Experts on Data
 
 

PM Summit 2019 Presentation Abstracts

2019 Persistent Memory Summit Presentations

Keynote: Converging Memory and Storage

Frank Hady, Fellow, Intel Corporation

Abstract

Order of magnitude advances in non-volatile memory density and performance are upon us bringing significant systems level architecture opportunities. Intel® 3D XPoint™ Memory, with much higher performance than NAND and greater density than DRAM has entered the platform to address the gap between the two. Intel® Optane™ SSDs are already available as fast storage volumes delivering significant application performance improvements to applications. Persistent Memory will take us further. With low latency and fine grained addressability this non-volatile high capacity memory is available directly, avoiding the 4kByte block size and multiple microsecond storage stack that accompany system storage. Making Persistent Memory requires a series of hardware and software advances and can be made accessible to the programmer in a number of different ways. The resulting high capacity, high performance, persistent memory creates opportunities for rethinking algorithms to deliver much higher performance applications. This presentation will explore this new memory technology, show how it fits within the platform and the resulting performance, describe the modes of operation in which it is being made available, and explore a few applications promise to benefit.

Characteristics of Persistent Memory: Performance, Capacity, Persistence – which ones?

Paul Grun, Open Fabrics Alliance Vice Chair; Advanced Technology Development, Cray, Inc

Abstract

Persistent Memory exhibits several interesting characteristics potentially including persistence, capacity, byte addressability, and others. This session explores the various characteristics usually mentioned in a discussion of persistent memory, and relates them to how applications might leverage those characteristics.

Impact on Application Development: SNIA NVM Programming Model in the Real World

Andy Rudoff, Member, SNIA NVM Programming Technical Work Group and Persistent Memory SW Architect, Intel Corporation

Abstract

In an ideal world, furnishing an application with access to Persistent Memory would result in an immediate improvement in application performance or efficiency. The reality is that may be true for some applications, others may rely on middleware, and still others may require modifications. This session explores some of these alternatives, and does so in the context of the NVM Programming Model, which provides a framework for accessing Persistent Memory.

Enabling Persistent Memory Through the OS

Stephen Bates, PhD, Chief Technology Officer, Eideticom; Neal Christiansen, Principal Software Development Lead, Windows Storage and File Systems Group, Microsoft

Abstract

One way to look at Persistent Memory is as another resource available to an application.  This session looks at how modern OS’ treat Persistent Memory.

Enabling Persistent Memory Use in Java

Steve Dohrmann, Intel Corporation

Abstract

Persistent memory offers new ways to program with long-lived data.  Java is a popular language for data center applications such as databases.  This session will describe ways to access persistent memory from Java and relate some of our application experience, including development of a persistent memory storage engine for the Cassandra database.

Elastify Cloud-native Spark Application with Persistent Memory

Yue Li, Co-Founder, MemVerge; Junping Du, Chief Architect, Tencent Cloud

Abstract

Cloud native deployment has become one of the major trends for large scale Big Data analytics. Compared to on-premise data center, cloud offers much stronger scalability and higher elasticity to Big Data applications. However, cloud is also considered to be less performant than on-premise alternatives due to virtualization and cluster resource disaggregation.

We present a new cloud native Spark application architecture backed by persistent memory technology. The key ingredient of this architecture is a novel acceleration engine that uses Intel’s 3DXPoint technology as disaggregated external memory resource. We discuss how the performance of multiple aspects of data processing can be improved using this new architecture.

As a key takeaway, audience will gain understanding on the benefits of latest persistent memory technology, and how such new technology could be leveraged in cloud data processing architecture.

What You Can Do with NVDIMM-N, NVDIMM-P

Rob Peglar, Advanced Computation and Storage; Wendy Elsasser, ARM

Abstract

NVDIMMs are currently the most popular and widely used Persistent Memory in computing systems.  The JEDEC industry standard NVDIMM-N is commonly available as the first modern Persistent Memory module.  It provides Persistent Memory to systems at a module-level using DRAM and Flash instead of Persistent Memory media.  The upcoming NVDIMM-P standard will now also enable use of Persistent Memory media and unlock its many new capabilities including higher capacity, lower power, and lower cost per bit along with persistence.  This session focuses on the new capabilities and benefits for system designers, developers and end users provided by these Persistent Memory module standards, NVDIMM-N and NVDIMM-P.

Persistent Memory in Mission Critical Architecture (How and Why)

Adam Roberts, Western Digital

Abstract

DRAM use cases can be elevated with a persistent memory layer. Larger capacity points, and lower price points combined with the ability to remove protection schemes needed to protect volatile memory allow for a denser, less expensive and more efficient solution for memory architectures in the datacenter

New Interconnects

Millind Mittal, Distinguished Engineer/Chief Architect, Xilinx; Kurtis Bowman, Dell; Steve Fields, IBM

Abstract

In this session we take a look at three powerful interconnect technologies.  A brief overview is provided for all three: CCIX, Gen-Z, and OpenCAPI, each presented by industry experts.  The focus will be on the current status of each interconnect, and how they are relevant to persistent memory.  The panel will then answer questions from the audience

Back to Top

Persistent Memory Performance Benchmarking and Comparison

Eden Kim, Chief Executive Officer, Calypso Systems Inc.

Abstract

Persistent Memory (PM) is becoming more of a mainstream potential storage architecture and solution as more applications begin to use PM. As such, there is increasing interest in benchmarking PM performance and comparing PM to other storage architectures. This presentation shows PM NVDIMM-N performance compared to NVMe, [NVMe-oF, RPM] and SAS SSD. 

Making Remote Memory Persistent  

Idan Burstein, Mellanox Technologies, Inc.

Abstract

This session will provide an overview of what is required to extend persistent memory across a fabric. Learn how control plane and resource management software need to change to support persistent memory. Understand new fabrics capabilities that support remote persistent memory, as well as the ongoing standardization efforts to support it. What are the tradeoffs between these technologies, and how can applications and operating systems take better take advantage of persistence?

Persistent Memory Media

Kevin Conley, President and CEO, Everspin; Amigo Tsutsui, Senior Business Producer, Sony Semiconductor Solutions Corporation; Danny Sabour, Vice President, Avalanche

Abstract

Persistent Memory technology is continuing to develop. There are today a number of ways of providing memory which is both byte addressable and persistent. This panel of technologists will discuss several promising emerging media technologies. They will explore how new applications like artificial intelligence and machine learning are demanding higher density media combined with at lower power and lower cost and what new memory technologies being implemented today, and forthcoming products, will address the cost/performance balance. 

Analysts’ Perspective

Tom Coughlin, Coughlin Associates; Jim Handy, Objective Analysis

Abstract

It’s easy to get excited over emerging technologies and the possible impact they may have on the trajectory of computer architecture development. In this session, analysts will bring a sense of perspective to the emerging field of Persistent Memory.