Storage Developer Conference

Software-based memory-to-memory data movement is common, but takes valuable cycles away from application performance. At the same time, offload DMA engines are vendor-specific and may lack capabilities around virtualization and user-space access. This talk will focus on how SDXI(Smart Data Acceleration Interface), a newly formed SNIA TWG is working to bring an extensible, virtualizable, forward-compatible, memory to memory data movement and acceleration interface specification. As new memory technologies get adopted and memory fabrics expand the use of tiered memory, data mover acceleration and its uses will increase. This TWG will encourage adoption and extensions to this data mover interface.

Learning Objectives: A new proposed standard for a memory to memory data movement interface,A new TWG to develop this standard,Usecases where this will apply to evolving storage architecture with memory pooling and persistent memory

Since the advent of the Smart Phone over a decade ago, we've seen several new "Smart" technologies, but few have had a significant impact on the data center until now. SmartNICs and SmartSSDs will change the landscape of the data center, but what comes next? This talk will summarize the state of the SmartNIC market by classifying and discussing the technologies behind the leading products in the space. Then it will dive into the emerging technology of SmartSSDs and how they will change the face of storage and solutions. Finally, we'll dive headfirst into the impact of PCIe 5 and Compute Express Link (CXL) on the future of Smart Acceleration on solution delivery.

Learning Objectives: Understand the current state of the SmartNIC market & leading products.,Introduce the concept of SmartSSDs and two products available today.,Discuss the future of Device to Device (D2D) communications using PCIe, CXL/CCIX.,Lay out a vision for where composable solutions leveraging multiple devices on a PCIe buss communicating directly.

The ZFS L2ARC is now more than 10 years old. Over that time, a lot of secret incantations and tribal knowledge have been created by users, testers, developers, and the odd sales or marketing person. That collection of community wisdom informs the use and/or tuning of ZFS L2ARC for certain IO profiles, dataset sizes, server class, share protocols, and device types.

In this talk, we will review a case study in which we tested a few of these L2ARC myths on an NVMe-capable OpenZFS storage appliance. Can high-speed NVMe flash devices keep L2ARC relevant in the face of ever-increasing memory capacity for ARC (primary cache) and all-flash storage pools?

Learning Objectives: 1) Overview of ZFS L2ARC design goals and high level implementation details that pertain to our findings; 2) Performance characteristics of L2ARC during warming and when warmed, plus any tradeoffs or pitfalls with L2ARC in these states; 3) How to leverage NVMe as L2ARC devices to improve performance in a few storage use cases.

5G Wireless technology will bring vastly superior data rates to the edge of the network. However, with this increase in bandwidth will come applications that significantly increase overall network throughput. Video applications will likely explode as end users have large amounts of data bandwidth to operate. Video will not only require advanced compression but will require large amounts of data storage. Combining advanced compression technologies with storage will allow a high density of storage and compression in a small amount of rack space with little power, ideal for placement at the edge of the network. NVMe based module provides the opportunity to use computational storage elements to enable edge compute and video compression.

This presentation will provide technical details and various options to combine video and storage on an NVMe interface. Further, it will explore how this NVMe device can be virtualized for both storage and video in an edge compute environment.

Learning Objectives: 1) Understand how NVMe can be used for both video and storage; 2) Understand how computational storage can be virtualized using NVMe; 3) Understand why combinational element modules such as Video Storage will become important after deployment of 5G networks.

By now, you have a good understanding of SNIA Swordfish™ and how it extends the DMTF Redfish® specification to manage storage equipment and services. Attend this presentation to learn what’s new and how the specification has evolved since last year. The speaker will share the latest updates ranging from details of features and profiles to new vendor-requested functionality that’s supporting the specification from direct-attached to NVMe. You won’t want to miss this opportunity to be brought up-to-speed.

Learning Objectives: 1) Educate the audience on what’s new with Swordfish; 2) Describe features and profiles and why they are useful; 3) Provide an overview of vendor-requested Swordfish functionality.

DMTF's Redfish® is a standard API designed to deliver simple and secure management for converged, hybrid IT and the Software Defined Data Center (SDDC). Both human readable and machine capable, Redfish leverages common Internet and web services standards to expose information directly to the modern tool chain. This presentation will provide an overview of Redfish, what’s new in the Redfish ecosystem, as well as adoption in the broader standards community. You’ll also learn more about the general Redfish data model, including the base storage models and infrastructure that are used by SNIA Swordfish extensions.

Learning Objectives: 1) Introduce the DMTF Redfish API; 2) Provide an update on the latest Redfish developments; 3) Understand how SNIA Swordfish builds on Redfish.

The SNIA NVMP TWG continues to make significant progress on defining the architecture for interfacing applications to PM. In this talk, we will focus on the important Remote Persistent Memory scenario, and how the NVMP TWG’s programming model applies. Application use of these interfaces, along with fabric support such as RDMA and platform extensions, are part of this, and the talk will describe how the larger ecosystem fits together to support PM as low-latency remote storage.

Whether you are a storage vendor, consumer, or developer, the performance of storage solutions affects you. Assessing the performance of large and complex storage solutions requires some level of performance testing lab, and there are many factors to consider. From network topology to load generator CPU, all components must be selected and configured with care to avoid unintended bottlenecks.

In this session, we will review a few best practices and lessons learned, including: whether virtual clients are feasible and my experiences attempting performance testing on several different hypervisors, best practices for network configuration, and how to use maximum effective data rates to avoid unintended bottlenecks. Finally, we will conclude with a review of data comparing different physical load generating hardware and its effect on measured performance.

Learning Objectives: 1) Effect of load generating client hardware on measured performance; 2) Avoiding unintended bottlenecks by using interconnect maximum effective data rates; 3) Best practices for configuring a performance lab network and load generators.

SMB 3.1.1 is the state of the art for secure remote file access, but deploying it for clouds and mobile users can be very challenging; TCP/445 is often blocked, networks are often slow, and edge file servers are often feared. The Microsoft SMB3 team has now built the first implementation of SMB3 over QUIC, a UDP/TLS transport pioneered by Google. This allows secure tunneling of SMB3 over internet-friendly ports. Furthermore, we have added compression for SMB3, which allows significant data savings over congested and low bandwidth networks. In this talk we’ll discuss these new options, as well as other recent security and feature capabilities nearing completion.

Learning Objectives: 1) SMB3 over new transport; 2) SMB3 over wide area networks; 3) SMB3 protocol update.

Attendees of this presentation will learn how to use several open source tools ( https://github.com/jharriga/ ) to evaluate object storage platforms. These tools provide automation and customer-based object storage workloads for activities such as filling a cluster, aging a cluster and running steady-state mixed operation workloads. One of the tools, RGWtest, automates pool creation, logs cluster statistics such as system resource utilization (CPU and memory) and submits workloads through COSbench - Intel’s open source object storage benchmark tool.

A demonstration of the tools will be part of the presentation.

Learning Objectives: 1) How to install, configure and execute the object storage workload tools; 2) How to interpret workload run results; 3) How to design and size object storage workloads.