Change is inevitable but, for the IT industry in recent years, it’s been nothing short of significant and disruptive. The era of big data is providing enterprises significant opportunities to make better-informed business decision by leveraging analytics, artificial intelligence, and machine learning technologies.
Businesses who take advantage of this to respond quickly to evolving market demands are finding success, as long as their underlying IT infrastructures are transformed into something agile enough to keep up. This agility often means a significant departure from legacy storage and movement towards all-flash arrays, software-defined storage, and converged infrastructure. The availability of cloud-based services in particular is driving change: creating expectations about deployment, making it easier to securely share data across different constituencies, and introducing IT to new ways of thinking about easy scalability, frequent and nondisruptive upgrades, and technology refresh for IT infrastructure in general.
But there’s A downside
These high-end enterprise storage system scale to millions of IOPS and petabytes of capacity, deliver “six-nines” of availability, and offer a full complement of storage management or “data” services that enable dense mixed workload consolidation while still meeting individual application requirements. But despite the advantages, there are still a number of challenges to overcome:
- Don’t address application service availability. Even if a storage system is 100 percent available, that doesn’t necessarily translate into application service availability. Storage causes 9.8 percent of the downtime in a typical datacenter, which means IT needs visibility into the other layers in the stack that deliver the application services and, when administrators use multiple tools to accurate diagnose and remediate these issues, they present more complex challenges and take longer to resolve.
- Designed for experts. The other side of the management flexibility coin is complexity. These systems often require very sophisticated storage management expertise for installation, configuring, and managing them to meet business objectives. Unfortunately, with storage management tasks moving away from dedicated administration teams and toward IT generalists, the knowledge can be lacking.
- Require constant tuning to address “noisy neighbor” concerns. Application changes, which happen on a regular basis, raise concerns about the ability to predictably meet performance requirements across a set of mixed workloads. If systems do not offer sufficient capabilities to meet service-level agreements, then manual tuning is required to rebalance the system configuration – a time consuming process that requires sophisticated knowledge many administrators managing storage today don’t have.
- Impose disruptive, time-consuming, or risky upgrades. Technology upgrades to these systems can be disruptive, time consuming and fraught with risk. Moving to the next technology generation still generally requires a forklift upgrade, manual data migration, and incremental upgrades to firmware. Pre-validating and forecasting the “impact” of these types of upgrades depends on human involvement – both of which add time and risk.
- Have outdated technology “baggage.” System architectures originally designed during the hard disk drive era will not necessarily support the massive parallelization necessary to optimize the performance and capacity utilization of the solid state media that has already come to dominate high-end system shipments, and this factor can hinder workload consolidation and limit a system’s ability to deliver predictably consistent performance.
A new way to look at high-end storage
Keeping in mind that agility is the key requirement for today’s digitally transforming enterprises, the design of high-end storage systems needs to change. The more purpose-built, statically defined scale-up designs of the past need to give way to more software-defined, modular scale-out designs. Systems that were originally designed and optimized assuming the use of HDDs need to be optimized to take advantage of the latest solid state storage technologies (NVMe and NVMe over Fabrics).
These systems need to leverage AI/ML technologies, in combination with big data analytics and automation, to improve and simplify routine management tasks at the scale of today's computing environments. They also need to be designed to operate in hybrid cloud environments, since this is clearly the blueprint for how IT infrastructure will be deployed going forward. And they need to offer all this while continuing to deliver the rich storage management functionality that enables secure, multitenant management from a proven storage OS that can scale to millions of IOPS, petabytes of capacity, and "six-nines plus" availability.
Reset the definition of high-end enterprise storage With HPE Primera
HPE offers significant choices to its customers in terms of deployment models for storage, delivering storage appliances; software-only, hyperconverged infrastructure (HCI); converged infrastructure; and cloud-based services in a storage portfolio that includes primary and secondary storage; block-, file-, and object-based platforms; scale-up and scale-out architectures; and the industry's most mature cloud-based predictive analytics platform (InfoSight).
In introducing its new high-end storage system, HPE Primera, HPE is melding the simplicity of Nimble Storage, the mission-critical heritage of 3PAR, and the intelligence of InfoSight into a single high-performance, highly available, and highly scalable large enterprise storage solution. HPE has designed a new system architecture around a number of new features and technologies to provide a platform that delivers management simplicity without compromising its ability to support an enterprise’s most mission-critical workloads:
- Scale-out architecture to support massive parallelism. As a storage protocol technology, NVMe supports at least three orders of magnitude higher parallelism that SCSI. HPE Primera embraces this with a design that his highly optimized for solid state storage.
- A modular, service-centric design for the storage OS. The HPE Primera storage OS is not a monolithic software platform. It is instead built around a service-centric design that allows any feature to be added or modified without requiring a recompile of the entire OS.
- AI/ML-driven self-optimizing systems management. HPE Primera features very comprehensive and granular monitoring, with all the collected data immediately uploaded to HPE InfoSight. HPE InfoSight uses AI/ML to evaluate system performance, capacity utilization, energy consumption, and other metrics and combines that data with data collected from servers, networks, virtual machines, storage infrastructure software, applications, and cloud storage to generate a global view of a system or workload state at any point in time. This global view provides a better overall opportunity to optimize to specific objectives and/or troubleshoot infrastructure issues.
- Streamlined multinode hardware design. Internode communication requires no separate cables — all nodes plug into a passive backplane for much simpler configuration and higher reliability. The HPE Primera node-level building block is a high-density design that can house up to 1PB of effective capacity in 2U (or 2PB in 4U), with additional external storage capacity expansion available in both form factors. This building block node is designed as the hardware foundation for storage systems that can deliver 100% availability. In fact, as part of the HPE Primera announcement, HPE introduced a 100% availability guarantee on every single Primera system that does not require stretch cluster configurations or even a maintenance contract.
Ready for high-end enterprise storage? Give WEI a call.
HPE is the first established enterprise storage vendor to leverage these technologies to this extent, and has done so in a way that will set a new customer experience bar for high-end storage. Their design approach will drive significant value for customers, and more enterprise storage vendors are expected to move in the same direction with their solutions over the next few years. If you’re ready to make the change now, contact WEI to begin building a custom solution for your business.
