Sergey Nivens - Fotolia
What digital transformation means for data storage
We survey the key planks of digital transformation and look at the storage technologies emerging to support it, such as NVMe and QLC flash, hyper-converged infrastructure and cloud storage
Digital transformation initiatives have been accelerated by the Covid pandemic, as organisations have been forced to intensify digital channels to their workforce and customers.
But digital transformation is a strategic trend visible across all verticals, driven by the gains that can be made by more responsive, real-time and agile working and use of emerging technologies and methods of delivery.
And storage is a key component of digital transformation. It may not look the same as it once did – with the rise of the cloud and architecture-morphing high performance media – but the need to store and access data remains.
In this article we will look at the key components of digital transformation and the storage technologies emerging to support them.
Digital transformation basics: First off, let’s recap on digital transformation. While clearly it will differ in detailed terms for every organisation, there are some common themes.
Data-centricity: Core to the concept is the centrality of data. That means increased use of data to drive customer interaction and business processes, with a heavy emphasis on analytics, often in real time or near.
A sometimes overlooked element is a general shift to digitisation of an organisation’s assets and the ability to derive value from data that was previously consigned to (digitally, or possibly literally) gather dust in archives.
More on digital transformation
- The top five digital transformation trends of 2021. In 2021, low-code, MLOps, multi-cloud management and data streaming will drive business agility and speed companies along in their digital transformation journeys.
- Ultimate guide to digital transformation for enterprise leaders. This in-depth guide explains what digital transformation is, why it is important and how enterprises can successfully transition to this new business paradigm.
Agile development: At the same time, data and information processes are also expected to work to a much higher clock speed, with agile development aimed at rapidly taking advantage of business opportunity.
Rapid scaling: Allied with this is an increased need to rapidly scale, including via emerging application delivery methods such as containers.
The cloud: Background to all this is increasing use of the cloud, often for less critical applications, but also to satisfy the need for rapid scaling and to burst to cloud capacity.
Key storage technologies
Rapidity of I/O and heavyweight throughput are major requirements of the data-centric aspects of digital transformation. That means increased use of solid state media in several emerging and already mainstream forms.
NVMe and NVMe-over-fabrics: Prominent among these is NVMe flash, which allies NAND flash with connectivity designed for solid state storage, unlike SAS and SATA-connected drives. Most major storage suppliers offer all-NVMe arrays now, with many offering them at the same price point as other media. NVMe-over-fabrics (NVMf) – a method of extending NVMe transports across the network – is also becoming more prevalent as a means of connecting hosts to storage.
NVMe and NVMf are aimed at primary storage use cases with heavy I/O requirements, which can include the kind of real-time processing-heavy tasks such as analytics and rapidly-scaling demand scenarios.
Persistent memory and storage-class memory: While NVMe is aimed at primary storage, emerging classes of media straddle backend storage and memory, with the likes of (the not-NAND) Optane 3D Xpoint, which can be DDR or block-connected to provide persistent memory and so-called storage class memory.
Again, analytics are a key focus for media that can retain sizeable datasets in quickly-accessible media close to compute.
QLC flash: At the other end of the solid state continuum is QLC flash, which offers the rapid access of solid state with a cost that threatens (but has not yet come close to parity with) spinning disk media. With good sequential I/O (but relatively limited endurance), QLC products are good for bulk storage of data that may need to be accessed quickly and in large chunks.
That makes QLC suited to batch analytics, and retention of historical data, such as backups, for which it suits rapid restores. NetApp and Pure Storage are key suppliers in this space. Some suppliers are innovating with storage architectures using such components by, for example, combining QLC and storage class memory.
Storage architectures and form factors
The emergence of these technologies that support digital transformation create choices in systems architecture and form factor.
Hyper-converged infrastructure: Key among these in terms of storage hardware is hyper-converged infrastructure, which marries server and storage (usually with a hypervisor) in nodes that can scale easily – though not always independently in terms of compute and storage – to create clusters.
The big advantage of HCI is the ease of scaling and administration. Disadvantages include potentially lesser performance than dedicated storage infrastructure. Nevertheless, HCI is a popular choice for those embarking on technology refreshes.
SAN and NAS: As hinted above, dedicated storage infrastructure – on-premise SAN, NAS and object storage – has not gone away and indeed innovation is constantly occurring, if not in fundamental architectures then in terms of offering the latest media, with consumption models and connectivity to cloud tiers available from all the big suppliers. Although it is possible to form the idea that HCI, the cloud etc have eclipsed “traditional” forms of storage, that’s not nearly the case in market terms for SAN and NAS.
As-a-service: While it is possible to describe emerging storage technologies and innovations in storage architecture, the key overarching change is in much increased use of the cloud and in consumption models on-premise that mirror as-a-service delivery methods.
What do these have to do with digital transformation? Well, pretty much all you can do on-premise storage-wise can be done in the cloud too. What’s different is that you can in theory – there are caveats and pitfalls – scale use up and down and do so rapidly according to business need. And potentially, use of containers – nowhere near universal yet – will allow portability of applications across cloud and on-prem.
Meanwhile, consumption models for on-premise storage hardware are now on offer from all the key suppliers, often with predictive AI-based telemetry. That too brings the kind of speeded-up scaling and as-a-service ethos that comes with digital transformation.
