DPDK is an optimised set of user libs for fast packet processing. We’ve added a generic compressdev API for compression. This presentation will describe how to integrate the compressdev service with a storage application, using SPDK as an example. It will show how the application can target either software or hardware accelerators without a change in application code. The examples used will be ISA-L, an optimised software library for storage and QuickAssist, an Intel hardware accelerator.

This booklet contains a collection of posts from Jim Handy's SSD Guy blog.

Using NVMe drives in a centralized manner introduces the need for high availability. Without it, a simple failure in the NVMe enclosure will result in loss of access to a big group of disks. Loss of a single NVMe disk will impact all hosts mapped to this disk. We will review the state of the industry in approaching these problems, the challenges in performing HA and RAID at the speeds and latency of NVMe, and introduce new products in this space.

Ceph is a distributed object store and file system designed to provide excellent performance, reliability and scalability. BTRFS with the compelling set of features is recommended for non-production Ceph environments.

People today are generating ever increasing amounts of data, with much of it is being stored in megascale datacenters such as those run by Microsoft Azure. Although freshly generated data is hot, over time it cools and ultimately many of the bytes stored are cold. Storing this data in a hot tier is expensive, so cloud storage systems need to trade performance for cost, and they currently do this by using tiers backed by different storage media. Ideally, we’d have a bedrock tier which would have close-to-zero cost to store data, while durably storing the data essentially forever.

While the “perfect” workload to test an array for a customer is the workload or combination that will run against the array in production, enterprises and vendors benefit greatly from a workload approach that more accurately models a production workload to be executed against an array. Enterprises benefit by seeing the characteristics of their own workload or combination, while vendors benefit by being able to scale emulated workloads to array limits and better able to deal with performance issues that occur only after an array is released.

The first storage testing session, “Getting it Right, Testing Storage Arrays the Way They’ll be Used”, discusses a best-practices methodology for testing modern solid state storage arrays. It stresses the importance of emulating real-world workloads and scenarios—including temporal and spatial locality, micro bursts, realistic block size combinations and data content. By emulating these characteristics, testers can do a better job of sizing arrays for potential applications and understanding the performance characteristics of these arrays.

The public release of the Gen-Z Core Specification 1.0 enables silicon providers and IP developers to begin the development of products enabling Gen-Z technology solutions. This presentation outlines how Gen-Z will address the multiple server challenges brought on by the explosion of data (180ZB annually by 2025) and the need for a high-speed, low-latency, scalable, memory-centric fabric. Customers have more data and require more processing of that data, and to be competitive, companies must monetize that data.

A Gen-Z supported system definitely will have Memory Management Unit (MMU). Processor MMU will be attached with DDR protocol engine and Gen-Z protocol engine. Further Gen-Z protocol can retrieve data from DRAM, SCM, DRAM + SCM or DRAM + Flash devices. A media controller supporting different media types and mix illustrates a variety of Gen-Z memory modules that support any type and mix of media types including DRAM, SCM, and Flash. Media is abstracted through the Component Media Structure which describes the memory component’s attributes and capabilities, and controls its operation.

High-performance SSDs with NVM Express* (NVMe) enable you to shrink your footprint, increase your storage density, support data-intensive workloads and enable storage solutions which are easier to deploy and manage. All flash software defined storage or hyper-converged solutions bring the best business value when coupled with NVMe. In this discussion we will present different ways to approach NVME through NVME over Fabrics in the data center. Join us for a collaborative discussion on the benefits of NVMe and how 2016 is a transformative year for storage.