“Move fast. Speed is one of your main advantages over large competitors,” said tech investor and entrepreneur Sam Altman.
Providing our customers with a fast data platform for both operational storage and analytical processing was the reason we decided to build Terracotta DB.
Terracotta DB is not just another database. Unlike traditional databases (Relational or NoSQL), Terracotta DB is in-memory, providing predictable latency for operational and analytical workloads at scale. It can handle high-speed operational storage and analytical processing requirements on the very same scale-out cluster. This is a significant differentiator for enterprises wishing to leverage an in-memory data store that is a natural fit for building smart applications that use insights gained in real time to drive their operational decisions.
Terracotta DB is built on top of the distributed in-memory data grid, Terracotta BigMemory, which is currently used by over 500 organizations worldwide including Octo Telematics, Kiabi, and CERN.
Our journey of building this new in-memory data store started about 2 years ago when we realized that more and more of our enterprise customers wanted to run compute and analytics right on top of BigMemory, our In-Memory Data Grid, without having to move the data to another analytic store. They longed to have the option to get rid of their reliance on a disk-based database of record.
Over the years, as we enhanced our current technology stack in terms of features and functions, we also embarked on re-thinking some core aspects of our platform and APIs in collaboration with our customers. To give you a peek into this journey of co-innovation, here are some broad areas where we invested while building Terracotta DB.
- Storage Engine & Platform: A storage tier that supports the flexibility of having a loose schema yet being strongly typed at the same time. This we believe is essential to address modern workloads. Also, a platform that lends itself to supporting multi-model storage is key for broader applicability.
- Simple & Elegant Storage API: A new simple data storage focused API (think database CRUD) to allow storage semantics with additional extensions for analytics.
- Search: A search framework that is purpose-built for a scale-out in-memory data store and supports search without pre-indexing data and optional secondary in-memory indexes.
- Stream-based data pipeline: A mechanism to create data pipelines for high-speed in-situ analytical processing. A domain-specific language (DSL) with portable server-side functions simplifies the programming model.
- Data connectors: webMethods hybrid integration platform provides various adapters to connect to enterprise application stores. A connector for Terracotta DB to facilitate ingestion from the traditional monolithic stores to move data to where it belongs – in-memory!
- Visual analytics out-of-the-box: Tight integration with a data visualization tool is important to enable visual analytics. MashZone NextGen integration with Terracotta DB allows you to create exploratory dashboards by connecting to a running Terracotta cluster.
- IoT streaming analytics: For the Apama Streaming Analytics platform, Terracotta DB serves as a data store for stream processing and enrichment.
- New and improved Ehcache API: Ehcache v3, based on JSR-107 standard, is a completely modern version of Ehcache API for data caching.
Terracotta DB 10.1 is now generally available (GA). A free trial of Terracotta DB is available on softwareag.com.