VAST Data, an AI-centric data platform company, has introduced its global data infrastructure offering, the VAST Data Platform. The platform merges storage, database, and virtualized compute engine services to meet the demands of future AI applications.
Moving past Large Language Models (LLMs), the VAST Data Platform's goal is to enable AI-assisted discovery. It allows machines to not only process but learn from data, potentially assisting in solving significant global issues, such as disease treatment and climate change.
The platform handles both unstructured and structured data types in real-time, bridging the gap between event-driven and data-driven architectures. Its underlying technology, the Disaggregated Shared-Everything (DASE) architecture, is designed to overcome conventional infrastructure tradeoffs, thus providing a solid foundation for deep learning.
The VAST Data Platform comprises the VAST DataStore, DataBase, DataEngine, and DataSpace. The DataStore is a scalable storage system for unstructured data, while the DataBase structures this data by introducing a semantic layer. The DataEngine refines the unstructured data into a structured format, and the DataSpace offers a global namespace for consistent data storage, retrieval, and processing.
Renen Hallak, VAST Data's CEO and Co-Founder, stated that the platform democratizes AI abilities and helps organizations unlock their data's value. Currently, the DataStore, DataBase, and DataSpace are available within the platform, and the DataEngine will be released in 2024.
