Preparing for your journey to hybrid multicloud: modernize the storage infrastructure you already have
By Ron Riffe | 5 minute read | October 21, 2019
This is Part 4 in a 6-part series on storage for hybrid multicloud. Part 1 set the stage – getting us all on the same page with the terminology of hybrid multicloud. In Part 2 we explored what’s behind this rapid industry shift toward hybrid multicloud infrastructure. In Part 3 we took a look at how IT organizations are approaching their journey to hybrid multicloud and the unique storage challenges that are emerging. We also posed the question, “So how do you move forward?” In Part 4 we’ll begin exploring some best practice guidance for navigating the change.
Before we get too far into the hybrid multicloud, let’s first talk about modernizing the storage infrastructure you already have. Fortunately, this is not something that requires you wait for a lease to expire or for data growth to drive a new hardware purchase. You can get started right away.
Strategic storage choices are more about the software foundation you select than the hardware it happens to be running on at the moment
The first idea I would like you to consider is that strategic storage choices are more about the software foundation you select than the hardware it happens to be running on at the moment.
Most of us have grown up with storage systems that have capabilities tied to them. Going forward, the capabilities can be portable. Here are some fundamental concepts to consider (ABC’s if you will…):
- Agility: There are all kinds of reasons for moving data without disruption among storage systems. Lease expirations, performance optimization, technology or vendor changes and tiering to place the right data on the right hardware just to name a few. The right storage software foundation can facilitate that regardless of your choice in hardware vendors.
- Bridge: The technology storage is connected to shifts. Unix is on the decline, Windows and VMware are big. Linux and containers and Kubernetes are on the rise. Changes like these don’t always wait for your storage hardware lease to expire. Software, on the other hand, can be updated and adapt much more quickly.
- Cloud: Unlike a physical piece of hardware, software capabilities and their associated application programming interfaces (APIs) can be deployed in traditional data centers and across a hybrid multicloud. We’ll talk more about why this is important in a few minutes.
- Data reduction: Data reduction techniques like deduplication and compression can really help you save infrastructure costs. But if you are like most IT organizations, you own or lease a lot of storage hardware that pre-dates data reduction. With the right storage software foundation, you can extend data reduction across all your storage, old and new, to store more data on the storage you already own.
- Encryption: The same can be said for encryption. The business cost and frequency of cyber incidents continues to rise. The right storage software foundation can improve cyber resiliency by encrypting data at rest across ALL of your storage.
Storage can simplify your operations
Most IT organizations today have storage systems from multiple vendors on their floor. Often, they also stand up a separate vendor-specific monitoring and management platform for each of these silos. It can be complicated, and complication can contribute to downtime.
For storage, you can think differently about monitoring and management tools. First, think heterogeneous. Why have a different set of APIs, procedures and interfaces for each vendor silo? Choose a consistent software-based approach that supports all the hardware vendors you might choose from.
Second, think cloud. There are a couple of important benefits you can derive here. Clouds can house data lakes with anonymized telemetry data representing thousands of storage systems from companies across your industry. Analytics can run against this data offering you some insights on best practices. Some clouds also have AI services that can be tapped to help you be more proactive in managing your environment and resolving problems. And finally, if your tooling is running in a software-as-a-service (SaaS) manner on the cloud, somebody else is responsible for keeping it up and running. It is like having an extension to your operations team.
Storage software is perhaps the fastest growing segment in the storage industry, but that doesn’t mean physical storage infrastructure is not important
These first two thoughts underline the idea that strategic storage choices are more about the software foundation you select than the hardware it happens to be running on. It is true that the software creating this storage foundation is perhaps the fastest-growing segment in the storage industry, but that doesn’t mean the physical storage infrastructure is unimportant. One of the factors driving this surge in software is that this software-defined style of building a storage infrastructure enables the fastest adoption of new hardware innovations.
Think about it for a moment. If your infrastructure is based on storage systems that each have their own APIs and procedures, it can be quite costly to insert “something new.” We often see IT managers who desperately want to take advantage of some new innovation but have to wait until their preferred vendor gets around to introducing it. On the other hand, if your infrastructure is based on a strategic software foundation that happens to support nearly every hardware platform, then you can react quite quickly inserting new elements into your infrastructure without having to change APIs or procedures.
As a quick case study from recent history, IDC detailed the economic benefits of investing in all-flash arrays versus traditional hard disk drive (HDD) systems. Their findings suggested savings in 6 areas:
- Far fewer devices are needed due to significantly larger-capacity devices
- Lower energy and floor space consumption on a watts-per-terabyte and terabyte-per-rack-unit basis
- Need for fewer servers due to the much better CPU utilization driven by latencies that were 10x better than HDDs
- Lower software licensing costs driven by the need for fewer servers
- Lower administration costs as the time spent tuning the performance of slow disk drives went basically to zero with all-flash array systems (AFAs)
- Better reliability because solid state media is electronic, while HDDs are mechanical
All told, the transition from HDD-based arrays to all-flash arrays for mixed primary workload consolidation saved most organizations 60-80 percent over the life of an enterprise storage array1. The point is that designing your storage infrastructure so that it can absorb the latest hardware innovations can be very important.
Two impactful innovations that are worth paying attention to are Non-Volatile Memory Express (NVMe) and Storage Class Memory (SCM)
Since probably the mid-1990’s, hard disk drive systems have been built around the small computer system interface (SCSI) protocol. With the rise of the kind of apps that run on hybrid multicloud and the performance demands they generate, a very small group of companies, including IBM, started shipping custom flash systems around 2012. With the speed of these systems, it became apparent that a new, much higher performance protocol was needed.
NVMe is the new standard, and it has now matured. One of the top reasons IT managers are adopting it is that it helps increase infrastructure density — because it delivers more predictable and faster throughput. Said another way, it helps reduce energy and floor space consumption to meet a particular application requirement.
The other is Storage Class Memory. SCM is an even higher performance tier of persistent storage than NVMe Flash — roughly 1/8th the latency — that over half of NVMe users expect to be testing before the end of the year.
With the right infrastructure-independent software foundation, you’ll be able to quickly adopt these new technologies and transparently move important data to them for maximum benefit.
Assuming you choose some of the modernizing steps we’ve discussed to this point, you’ll likely generate some meaningful savings. That gives you the freedom to make some choices. One perfectly valid choice that we often see is to “bank” the savings and be proud of your progress.
Another choice we see customers make is to take the margin they’ve created and invest it in some innovation around storage for hybrid multicloud. That’s where we’ll focus in the next post. Stay tuned for Part 5!
What do you think? Which of the ideas for modernizing the storage you already have stands out the most for your organization? Why? Join the conversation using one of the social links below!
Click here to read Part 5.
 IDC, March 2019, IDC TechBrief: Persistent Memory and Storage-Class Memory, Doc # US44891819