Six Observations about Today’s High End Storage Arrays

Enterprises investing in today’s high end storage arrays understand the value that these arrays offer in regard to their availability and performance as it can cost upwards of $5,000 for every minute that an application is offline. Applications and data must be available all of the time as any interruption in service can seriously impact a corporation’s revenue and reputation.

Selecting the appropriate high end storage array that offers the correct combination of features and functionality for an enterprise can mitigate the possibility of outages and the costs associated with them. This explains why high end storage arrays, even many years after their introduction, remain more than a popular choice to host today’s centralized, virtualized applications. They are, in essence, experiencing a rebirth of sorts.

However, choosing any high end storage array requires a substantial investment in both time and money to research and implement. Further, there are notable differences between each array that DCIG classifies as “high end.” This is why DCIG is producing a Buyer’s Guide on High End Storage Arrays that it anticipates releasing in the very near future.

In that vein, as DCIG has done research on these arrays, it has made the following six observations about them and the environments into which they are going into.

  • High application availability. Possibly the most desirable feature on these arrays that prompts so many enterprises to deploy them is their high availability (HA). Yet what differentiates them is that vendors employ various methodologies to deliver HA with options to scale up, scale out or both to deliver HA. At the most fundamental level, these arrays support multiple pairs of Active-Active controllers (also implemented as “blade pairs” or “processor pairs” on some of these arrays) on the same physical array that are all part of the same logical array configuration.
  • Large hardware capacities. Each high end array also has high capacities in regard to its cache, raw storage and processing. Over half of the arrays scale to support upwards of 3,000 GB (3 TBs) of cache, 67 percent of the arrays support at least 4,500TB (4.5 petabytes) of raw storage capacity and 60 percent of the arrays scale out to support at least 64 processor cores.
  • Multiple storage networking interfaces. All of the storage arrays covered in the forthcoming Buyer’s Guide have a minimum of 20 storage networking ports available while 80% had up to 64 networking ports. The interface types vary by vendor and product, but all of the arrays support 8Gb Fibre Channel (FC), 75 percent support 16Gb FC, 87 percent offer 10Gb Ethernet and one-third of the arrays support 8Gb FICON (used in mainframe environments.)
  • Robust VMware integration. Given VMware’s predominance in enterprise data centers today, it follows that they want storage arrays that can integrate with VMware vStorage APIs such as VAAI (vStorage APIs for Array Integration) and VASA (vSphere Storage APIs for Storage Awareness) to take advantage of the “force multiplication” that these APIs provide. The good news is that all of the high end storage arrays in this upcoming Buyer’s Guide support all of the VAAI v4 APIs and the majority of the options in VAAI v5. Similarly, all of the arrays support VASA.
  • OS and application performance monitoring. Managing any enterprise data center is challenging, but managing one without visibility into each application’s performance so that one can diagnose and even anticipate pro-actively anticipate problems can be foolhardy.

Using these performance monitoring and management tools, administrators can quickly pinpoint performance bottlenecks or what piece of hardware inside of the array is malfunctioning.All of the arrays provide some level of performance monitoring, with over 80 percent of the arrays providing physical drive monitoring while over 50 percent provide monitoring on a per application level. However, there is still some disparity in the ability of these arrays to monitor performance at the OS (Operating System) and VM (Virtual Machine) level so enterprises need to exercise some caution in which one they select as not all of these arrays may offer the full suite of software that they need to fully monitor and manage performance for all OSes and applications.

  • Automation. Data center automation is another growing area of emphasis for many enterprises as it facilitates efficient management of their data center infrastructure and more agile responses from IT to changing business requirements. Ultimately, automation means more staff time can be spent addressing business requirements rather than managing the routine tasks of a data center.

Currently, 71% of the high end arrays support policy-based storage selection. However, only 28% expose their APIs for third-party automation tools, while 42% provide an SDK for integration with management platforms. As more enterprises place a premium on automating their storage environment, look for these numbers to increase.

Chuck Cook

About Chuck Cook

Chuck Cook is an Analyst at DCIG, an independent storage analyst and consulting firm, which provides informed, insightful, third party analysis and commentary on IT hardware, software and services. Chuck joined DCIG in July of 2014 bringing 20 years of information technology and unified communications experience to the DCIG team.

Leave a Reply