New tool
Some Spark jobs accelerate better than others. Find yours
New tool
Beyond performance
Lower TCO at scale
Reduce the total cost of ownership of your longest, most compute-intensive workloads. Better ROI, no trade-offs.
More intelligence per watt
Higher CPU throughput, lower energy draw. The same fleet, working harder.
Expand AI Capacity
The headroom for your next AI initiative exists today. FuseCore makes it available.
Unlock your hardware
MLIR JIT compiler
FuseCore's compiled, not interpreted. No middle layer, no wasted cycles.
In-process execution
All your APIs, libraries, and job logic remain exactly the same, just faster.
Compute or memory?
Reduce execution time, or reduce memory pressure - your call.
Adoption without disruption
Plug-and-play
Add the Xonai FuseCore JAR like any standard Spark plugin and switch it on with one flag.
Production grade
Maintain full control and revert to default Spark anytime.
No learning curve
No new dashboards, syntax, or operational rituals. Same Spark, minus the wait.
FuseCore accelerates Spark within existing execution boundaries, using the same security controls already in place. All without expanding access scope or data exposure.
FuseCore for Apache Spark
Xonai FuseCore integrates with open-source Apache Spark up to 3.5.3 and the following data platforms:
- Amazon EMR up to 6.12.0
- Databricks up to 15.4 LTS
- Dataproc 2.0.X, 2.1.X and 2.2.X release line of versions
Note that FuseCore is frequently being updated to support new Spark versions.
FuseCore runs as a Spark 3 plugin. You activate it by adding the JAR we provide and enabling the spark.plugins property.
When you run a job via spark-submit, Spark still selects the physical plan as usual and FuseCore executes an equivalent plan underneath.
Existing solutions tackle cloud spending reduction by improving resource provisioning and/or tuning application parameters, and may only have a one-time benefit.
FuseCore accelerates Spark data processing speed far beyond the default Spark engine (Catalyst), and delivers seamless hardware acceleration and reduced resource utilization regardless of how optimally deployed Spark workloads already are.
No. We intentionally designed FuseCore to be API-compatible with existing runtimes for Spark, including proprietary ones that may modify query plans to improve performance, such as the Databricks and EMR runtime.
As Spark is an in-memory compute engine, the more time queries spend on doing physical computations between reads and writes, the more benefit they are expected to get by using FuseCore. These are typically high compute data transformation jobs with heavy aggregations, joins and sorting stages.
