DPU Use Cases
Powered by MPPA® architecture, Kalray DPUs are deployed where performance, efficiency, and scalability matter most — enabling smarter AI, faster storage, and more secure networks across industries.
AI & compute
AI workloads—from inference to real-time decision-making—require compute architectures that are both efficient and adaptable. Kalray’s DPU accelerators are built to execute AI tasks where latency, power, and deterministic performance matter.
Common Use Cases
- Edge AI inference (e.g., telecom, industrial systems)
- AI offload for data centers
- Low-latency analytics and decision-making pipelines
WHY MPPA?
- Deterministic, real-time response
- High performance per watt
- Programmable for evolving AI models
Storage Acceleration
Data movement, encryption, and I/O processing are major bottlenecks in modern storage infrastructures. Kalray’s DPUs bring parallelism and dedicated processing to eliminate those bottlenecks—especially for NVMe and object storage environments.
Common Use Cases
- Inline encryption & compression
- Data integrity and replication at wire speed
- NVMe-oF and object storage offload
WHY MPPA?
- Removes CPU/GPU load for storage functions
- Scales to high-throughput workloads
- Secure, deterministic packet and data processing
5G Radio Access Networks (RAN)
MPPA-powered DPUs are ideal for 5G and next-generation telecom workloads. Whether for FEC offload, real-time scheduling, or distributed compute, Kalray delivers the performance RAN infrastructure demands.
Common Use Cases
- Forward Error Correction (FEC) acceleration
- Data plane offload in O-RAN environments
- AI-enhanced radio resource management
WHY MPPA?
- Real-time, low-latency execution
- Deterministic packet scheduling
- Software-programmable updates for evolving 5G workloads
Security Acceleration
Security workloads—from encryption to zero-trust packet processing—require processing architectures that ensure speed without compromising control. Kalray’s DPUs are used in critical infrastructure and defense systems for exactly this reason.
Common Use Cases
- Inline encryption/decryption
- Packet inspection and filtering
- Secure boot and trusted compute
WHY MPPA?
- Designed for inline, real-time security tasks
- Software-defined, adaptable policy engines
- Used in defense, aerospace, and telecom applications
Custom Acceleration
From aerospace to deep tech, not every workload fits a predefined category. Kalray’s MPPA-based DPUs can be custom-designed for virtually any compute-intensive or latency-sensitive task. Whether you’re building specialized pipelines, protocol stacks, or unique AI flows, our architecture gives you full control over performance, power, and flexibility.
Common Use Cases
- Proprietary protocol offload
- Specialized AI model inference
- Sensor fusion & data pre-processing
- Industry-specific encryption or encoding schemes
WHY MPPA?
- Massively parallel architecture tailored to customer-defined workloads
- Flexible software-defined pipelines and deterministic scheduling
- Proven track record in aerospace, defense, and HPC co-design projects