How Smart Caching Lets You Run Business at the Speed of Memory

How Smart Caching Lets You Run Business at the Speed of Memory

Enterprises are in a never-ending quest to make their applications, i.e. their business, run faster. The rising expectations of human nature condition us to expect instantaneous response times. After all, faster IT services ultimately improve worker productivity, enhance customer service, and increase revenue.

To seek out additional performance, IT managers seek faster CPUs, add more memory, re-architect application workflows, etc. But chances are, their old storage systems probably deserve most of the blame.

Wouldn’t it be nice if your storage performed at near the speed of memory?

Enter Persistent Memory

Persistent memory such as NAND flash, 3D XPoint™, and other Storage Class Memory (SCM) are fundamentally changing the storage industry. These are non-volatile solid-state memory that can retain data for a sustained period without refreshing the data in the device.

Persistent memory comes in various form factors, including DIMM (e.g., NVDIMMs) and SSD (e.g., NAND flash, 3D XPoint). Just take a look at our portfolio of flash-based devices, platforms, and systems to get a feel for the depth and breadth of our persistent memory solutions.  Of course, these solutions deliver various levels of performance, endurance, capacity, etc. at different price points. Since customers’ needs differ, storage cannot be effectively delivered by a one-size-fits-all approach.

Taking Advantage of Multiple Grades and Form Factors within a Single System

Western Digital operates on a Silicon-to-Systems design approach. Since we manufacture the basic components of our flash solutions and design the form factors in which it is placed, our deep knowledge of flash enables us to crate optimized solutions at the die, component, platform, and system level. We combine DRAM, NVDIMMs, NVMe™ flash, and other high-density storage media in our IntelliFlash™ arrays. This combination of storage media enables IntelliFlash arrays to deliver near memory-speed performance to our customers and with much better economics.

Within each of our arrays, there’s a performance tier and a capacity tier. In the performance tier we deploy DRAM, NVDIMMs, and NAND flash. This provides the highest performance at low latency for mission-critical applications. The capacity tier takes full advantage of the evolution in media densities. In this tier we use either NVMe-flash (IntelliFlash N-Series), performance SAS-flash (T-Series all-flash), dense flash (HD-Series), or a combination of flash and HDDs (T-Series hybrid flash).

As a result, customers can select a combination of storage media to help align the expense of storing data with its business value.

But this flexibility doesn’t come at the loss of performance. “How can this be?”, you might ask. Let’s take a look.

A Little Smart Caching Goes a Long Way

The performance tier is logically divided into three sections: A Write Cache, Read Cache, and Metadata. Let’s dive deeper into what each of these sections does:

1. Write Cache

Since IntelliFlash arrays feature a combination of DRAM, NVDIMM, and NAND flash in the performance tier, they can process a write I/O extremely fast. The moment I/O hits memory it’s immediately synchronized to NVDIMM in the Write Cache and an acknowledgement is sent from the controller to the client. This process takes 100s of microseconds, which is a significant performance gain over storage systems that send an acknowledgement when the data hits NAND flash.

Meanwhile, the data is processed in memory, calculating a checksum, compressing when advantageous, checking dedupe pointers, and so on. Then the data is coalesced and streamed sequentially to the capacity tier.

To ensure availability, the NVDIMMs are mirrored across controllers, and in the rare occurrence of a controller failure, the I/O is sent directly to the NAND flash.

2. Read Cache

There are two levels of Read Cache in an IntelliFlash array—both are self-tuning. The first level is the primary DRAM cache. The second level resides on NAND flash. The array proactively fetches and populates the Read Cache with the hottest data. This all happens dynamically in real time, using intelligent, pre-fetch algorithms—without user intervention. In most customer cases, the cache-hit ratio is well over 90%. That means that most read requests are served from DRAM, NVMe-flash or high-performance flash, even if the underlying storage is dense flash or HDDs. The most popular data is available from DRAM with the balance of the cache in high-performing flash. Only truly cold data is retrieved from the capacity tier, which could be any media from NVMe-flash down to HDD, depending upon the IntelliFlash model.

3. Metadata Acceleration

IntelliFlash arrays dedicate a portion of the performance tier exclusively for metadata management. This includes block pointers, dedupe pointers, compression type, and the like. All the metadata is organized, aggregated, and remains in the performance tier. This performance-optimization technique stands in stark contrast to traditional storage systems that intersperse metadata with the rest of the data on disk. Over time, as data inevitably gets modified, deleted, and rewritten, the metadata becomes very fragmented, which negatively impacts performance. Metadata placement in the performance tier dramatically improves the lookup of any data regardless of the underlying media.

ONE Platform, Multiple Grades of Storage

IntelliFlash arrays are designed so we can essentially add any type of storage media to the performance and capacity tiers through NVMe interfaces. We talked about using NVDIMMs and NVMe-flash today. In the future, our design would permit incorporation of other Storage Class Memories as they become available.

Run Your Business at the Speed of Memory

Would you like to dramatically accelerate all your applications to achieve substantial gains in business agility and competitiveness? I would think so. So why not take a leadership position in your marketplace through unparalleled business acceleration with economics that are unmatched by other storage vendors?

Discover how our portfolio of IntelliFlash arrays with PCIe-connected non-volatile, flash memory for data storage coupled with a highly optimized storage operating system (IntelliFlash OE) deliver very high performance at low latency for your data center workloads.

I invite you so See how Western Digital makes ONE Flash Platform, ANY Workload a reality so you can Make Your Data Come Alive and Run Your Business at the Speed of Memory.

Learn more about our flash and hybrid arrays designed to deliver extreme performance and exceptional economics for a wide range of demanding workloads by visiting our Storage Systems page on now.

Clay Ryder: Clay has 25+ years in the enterprise storage industry, driving business via strategic marketing, product positioning and analyst relations.

Related Posts

5 Reasons it’s Time to Consider NVMe-oF™ Today
5 Reasons it’s Time to Consider NVMe-oF™ Today

End-to-end NVMe is likely in your future, but when and how do you take the next step? Here are five reasons to start the transition now.‌

NVMe™ Spec Ratification and New ZNS Milestones
NVMe™ Spec Ratification and New ZNS Milestones

ZNS SSDs and Zoned Storage have just achieved several significant milestones, including NVMe spec ratification. Here's a summary for data center architects and developers