dgraph-io/ristretto
Ristretto
Ristretto is a high-performance in-memory cache and concurrent key-value store for Go applications. It provides a thread-safe memory store that manages strict memory bounds and employs probabilistic set filters to reduce lookup overhead.
The system is distinguished by an admission-policy cache that utilizes frequency sketches and cost-based eviction to maximize hit ratios. It minimizes contention and improves throughput through the use of striped ring buffers and concurrent map sharding.
The project covers a broad range of data management capabilities, including time-based expiration, item frequency estimation, and batch cache updates. It also includes operational telemetry for tracking cache performance metrics and hit ratios, along with the ability to serialize filter states.
Features
- Concurrent Caches - Implements a thread-safe data structure designed for high-concurrency environments to manage object lifecycles.
- Go In-Memory Caches - Provides a high-performance in-memory key-value store specifically optimized for Go applications.
- Cache Eviction Policies - Implements cache eviction policies based on cost and frequency to maximize hit ratios within memory limits.
- Data Caching - Provides a high-performance system for storing temporary data and session state to improve application performance.
- Application State Caching - Provides high-speed in-memory caching for application state and session data to reduce database load.
- Memory-Bounded Caching - Manages strict memory bounds by automatically evicting low-value data to prevent out-of-memory errors.
- In-Memory Caches - Implements a high-performance in-memory cache system for accelerating data access in Go.
- Key-Value Stores - Provides a high-throughput, thread-safe in-memory key-value store.
- Probabilistic Membership Filters - Employs memory-efficient structures like bloom filters to check if an item exists in a set without storing full keys.
- Trending Item Tracking - Uses memory-efficient probabilistic structures to identify and track the most frequent items within the cache.
- Cache Cost Eviction - Removes items based on a weighted combination of access frequency and memory cost to maximize hit ratios.
- Concurrent Data Structures - Implements thread-safe concurrent data structures to handle simultaneous read and write operations in Go.
- Frequency-Based Admission - Implements frequency-based admission logic to determine which items should enter the cache to maximize hit ratios.
- Concurrent Ring Buffering - Buffers incoming data in a ring structure to handle bursts and coordinate the transfer of items to a consumer.
- Lock-Striped Sharding - Utilizes lock-striped sharding to split data across segments, minimizing global lock contention and improving throughput.
- Striped - Buffers updates into multiple independent rings to reduce lock contention and improve write throughput.
- Concurrent - Buffers data updates into striped rings and drains them in batches to reduce contention.
- Expiration Policies - Automatically excludes items from retrieval if their predefined time-to-live expiration has passed.
- Value Update Policies - Modifies existing entries in the cache using a custom function to determine if a value should be replaced.
- In-Memory Cache Invalidations - Provides the ability to remove specific entries from the memory store to free space or invalidate stale data.
- Thread-Safe Collection Operations - Supports adding and checking for elements across multiple concurrent threads using synchronized methods to prevent data races.
- Cache Performance Metrics - Records internal performance data and admission statistics to evaluate cache efficiency.
- Time-To-Live Expirations - Automatically removes stale data based on predefined time-to-live durations to maintain cache freshness.
- Caching Libraries - High-performance cache library.
