Realize faster time to data insights for better outcomes
Integrated into Intel® Xeon® processors, Intel® Data Streaming Accelerator (Intel DSA) offloads data movement from the CPU to accelerate workloads and free resources for analytics. Built-in tools help simplify data movement and transformation.
Lower TCO and enhance sustainability
Intel DSA not only increases performance per watt1 to help organizations meet their sustainability goals, but also allows businesses to optimize workloads so fewer servers are needed for the same performance envelope, helping to reduce TCO significantly.
Enable high performance
Intel DSA includes resources to accelerate performance, including the open-source Intel® Data Mover Library (Intel® DML), which improves database, communications, and application performance. It helps analysts and admins access data faster with higher IOPS.
Streamline integration
As a built-in accelerator in Intel Xeon Scalable processors, Intel DSA simplifies integration by eliminating the need for specialized hardware, additional vendors, and disparate support. We also provide developers with access to resources to streamline implementation.
Explore Intel DSA in action
Intel DSA helps optimize data movement and transformation operations for demanding workloads.
Data center
Offload data movement to reduce the data center burden when performing memory copying or zeroing and free up CPU cycles for advanced analytics workloads.
Storage
Use for data movement in storage appliances using non-transparent bridging (NTB), cyclic redundancy check (CRC) generation, and DIF generation, with or without simultaneously moving data.
Networking
Copy data in packet-processing pipelines. For example, use as a virtual switch offload for inter-VM packet switching.
VM migration and fast checkpointing
Use when a VM manager (VMM) must identify a VM’s modified pages and efficiently send them to a destination machine.
Data movement between devices
Optimize data movement between a peer accelerator device and host memory or two peer devices to free up CPU cycles.
Data movement to / from / between VMs
Free up CPU cores from performing routine infrastructure tasks such as moving data between VMs, containers, and bare-metal hosts.
1 Up to 2.06x higher Perf/W for large packet sequential read with 5th Gen Intel Xeon Platinum 8592+ using integrated Intel DSA vs. 3rd Gen Intel Xeon Platinum 8380 oob.
