The convergence of Mobile, Social, Big Data, and Cloud has placed increasing demands on today’s applications to react instantaneously to changes in data at a large scale. A delay of a few seconds can mean the difference between engaging or losing a customer for a retailer, increased liquidity or fraud risk for a financial institution, or escalation of an adverse occurrence in a manufacturing process or IoT network.
In-memory data grids (IMDGs), such as Red Hat JBoss Data Grid, are well-positioned to meet this goal of instant response. IMDGs offer the same flexibility of data formats as leading NoSQL databases, and additionally can deliver far superior performance and extreme scalability via auto-sharding across dozens of random-access memory (RAM) nodes on commodity hardware. You can configure replication to ensure that data is highly available across servers, racks, or data centers. Additionally, IMDGs can be configured for ACID-based data integrity like relational databases, whereas most NoSQL databases offer only eventual consistency.
Red Hat JBoss Data Grid (JDG), rated as one of the Leaders in the latest Forrester WaveTM, is an excellent option for event-driven computing. By setting a cluster-wide listener in JDG, your application can run arbitrary business logic in real-time, triggered by creation, update or deletion of data in the grid. For example, consider an e-commerce platform which stores all open customer orders in JDG. When the status of an order is updated from ‘Processing’ to ‘Ready for Shipping’, a listener can update a metric that tracks the average time to fulfil orders. Another use of this model is near-caching, in which your client application can store often-accessed data entries locally for best performance. When the value of the entry on the grid changes, an invalidation event is sent to the client.
JDG 6.6, now generally available, brings key new features in this area. We have added support for listeners to react to Expiration events, which occur when the specified time-to-live for a data entry elapses and its value is no longer valid. We have also introduced the ability to specify Continuous Queries, so that the result set of a query is kept up to date as the data stored in JDG changes, without having to rerun the query. Both features are supported in Library Mode and Client-Server mode.
Thinking about a NoSQL database such as MongoDB, CouchBase, or Cassandra for your next project? Press the fast button instead for your NoSQL data – store and compute in-memory with Red Hat JBoss Data Grid, get blazing performance without compromising scalability, high availability, and consistency!