Cloud Native Orchestration — Lifecycle Management on Kubernetes

Kubernetes Operator Support

Deployment of the GigaSpaces Kubernetes Operator allows the following features in Kubernetes:

• Vertical Scaling — InsightEdge enables auto scaling up/down (more/less CPU/RAM) in 1-click per service, with no downtime and auto recovery

• Horizontal Scaling — InsightEdge provides auto scaling out/in (more/fewer nodes) in 1-click.

• Deploy Service — InsightEdge now provides one-click deployment of a microservice

• One-Click Installation and Auto Provisioning

Enhanced SQL Support

The introduction of quickly-accessed colocated data, improved JOIN support, and additional SQL subqueries provides a significant improvement in query performance and versatility.

Colocated Data — Broadcast Objects

• Broadcast Objects are Space Objects that are duplicated for each Processing Unit (code plus data) in the cluster. This ensures that the Broadcast Object data is available immediately in the same memory space as the executing code, without the need for a cross-node data request.

• SQL statements that reference Broadcast Objects, such as joins of a partitioned Object with a Broadcast Object, can run up to ten times faster than SQL that requires non-broadcast data.

Improved JOIN Support

• SQL JOINs that are joined using the routing key of the Space Object run on the server side, using colocated data for faster performance.

Additional SQL Subqueries

• In addition to existing support in the SQL WHERE clause, subqueries are now also supported in the SQL FROM clause.

RAM Footprint Optimizations

InsightEdge's new Storage Optimization feature reduces the RAM footprint of your in-memory data store with minimal performance impact.

Storage Optimization can be implemented selectively, on an entire Space Object (table) or for specific properties (fields), to optimize only the data that will benefit most. It can be invoked through the Ops Manager UI, or by Object annotations to customize the generated code.

Kafka Connect

InsightEdge has introduced a Kafka Connect sink connector in the Confluent Hub. This allows easier mapping and integration of data streams, allowing consolidation of multiple heterogeneous data sources to a unified in-memory Data Space. The readme file is available here.

 

Administration and Security

Enhanced troubleshooting and data analysis provide the GigaSpaces system administrator with the tools they need to keep the system running at optimal efficiency.

• Troubleshooting - Ops Manager now provides General Availability for Explain Plan in the SQL editor.