Pure Storage has announced the FlashArray//C20, a new entry-level system aimed at smaller initial deployments of low-cost flash. This latest offering appears to be a response to the NetApp capacity solutions introduced in 2023.
Background
Pure Storage, Inc. has announced a new member of the FlashArray family, branded FlashArray//C20. This entry-level model delivers between 148-260TB of raw capacity in a 3U chassis, much like the equivalent products of //C50, //C70 and //C90 (side note: we’re not sure why there was a move from odd-numbered decimal digits and why this platform isn’t the //C30, but there you go). The FlashArray//C was first announced in 2019.
Configuration
Although not explicitly discussed in the press release, we believe the //C20 will use 18TB DFMs (Direct Flash Modules), probably similar to those that were announced in the launch of FlashBlade//S100 in October 2024. As such, the //C20 isn’t a new product but a new configuration option based on a specific storage capacity. In common with all FlashBlade and FlashArray products, the //C20 will be upgradeable to higher capacity models, turning the system into a //C50 or higher if that level of storage capacity is needed.
Competition
The //C20 is a straightforward product announcement, so why does it justify a Research Note? We can answer the question by looking at the announcement from NetApp in November 2023 of the C-Series, a new capacity-optimised platform that scales down as low as the entry-point 122TB ASA C250. The C-Series was first launched in February 2023 with the AFF range, while the ASA “All SAN Array” models were released later in the year.
For capacity flash, NetApp offers a lower entry point than Pure Storage, which, until the announcement of the //C20, was 2/3 of the capacity of a //C50. We see the //C20 as a direct response to NetApp’s capacity models, a process of “leap-frogging” we discussed in this research note.
The Architect’s View®
Both NetApp and Pure Storage clearly view each other as the leading competitors for Flash in the data centre. As such, we see competitive manoeuvring to ensure that neither vendor gives the other any market advantage by not having a competitive product in a similar space based on capacity and price.
Elsewhere, Dell now offers the 3200Q, which has an entry point of 169TB, while Hitachi Vantara recently announced a QLC model of the VSP One platform. HPE may have QLC-based systems and low entry-point capacity, but their website is too impenetrable to work out if this is true.
Should we care whether systems contain one NAND flash type over another? Despite living in a cloud world, hardware specifications are still important, arguably more important than ever, as IT organisations compare, for example, the capability of specific vendor GPUs.
QLC-based systems offer a better cost point in $/GB while maintaining more than adequate levels of performance. Building media-specific systems also means being able to avoid issues such as tiering, as we discuss in this recent Commentary article.
Where could this market go next? There must be a natural minimum capacity that makes sense for vendors to offer, based not on the NAND cost but on the other components of shared storage systems. When taking data redundancy requirements into consideration, the 100TB mark seems to be a reasonable absolute minimum capacity. As a result, we could already be at the lowest point in the cycle (although Pure Storage has left room in the nomenclature for a //C10 system).
The announcement of the //C20 signifies another step forward in the goal of an all-flash data centre, offering yet another lower entry point into capacity flash systems. Perhaps this is why Pure Storage believes there will be no new hard drives sold after 2028 – something the company is working hard on bringing to bear.
Copyright (c) 2007-2024 – Post #a1a2 – Brookend Limited, first published on https://www.architecting.it/blog, do not reproduce without permission.

