Wednesday, August 9th
9:45-10:50 AM
SARC-202-1: CXL RAS (System Architectures Track)
Organizer: Jim Pappas, Director, Technology Initiatives, Intel

Paper Title: Server-grade RAS and advanced ECC functions for resilient CXL-based systems

Paper Abstract: Modern data-centric applications such as AI and ML require various CPUs, GPUs, memory, storage, networking, and software. To address performance and capacity requirements in these new architectures, CXL was introduced. With CXL, the scale of deployment of data-centric applications is expected to grow extremely fast. In addition to performance improvements, mission-critical servers with CXL memory controllers must provide robust error detection, correction, containment, and recovery across the ecosystem since any failure or system downtime can cost hundreds of thousands to millions of dollars per hour. To address these fault tolerance requirements, CXL provides resilience from the hardware all the way to the application software through three RAS features: 1 Reliability: detect and correct faults 2 Availability: aid uninterrupted operation with minimal degradation 3 Serviceability: proactively diagnose and repair failed components Attendees will gain key insights into CXL 2.0 RAS features, including advanced Error Correction Capabilities (ECC), that ensure robust operation of the overall system and enable data-centric workloads to run with minimal interruptions.

Paper Author: Sandeep Dattaprasad, Senior Product Manager, Astera Labs

Author Bio: Sandeep Dattaprasad is a Senior Product Manager at Astera Labs with 15+ years of experience in semiconductor, software diagnostic tools, developing security strategies and firmware development for complex SoC product lines including Compute Express Link™ products, SAS RAID controllers, SAS expanders, and PCIeŽ switches. He is also a contributing member of the CXL Consortium. At Astera Labs, Sandeep focuses on driving product strategy for new market segments by translating data center bottlenecks into profitable and competitive hardware and software product roadmaps using CXL technology.