A NEW STANDARD IN HIGH-PERFORMING STORAGE FOR FEDERAL AGENCIES
Technology is advancing at breakneck speed. For many federal agencies, these advances have yielded tremendous new transformational capabilities, including big data analytics, artificial intelligence and machine learning, to assist them with especially large, complex, or fast workload requirements. But as agencies transform their operations with more powerful applications and computing resources, they are pushing their data storage capabilities well beyond their limits.
Federal data centers typically employ a mix of traditional hard disc drive (HDD) and solid state drive (SSD) technologies for their enterprise storage needs — but their limitations in speed and capacity are creating performance bottlenecks for many agencies. This is understandable for HDD storage, which, after all, is older technology and far slower than SSD technology. But why would federal data centers encounter bottleneck problems with the newer, faster SSD storage (also known as flash) as well?
The answer lies in the technology used to move data back and forth between the storage media and the application. Many federal agencies have modernized their data centers with flash storage, which promises performance metrics many times better than spinning disk media. But in their transition to SSD, data centers have continued to access their flash storage arrays through interfaces that have been historically used with slower HDDs. Those interface protocols are called SAS (Serial Attached Small Computer System Interface) and SATA (Serial Advanced Technology Attachment). And when fast flash-based storage relies on legacy data transport technologies, the result is bottlenecking.
In recent years, a non-profit consortium backed by the leading storage players in the market — Dell EMC, Cisco, NetApp, Seagate and Western Digital, among others — called NVM Express, Inc., has developed a new interface protocol that is expressly built to efficiently leverage and optimize the power of flash storage. This interface protocol is called NVMe (Non-Volatile Memory Express). Unlike SAS and SATA storage devises, NVMe storage can fully accommodate the ridiculous speeds that SSDs are capable of.
This is great news for federal agencies with highly demanding or mission-critical workloads that have found traditional SSD storage underwhelming. The most impressive of today’s NVMe storage solutions for such workloads is the Dell EMC PowerMax, the world’s fastest storage array1. We at Wildflower have found that federal agencies modernize with the Dell EMC PowerMax for four key reasons:
Featuring end-to-end NVMe and a built-in machine learning engine for automated data placement, PowerMax delivers new levels of performance: up to 10 Million IOPS2, 150 GB per second bandwidth3, and predictable response times of 290 microseconds4 for mixed workloads. It also employs data-at-rest encryption, massive scalability, and best-in-class data protection, including Symmetrix Remote Data Facility (SRDF), the gold standard in remote replication. PowerMax represents true modern storage designed for the mission-critical applications of today and tomorrow, including databases and applications as well as real-time analytics that demand uncompromising uptime and extremely low latency.
One of the striking features of PowerMax is its queue depth, which refers to the number of operations it can handle. SATA flash drives work on a single command queue that handles 32 commands. By comparison, PowerMax is able to leverage NVMe’s robust ability to operate 64,000 command queues in parallel, with each queue holding 64,000 commands. This contributes to PowerMax’s exceptionally high IOPS performance.
PowerMax also comes with stringent security features to ensure agencies can meet governance and compliance requirements and work across all available data services so that no tradeoffs need to be made. PowerMax data-at-rest encryption (ready for FIPS 140-2 validation) secures every drive and delivers integration with external key managers, enabling customers to simplify security through a centralized key management platform. Tamper-proof audit logs allow IT managers to quickly identify unwanted activity and feel confident in the accuracy of the original PowerMax logs.
Moreover, PowerMax can scale to meet most federal agency needs: Its multi-controller scale-out architecture can accommodate up to 4 PB of effective capacity.
Designed for six nines of availability (or 99.9999% of the time), PowerMax sets the standard for uptime. This incredible standard translates to a mere 31.56 seconds of downtime in a year. Most service level agreements (SLAs) today offer four nines of availability (more than 52 minutes of downtime in a year). It achieves this by relying on, not one, but eight performance controllers, which are embedded processors that perform operations related to reading and writing data. If one controller somehow goes bad, there are seven others available to take over. PowerMax also offers no single points of hardware failure and hot swappable components to extend availability
A future-proof platform
PowerMax can work with any storage technology, whether spinning discs, solid state drives, NVMe, and even storage-class memory (SCM), also known as persistent memory, which may be the most disruptive storage technology innovation of the next decade. SCM has the potential to be even more disruptive than flash, both from a performance perspective — it is orders of magnitude faster than flash — and in the way it will change both storage and application architectures.
In addition, PowerMax is part of the Dell EMC Future-Proof Loyalty Program, which provides investment protection through a set of world-class technology capabilities and programs that enable Dell EMC storage products to provide value for the entire lifetime of customers’ applications. The unique program is available to customers at no additional cost in terms of maintenance or product prices.
There are numerous features contributing to PowerMax’s extremely efficient architecture and operation. These include inline deduplication and compression that delivers up to 5:1 data reduction (vice the 3:1 average) that have virtually zero impact on performance and can be used with all data services.
It also employs thin provisioning (TP), which optimizes the efficiency with which the available space is utilized in storage area networks (SAN). TP allocates disk storage space in a flexible manner among multiple users, based on the minimum space required by each user at any given time. This compares to conventional storage provisioning models in which storage space is allocated beyond current needs in anticipation of growing need and increased data complexity. This approach leads to low storage utilization rates where large amounts of storage space are paid for but may never be used. This also translates into reduced energy consumption, smaller hardware space requirements, and reduced heat generation.
The combined effect of these benefits makes PowerMax a top option for any federal agency grappling with extremely demanding or mission-critical workloads. Any agency considering virtualization, VDI, data analytics, multi-use, or large-scale (e.g. 100 terabytes or more) workloads, for example, will find PowerMax more than ably addresses its requirements.
Agencies are increasingly recognizing data as a strategic asset that must be more smartly leveraged — and that means they must pay more attention to where data resides, how it is managed, and how it is protected. Legacy storage platforms are increasingly incapable of fully meeting the robust requirements of modern workloads. By leveraging powerful and innovative technologies like NVMe storage, federal organizations can speed digital transformation as they strive to improve operational efficiencies and advance their missions with data-driven insights.
1 Based on Dell EMC internal analysis of published bandwidth for PowerMax 8000 versus competitive mainstream arrays, March 2018.
2 Based on Dell EMC internal analysis of max IO/s per second (within a single array) for the PowerMax 8000, March 2018.
3 Based on Dell EMC internal analysis of GBs per second (within a single array) for PowerMax 8000, March 2018.
4 Based on Dell EMC internal analysis of an OLTP heavy-write mixed workload for PowerMax, March 2018.