2018 SDC EMEA Abstracts

Main Stage Abstracts

Transforming Storage with Innovations in Non-Volatile Memory

Bill Bollengier, Director of Application Engineering, Non-Volatile Memory Solutions Group (NSG)


The smart and connected world is driving a massive flood of data – and the need to manage it.  Intel is evolving to address this by moving towards a world where the boundary between digital and physical is eroding, computing is truly mobile and ubiquitous, and everything is smart and connected.  This presentation will review Intel’s strategy and focus on its transformative storage innovations, including Intel® Optane™ technology and Intel® 3D NAND, that reside at its core, as well as the opportunities presented for OEMs and ISVs.

Enhancing NVMe-oF Capabilities using Storage Abstraction

Yaron Klein, System and Architecture Leader and
Verly Gafni-Hoek, Inbound Product Manager


NVMe protocol is optimized for NAND media, providing high performance in comparison with legacy protocols, such as SAS and SATA. Furthermore, it replaces the conventional volumes with namespaces and subsystems.

NVMe-oF protocol enables sharing the NVMe resources within the rack and across racks in the datacenter, while maintaining the benefits of NVMe.

In our talk, we will describe how NVMe SW abstraction further increases the potential of NVMe-oF. The abstraction enhances the ability to share storage resources. This increases the storage utilization and enables balancing workloads across a pool of SSDs. We will discuss different abstraction configurations, their advantages and disadvantages.

Learning Objectives

  • NVMe-oF Subsystem building blocks: review of a typical NVMe-oF system architecture; subsystems, controllers and namespaces terminology
  • What is SW storage abstraction and what are its advantages
  • Different abstraction configurations for virtualized NVMe Subsystems, their advantages and disadvantages

Paving the way for All-Flash with new SSD technologies

Kristian Vatto, SSD Product Manager


Samsung will explain the data flood problem that is facing the IT industry and how existing technologies and architectures are inefficient for modern data centers and edge computing. 

Object Storage: Storage for Developers

Dr. Michael Factor, IBM Fellow, Storage and Systems, IBM Haifa Research Lab


Machine Learning, Video Processing, Analytics, and IoT all run on data, and typically, they require massive amounts of data. This data is needed by data scientists, line of business, developers and others. Object storage, which is the default storage for the PBs and EBs of unstructured data in clouds, has brought huge efficiencies to these application-centric use cases. In this presentation, we will provide an overview of object storage, using IBM's Cloud Object Storage (formerly Cleversafe) as an example. In particular, we will describe the differences between object storage and more traditional big data storage solutions from the perspective of the developer. We will then discuss how developers can take advantage of object storage with modern applications via integration with software such as Apache Spark and Kafka.

Track 1 Abstracts

Controller Reliability Challenges in Emerging NAND Memory Technologies

Hanan Weingarten, Head of Haifa Site, Toshiba Israel Storage Design Center


The requirement for ever denser memories is answered by the introduction of new NAND memory technologies. New multi-layer three dimensional stacking (BiCS4) as well as four bits per cell (QLC) technologies are introduced by the NAND memory FAB companies. The introduction of these technologies brings into bear reliability challenges which are answered with new and unique methods.

Any memory controller must contend with the variability between the layers in 3D manufacturing technologies as well as the high error rates and stress condition sensitivity introduced by QLC. We answer these challenges with new ECC, DSP and memory management schemes. The high performance, low power and price requirements from storage devices lead to unique solutions that are not available in standard communication systems.

In this talk we will introduce some of the challenges and will discuss the principles behind some of the solutions.

Estimating Compression and Deduplication ratios

Dr. Danny Harnik, Research Staff member, IBM Research Haifa


The talk will examine the following challenging question: how to predict what data reduction ratio one should expect when storing his data in a storage system that provides compression and deduplication? We will describe cutting edge results on how to perform these tasks and explain the obstacles in doing so efficiently. Specifically, the answer to this question is very different for compression and deduplication - the task is significantly harder for deduplication. The talk will describe solutions for both tasks that are based on sampling and give estimations rather than actual results.

What's New in NVM Express

Dr. J Metz, R&D Engineer, Office of the CTO, Cisco Systems


NVMe has taken the storage world by storm, and is arguably one of the hottest topics around. The NVM Express group, however, is not just resting on its laurels. With over 30 active projects, new features and capabilities are being developed on an ongoing basis. Some of these involve some important additions to how NVMe will work in both local, client, and enterprise environments.  This presentation will discuss some of the more prominent developments in NVMe.

Achieving Predictable Latency for Solid State Storage

Mark Carlson, Principal Engineer, Industry Standards, Toshiba


Your typical SSD is very badly behaved with regard to latency. Background tasks interfere with I/O requests. As drives increase in size beyond a typical working set, multiple users interfere with each other (noisy neighbor). Scale out systems such as those assembled by Hyperscalers, are not tolerant of today’s long tail of SSD latency.
In order to address these requirements, the storage industry through the NVM Express standard is adding some new features. This talk will explain those features and how they address these issues.

Accelerating Storage with RDMA - Leveraging RDMA for higher performance File, Block and Object Storage

Max Gurtovoy and Idan Burstein, Storage Architects, Mellanox


File, Block and Object storage is able to take advantage of current NAND flash to get better performance. But much more performance is possible as RDMA based storage technology originally developed for the HPC industry moves to the main stream. By enhancing a storage system’s network stack with RDMA users can see an even more dramatic improvement than by just adding flash to their storage. The technology increases the performance of the entire storage system allowing File, Block and Object based applications to take more advantage of much higher performance solid state storage. With even faster Persistent Memory on the way, RDMA is even more important to eliminate the network stack bottleneck. This session will explain how this technology utilizes ultra-low latency network interfaces to achieve this.

Accelerating Remote Virtual Machine Access with the Storage Performance Development Kit (SPDK) - Vhost Deep Dive

Anu Rao, Product Line Manager, Data Center Group, Intel® Corporation


Although new non-volatile media inherently offers very low latency, remote access using protocols such as NVMe-oF and presenting the data to VMs via virtualized interfaces such as virtio adds considerable software overhead.  One way to reduce the overhead is to use the Storage Performance Development Kit (SPDK), an open-source software project that provides building blocks for scalable and efficient storage applications with breakthrough performance.  Comparing the software paths for virtualizing block storage I/O illustrates the advantages of the SPDK-based approach.  Empirical data shows that using SPDK can improve CPU efficiency by up to 10 x and reduce latency up to 50% over existing methods.  Future enhancements for SPDK will make its advantages even greater.

Persistent Memory - What developers need to know

Mark Carlson, Co-Chair, SNIA Technical Council


Persistent Memory (PM) is here now. System architectures are incorporating PM and Operating Systems, Hypervisors and other system software have been altered to accommodate this new, much faster storage. What are the implications for developers? This session will detail the new programming model and detail both what is available now as well as what can be expected in the near future.

Track 2 Abstracts

Intel® Optane™ Technology – Use Cases for Enterprise and Cloud Storage

Behnam Eliyahu, EMEA AE Lead, Non-Volatile Memory Solutions Group (NSG), Intel® Corporation


Intel® Optane™ is the world’s most responsive data-center SSD based on 3D XPoint™ technology. It brings an industry leading combination of low latency, outstanding QoS, and high endurance, delivering performance needed for both memory and storage workloads. In storage applications, Intel® Optane™ SSD allows breaking storage, network and caching bottlenecks, thus it enables new use cases and architectural models which bring faster performance and increasing scalability. The session will cover various use cases, for example: persistent memory, write buffer, scale out storage and networking acceleration (NVMeoF). In addition, we will demonstrate how storage reliant applications can benefit from Optane™ SSD, such as: data base workloads, hyper converged and software defined storage. 

Techniques for Maximizing Shingled Disks, A View from Inside

Andrew Kowles, Principal Engineer, Global Firmware, Seagate Technology


For too long, SMR disks have remained a mystery to everyone but HDD makers.  SMR HDDs have unrealized potential because of the veil of secrecy. This presentation will attempt to lift the veil from one major HDD vendor's perspective with a goal of providing needed clues to storage developers for managing and maximizing shingled disks. Techniques introduced will include: Understanding Zoned Block Devices (HM-SMR) as well as Flex/Hybrid CMR/SMR disks and their potential, internal management concepts for Drive Managed (DM-SMR) disks and use case examples. 

Enabling Tape In The Storage Stack Using Open-Source Software

Dr. Ioannis Koltsidas, Researcher/Engineer, IBM


Magnetic tape has some very attractive properties including low cost, energy efficiency, long-term reliability and security, that make it particularly well-suited to storing the vast amount of data the modern world generates as it becomes cold. However, tape systems have inherent complexity in terms of management and operation. In the enterprise world these complexities are typically hidden behind proprietary, monolithic software and data formats; these are not a great fit for open-source based storage systems and cloud architectures.

In this presentation, we discuss how developers can build a storage stack with a tape backend that relies on open formats only and uses no proprietary software. We present new open-source software that has been designed to hide the complexity of tape from the storage developer. They build on the standardised Linear Tape File System (LTFS) and provide the capability to integrate a tape backend at different levels of the storage stack (e.g., as an extension to a disk filesystem or as a cold tier for an object storage system)  providing ease-of-use without giving up flexibility and without dependency to specific tape technologies. We also discuss how frameworks like Kubernetes can be used to orchestrate tape resources.  The main goal of the presentation is to familiarise the audience with the new software projects, explain how they can be used, and invite developers to participate with their own ideas and requirements to shape them.

Where Network meets Storage. Three ways of integrating Network Protocol into a Storage platform

Mark Rabinovich, CTO, Visuality Systems


When integrating a Network Protocol into Storage one of the key questions is whether the Network component or the Storage platform is responsible for file system semantics. This question applies to the semantics in general and to the cross-node replication in particular.  As usual the truth must be somewhere between the two sides. In this presentation we will discussvarious options exploring the pros and cons of each of them.

ZUFS - How to Develop an Efficient PM-based File System in User Space

Dr. Amit Golander, Technical Director, NetApp


Persistent-Memory (PM) based file systems are key to bringing the low-latency value of PM hardware to applications and thus to businesses. Developing a POSIX-compliant file system in Kernel space is expensive, but developing it in user space using the traditional Linux FUSE bridge results in high-latency access, and thus defeats the purpose. To bridge this gap we have developed the ZUFS bridge. ZUFS was designed with PM latencies and byte addressability in mind. It enables the development of a new breed of file systems that will be both efficient and in user space. ZUFS is an open source project, so users and contributors are welcome to join.

Optimizing Storage Architecture for Modern Data Analytics

Seongyong Lee, SSD Product Planning Engineer, Samsung Electronics


In this presentation, Samsung will explain the concepts of NGSFF, Z-SSD and KV-SSD and how this technologies can be implemented to optimize storage architecture for modern data storage and analytics requirements.

Developments in Swordfish

Alex McDonald, SNIA Technical Council


Swordfish™ is an extension of the DMTF Redfish specification developed by the Storage Networking Industry Association (SNIA) to provide a unified approach for the management of storage equipment and services in converged, hyper-converged, hyperscale and cloud infrastructure environments, making it easier for IT administrators and DevOps to integrate scalable solutions into their data centers. In this presentation, we’ll show how Swordfish extends Redfish and provides an overview of basic Swordfish storage management concepts, along with why you should use it as a management interface in your products.