HPKV is a high-performance key-value store offering ultra-low latency, disk persistence, and scalability at a low operational cost.
Features
Ultra-low latency
Delivers sub-microsecond response times across all operations ensuring quick access to stored data.
Persistent Storage
Utilizes advanced storage layouts and techniques for reliable disk persistence with direct I/O and zero-copy methods.
Scalability
Enables linear scaling with data growth without compromising performance, perfectly handling increased loads.
Developer Friendly API
Offers a simple and easy-to-use API compatible with popular programming languages and frameworks, facilitating seamless integration.
WebSocket API for Real-time Applications
Allows real-time collaboration and instant updates in applications using WebSocket capabilities.
REST API
Provides a simple and efficient way to interact with your stored data using HTTP requests.
gRPC API
Enables faster communication with your data through a high-speed remote procedure call framework.
Latencies as low as 250 ns
Offers incredibly low data access latencies, ideal for time-critical applications, ensuring rapid data retrieval and processing.
Remote IO Control (RIOC)
Allows fast and secure style remote procedure call protocol to access HPKV, enhancing performance and security through TLS 1.3.
Data Recovery Guarantee
Ensures your data is safely stored and recoverable in case of server issues, enhancing data security and integrity.
Enterprise Ready Features
Includes high availability, disaster recovery, and security features tailored for enterprise-grade applications.
Cost Efficiency
Reduces operational costs by 4-5 times compared to leading in-memory key-value stores through advanced techniques and algorithms.
API Key Authentication
Ensures secure access to the APIs using API keys, supporting HTTPS/WSS encryption for robust security.
Live Demo Application
Offers a functional Todo application as a demonstration of the WebSocket API capabilities, allowing multiple users to see real-time data changes and collaboration.
Record Insertion/Update
Allows you to insert or update a key-value pair in the database using a simple HTTP POST request. You can choose to partially update an existing value or replace it entirely.
Fetch Record
Enables you to retrieve the value associated with a specific key using HTTP GET. This is essential for accessing stored information quickly and efficiently.
Delete Record
Facilitates the removal of a specific key-value pair from the database using an HTTP DELETE request, ensuring that outdated or unnecessary data can be discarded.
High-Performance Key-Value Store
Optimized for ultra-low latency, HPKV offers consistent sub-microsecond response times even under heavy load, providing 2x faster performance than leading in-memory solutions.
Nanosecond-Level Performance
Delivers consistent high-speed data access at nanosecond-level latency, ideal for applications requiring rapid data processing.
Innovative KV Store Architecture
Utilizes a redesigned data storage and retrieval method that maximizes performance and efficiency on modern hardware.
Sophisticated Caching Mechanisms
Employs advanced caching techniques to enhance performance by reducing data retrieval times.
Disk persistence
Provides disk-based storage with immediate consistency and low latency for reliable data management.
High Throughput Operations
HPKV offers significantly higher throughput compared to Redis, achieving up to 1.73 million operations per second for GET requests, making it much faster for single operation read and write tasks.
Low Latency Disk Persistence
With its local vectored call interface, HPKV maintains low latencies, even in disk persistence, achieving median latencies under 1 microsecond for batch operations.
Batch Operation Efficiency
HPKV excels in batch operations, performing multiple operations in a single batch more efficiently than Redis, with improvements of up to 6.1x for delete operations.
Consistent Performance
The system demonstrates consistent performance with low variability in latency between different operations, making it reliable under various workloads.