What is Kinetica?
Kinetica is a real-time analytical database designed to deliver high-speed, complex analytics at scale on modern GPUs and CPUs. It enables organizations to build applications that combine vector, graph, spatial, time-series, and OLAP operations using SQL, providing a unified platform for multimodal analytics. The database supports real-time data ingestion from various sources, including streaming and batch data, and offers GPU acceleration for enhanced performance on demanding workloads.
With features like vector search, time-series analytics, spatial intelligence, and graph analytics, Kinetica empowers AI agents and generative AI applications by unlocking both text and tabular enterprise data. It offers flexible deployment options, including cloud-managed services and on-premise solutions, with enterprise-grade security and scalability for mission-critical use cases across industries such as telecommunications, logistics, financial services, and defense.
Features
- Real-Time Analytics: Complex analytics on streaming and batch data with high-speed query execution
- GPU Acceleration: Leverages NVIDIA GPUs for 10X performance on complex workloads
- Multimodal Analytics: Combines vector, graph, spatial, time-series, and OLAP operations in a single query
- AI Agent Support: Empowers AI agents with real-time, real-world data retrieval and fusion
- SQL Compliance: ANSI SQL compliant for easy temporal and spatial analytics
- Flexible Deployment: Options include free cloud, developer edition, dedicated cloud, and enterprise on-premise solutions
- Enterprise Security: Cell-level security with support for LDAP, Active Directory, and Kerberos authentication
- High Availability: Node and process failover with cluster resiliency for continuous operation
Use Cases
- Detecting financial crime and fraud through real-time pattern analysis
- Optimizing logistics routes and tracking goods movement in real-time
- Analyzing network data and coverage patterns for telecommunications
- Assessing insurance claims based on catastrophic weather events
- Building common operating pictures for defense situational awareness
- Optimizing energy usage in industrial facilities with spatial and time-series analytics
- Powering generative AI applications with real-time enterprise data retrieval
- Performing complex spatial queries for geographic data visualization
FAQs
-
What are the differences between Kinetica PAYG and BYOL options?
Kinetica PAYG (Pay As You Go) is best for new customers or unpredictable usage, with license fees based on cluster size and usage hours billed to your CSP account. BYOL (Bring Your Own License) is for customers with predictable workloads or existing licenses, where you input a license key and only pay for AWS infrastructure costs. -
When should I choose GPU accelerated clusters over CPU-only?
GPU accelerated clusters are recommended for use cases benefiting from GPU performance, such as interactive visualization of large-scale geographic data or high-cardinality joins. They cost more due to GPU compute nodes and additional license fees. -
How do I get a license key for Kinetica BYOL?
Contact [email protected] to obtain a license key for the Bring Your Own License option. -
What size cluster should I choose for my deployment?
Start with smaller cluster sizes for testing and evaluation, then scale to larger sizes for production as needed. Each size increase roughly doubles performance, with infrastructure costs varying based on cluster size, hours, region, and CSP discounts. -
What additional benefits does Enterprise Edition provide?
Enterprise Edition offers custom sizes and configurations, external authentication and authorization (LDAP, Active Directory, Kerberos), and full HA replication for mission-critical workloads.
