Sorry, you need to enable JavaScript to visit this website.

SNIA Developer Conference September 15-17, 2025 | Santa Clara, CA

Technical Staff

Dell

Ugur is Technical Staff at Dell Storage CTO. Her current focus on storage for AI and object storage. She drives research, incubation, positioning, and adaptation of emerging technology and architecture for AI/ML storage. Prior to her role at Dell, Ugur completed her PhD, specializing in distributed caching for object storage in datacenters. Her interests lie broadly in the fields of storage systems, object storage, caching, distributed systems, and AI.

KV-Cache Storage Offloading for Efficient Inference in LLMs

Submitted by Anonymous (not verified) on

As llm serve more users and generate longer outputs, the growing memory demands of the Key-Value (KV) cache quickly exceed GPU capacity, creating a major bottleneck for large scale inference systems.

In this talk, we discuss KV-cache storage offloading, a novel technique that enables inference acceleration by relocating attention cache data to high speed, low latency storage tiers. This approach alleviates GPU memory constraints and unlocks new levels of scalability for serving large models.

‎ On-Prem Object Storage: S3 Ecosystem and Its Role in Various Workloads

Submitted by Anonymous (not verified) on

Over the years, on-prem object storage has significantly evolved, with S3 and its API compatibility being key topics at S3 Plugfest. This presentation will explore the comprehensive S3 ecosystem, highlighting how various S3 features support diverse use cases, particularly in the context of rapidly advancing AI workloads. Additionally, we will delve into the role of S3 extensions in enhancing this ecosystem.
 

Subscribe to Ugur Kaynar