Timezone isn't accessible, please provide the correct parameters
eventFeedUrl=http://realintelligence.com/customers/expos/00Do0000000aAt2/FMS_xmlcreator/a0J5c00001MW1eJ_specific-event-list.xml
trackCategory=Session
eventID=a0J5c00001MW1eJ
timezone=
duration=PTH
, NaNth
3:10-4:15 PM
AIAP-302-2: Storage for AI Part 2 (Artificial Intelligence Applications Track)
Paper Title: Accelerating Big Data and AI with a Remote Persistent Memory Pool

Paper Abstract: Big Data and AI converging on the same platform has become an increasingly popular trend as it is eliminating the need for data movement between two dedicated, separate clusters, with a resultant improved workload performance and reduction in costs. However, this trend poses some new challenges, especially in the compute and storage disaggregated environment. Persistent memory is able to offer high performance, high capacity and data persistence at a lower cost, and thus how to leverage persistent memory to address the aforementioned challenges remains an interesting topic. This talk will introduce the notion of a remote persistent memory pool - a solution created to accelerate ephemeral data access for Bid Data and AI workloads in the unified cluster, such as for data shuffling and caching. We will also introduce the architectural design, implementation, and resultant performance of this solution, as well as the results of how it improves key analytics workloads' performance by 4.5x with an improved failover rate of 30x for mission critical jobs.

Paper Author: Jian Zhang, Software Development Engineer, Intel

Author Bio: Jian Zhang is a senior software engineering manager at Intel, where he and his team primarily focus on end-to-end data analtyics and AI pipeline optimizations on Intel platforms. He has over 10 years of experience in performance analysis and optimization for many open source projects such as Xen, KVM, Swift, Ceph, Spark, and Hadoop. He earned a master's degree in Computer Science and Engineering at Shanghai Jiaotong University. He has mulitple publications and has also presented at the OpenStack Summit, Vault, Strata Data conference, Cephalocon, OFA workshop, Flash Memory Summit, Data + AI Summit etc.