How Air-cooled HPC Data Centers & Aerospike Meet the GenAI Needs of 6G | Real-time Data Summit 2024

Learn about the research on developing an exascale AI pipeline for the telecommunications industry. Try Aerospike Cloud for free: https://auth.control.aerospike.cloud/...

Join Theresa Melvin, JD, PhD, and Principal Consultant at Aerospike, as she explores the cutting-edge world of exascale AI pipelines for the telecommunications industry. Drawing on her extensive background, including a decade as Global Chief AI Solutions Architect at HPE and a PhD in Quantum Machine Learning (QML), Theresa presents a powerful vision for the future of AI infrastructure, 6G networks, and mixed reality applications.

In this insightful session, Theresa introduces her award-winning research on developing an end-to-end, high-performance AI pipeline, designed to meet the ultra-low latency and extreme throughput demands of next-generation wireless communication systems, AR/VR devices, and digital twins in smart manufacturing. She highlights the limits of traditional networks and the emerging role of quantum networks, as well as the growing need for NoSQL databases that can handle nanosecond-level ingestion speeds.

Enter Aerospike: a high-performance NoSQL database platform that Theresa brought into the telco use case to solve critical performance challenges. She walks through real-world benchmarks achieved using Aerospike on HPE Superdome Flex (now CSS 3200), demonstrating impressive results: sub-millisecond latency, 400ns transfer times, and billions of transactions per second, all optimized for high-performance computing (HPC) and AI/ML workloads.

You’ll learn about:

✅ The future of 6G and exascale AI in telecom
✅ Why classical networks are nearing their limits
✅ How quantum computing and QML are reshaping data infrastructure
✅ Building end-to-end generative AI pipelines with real-time analytics
✅ Key advantages of using Aerospike with Superdome Flex and ScaleFlux CSDs
✅ Designing AI systems that serve both engineering teams and C-level executives via multimodal dashboards
✅ Achieving HPC scalability with flexible compute/storage separation
✅ Deployment options across on-prem, edge, and government cloud environments

Theresa also unveils the solution building blocks required for mission-critical AI systems, especially those needed by Fortune 2000 enterprises, hyperscalers, and exascale customers. Whether you're building smart factory infrastructure, edge AI inference systems, or federated learning networks for secure sectors, this talk offers rare insight into the design and benchmarking of next-gen AI systems.

If you are an architect, engineer, executive, or technologist looking to future-proof your AI and data infrastructure, this session delivers real benchmarks, actionable frameworks, and future-ready insights.

🔧 Featured Technologies:

Aerospike Database 5.7+

HPE Superdome Flex (CSS 3200)

ScaleFlux CSDs (FPGA-based storage)

Generative AI, Graph AI, and Vector Search

HPC, NUMA architectures, 100GbE & Infiniband

HPE Electra 4110/4140, EdgeLine, and GreenLake support

S3-compatible object storage integration

Explore Aerospike: https://www.aerospike.com

Aerospike Database
https://aerospike.com/products/database/

Aerospike Documentation Guide
https://aerospike.com/docs/

Build with Aerospike
https://aerospike.com/docs/develop/

#AIpipeline #Aerospike #QuantumAI