.

You can find the code here.