计算存储之旅：动机、经济性和可消费性.pdf

1、OCP Global Summit October 18,2023|San Jose,CAGary Grider,Los Alamos National LabLA-UR-23-29917A Computational Storage Journey:Motivation,Economics,and Consume-abilityNeed flexibility in where computation is done(host,network,device)as economics will change over time Its not just energy and time to i

2、nsight,some of these analytics require the same size analytics footprint as the simulation footprint(petabyte of ram)making analytics not always as feasibleData Agnostic OffloadsServer memory BW does not allow many passes over streaming dataData Aware OffloadsAnalytics is often multiple orders of ma

3、gnitude less reading than writingYou just have a hard time finding what you are looking for(filter/index/histogram/etc.)Can we add metadata/indexing/ordering to data as it is written with almost no overhead and reap huge wins on read(time,hdwr resources,energy)For ScienceParticle methods-“Ordered”ro

4、w-based analytics(KV)Grid methods-columnar-based analytics Large Complex Grid methods -THE KITCHEN SINKComputational Storage WhyData Agnostic Offload(ABOF)Data Aware Row And Col Based AnalyticsCS data agnostic/data aware learningsWhy Columnar and why Offload to near Storage?A Columnar end-to-end demo with Object CSLANL use case(high level)OCS Initiators/Targets Open Ecosystem Demo(s)Partnering has been the key to this exploration!What does this notionally look like?Replace this image with your full-screen photoThanks for your time!OCP Global Summit|October 18,2023|San Jose,CA