Pure Storage’s fifth annual Pure Accelerate conference for customers, partners and industry influencers drew more than 3,000 attendees and was held September 16 – 19, 2019, in Austin, Texas. Evaluator Group attended, participating in a number of keynote and breakout sessions and engaging with various Pure Storage executives, customers and partners.
Pure Accelerate 2019 kicked off with CEO Charlie Giancarlo reflecting on the company’s ten years in existence. Pure gained market traction by accelerating and simplifying management of Tier One workloads for customers with its all-flash arrays and flexible application programming interface (API)-driven management. The company is retaining this focus while building into broader ambitions. Pure is building competencies to allow customers to consume multiple storage classes as hybrid cloud-based services – with the ultimate goal of allowing customers to consume all applications, storage protocols and clouds as-a-service. The company describes its vision as enabling a “modern data experience,” characterized by a common data fabric that is:
- Simple, in being API-defined, consumed as-a-service, and driven by artificial intelligence (AI)-based automation.
- Seamless, in supporting multiple storage classes and protocols and clouds.
- Sustainable, in being consumed as-a-service with continuous non-disruptive upgrades.
Pure’s vision has facilitated a product roadmap and announcements at Accelerate 2019 that are focused on consumption-based storage delivery, bringing more sophisticated support for cloud-based and containerized environments, and facilitating storage analytics.
Technology Announcements – FlashArray
Pure’s product enhancements and launches at Accelerate 2019 included the extension of its flagship storage array portfolio, FlashArray, to accommodate a broader range of workloads. Pure’s objective is twofold – to displace hybrid and all-spinning disk arrays for capacity-oriented workloads, and to displace in-memory computing for highly latency-sensitive workloads. Pure is building a portfolio that allows customers to create multiple tiers of flash in their on-premises storage environments to balance cost and performance requirements.
- For capacity-oriented workloads, Pure introduced its new FlashArray//C, an all-NVMe array that was designed to use lower-priced and denser QLC NAND memory (FlashArray//C will ship with TLC that performs with QLC characteristics until being replaced with QLC some time in 2020). According to Pure, the device provides milliseconds of latency alongside 99.9999% availability and an average data reduction ratio of 5.1 average data reduction, storing up to 5.2 petabytes (PB) of effective capacity in three nine-rack units. FlashArray//C uses the same controller and enclosures as the faster-performing FlashArray//X offerings. Target workloads include virtual machine (VM) tiering, disaster recovery and test and development.
- Serving the other end of the spectrum are Pure Storage’s new DirectMemory Modules. These are Intel Optane storage class memory (SCM) modules that plug directly into FlashArray//X70 and //X90 arrays, serving as a read cache that accelerates performance. According to Pure, DirectMemory Modules can reduce latency by 25%-50%, achieving 100 microseconds of latency compared to the FlashArray//X’s base capability of 250 microseconds, for up to 2x performance acceleration for latency-sensitive applications such as SAP HANA. At the same time, costs are reduced through a consolidated infrastructure footprint, including the DRAM, servers and software required.
In addition to its technology announcements, Pure announced that all of its products are now offered as-a-service through its Pure-as-a-Service program, a re-branding of its Evergreen Storage Service (ES2). It is relevant to note that Pure’s Evergreen initiative, which allows customers to upgrade their storage controllers once every three years, still exists. The gamut of Pure’s FlashArray and FlashBlade scale-out file and object storage devices, including the new FlashArray//C devices, are Evergreen.
Technology Announcements – Hybrid and Multi-Cloud
In addition to core storage array updates, Accelerate 2019 saw the expansion of Pure’s Cloud DataServices portfolio. Pure launched its new Cloud Block Store (CBS) service, which is its Purity Operating Environment software re-architected to run in Amazon Web Services (AWS) public cloud (whether native or deployed on VMware Cloud on AWS; future support for Microsoft Azure and Google Cloud Platform were alluded to), with the objective of improving resiliency (e.g. introducing high availability and replication) and bringing enterprise storage services to AWS storage implementations. At the same time, because the APIs and interfaces are consistent, CBS facilitates bi-directional mobility, for data access, migration and management in hybrid on-premises and AWS environments.
Alongside CBS, Pure also expanded its CloudSnap capability, previously exclusive to AWS, to support Azure. CloudSnap is a functionality built in to Purity that handles data transfer to and from on-premises and cloud environments, for backup and recovery use cases (customers can recover either to the on-premises Pure array, or via CBS to the cloud, depending on their performance service level agreements (SLAs) and their compliance and security requirements. It allows entire arrays and volumes to be recovered.
Technology Announcements – Storage Analytics
As a function of its ongoing maturity, Pure is continues to build telemetry data analytics capabilities into Pure1, largely through its Pure1 Meta cloud-based predictive artificial intelligence (AI) engine. Pure1 Meta powers a number of services intended to simplify management and support, improve uptime, and improve resource planning across multi-cloud storage environments. Launched at Accelerate 2019 was the new Pure1 VM Analytics Pro service, which extends data retention from seven days to three years and performs functions such as checking for and identifying root causes of anomalies.
Technology Announcements – AI Data Hub and AIRI-as-a-Service
As part of its effort to continue expanding its market traction, Pure has been increasing emphasis on artificial intelligence (AI) workloads over the past 18 months, primarily with a focus on helping customers to scale and to bring into mainstream production their AI workloads more easily. Initially launched in March 2018, Pure’s AI-Ready Infrastructure (AIRI) converged platform developed jointly with NVIDIA and based on FlashBlade, is now available as-a-service through managed service and colocation provider partners. Target customers are those that are encountering cost and performance challenges deploying AI applications at scale in the public cloud, but that do not have the ability to deploy and manage AI infrastructure in-house. Pure additionally unveiled its AI Data Hub, a FlashBlade-based and Kubernetes-orchestrated solution designed to streamline deployment of AI applications.
Evaluator Group Perspective
Accelerate 2019 served as both a look back at Pure’s ten years in operation, and a sneak peek at what the next ten years might look like for the company and for its customers. Pure’s DirectMemory Modules and its ongoing emphasis on its Evergreen initiative reflect an ongoing fit in bringing simplicity and cost-efficient performance to Tier One workloads.
At the same time, Pure is developing into a suitable fit for other workloads as well, supported by FlashArray//C and its maturing ability to tier data and to facilitate data mobility across various flash storage types and public cloud storage services. Extending its foothold in existing environments and continuing to attract new customers for Pure will be predicated on its ability to convince customers of the value proposition that it brings to cloud-native applications, and to more effectively manage heterogeneous (multi-cloud, multi-protocol, etc.) storage environments.
Download the free report now!