Trending Articles
Trending Topics
Trending
An organization’s storage setup can be daunting, but there are smart strategies to help untangle the chaos, optimize existing resources, and give you the grounding to take advantage of AI-driven insights.
Edge-to-cloud, built to transform your business. Learn more about it here.
Businesses have a problem. They operate in a world where data is increasingly important and where ready access to that data has become essential. However, in too many cases organizations have a complicated mess when it comes to storage provision.
It’s easy to see why such situations occur. Different storage technologies come and go over the years and sometimes it’s simply easier to stick to that you’ve got, even when newer technologies have superseded them.
For Tony Lock, distinguished analyst with Freeform Dynamics, there are three questions that need to be asked. “What do we have and who is using it? What do people do with the storage? And the big one: what’s available to help me and what can fit in the current infrastructure?
In other words, there’s a need to get to grips with what’s available. Many companies don’t have that level of insight.
In many cases, storage products are bought by departments for their specific needs. A siloed approach means companies are missing out on efficiencies.
But there is a solution. “Companies can look to juggle workloads to get the best of them,” says Lock. “It may be the case that Department X paid for a storage system but, in reality, it can be better used by Department Y.”
Nor is it simply a matter of more efficient use of existing products: organizations should explore the possibilities of virtualized storage, which could be used for multiple applications. It’s all part of the process of making more effective use of the systems that companies have.
In other words, get to grips with what’s you currently own. This immediately raises one problem: many companies don’t have that level of insight. Gaining that insight should be a priority.
This may flag another issue, namely the length of time that products remain active.
There used to be an understanding that devices had a write-off life of three years before they were renewed; over the years this has stretched to four or five years. Lock warns that when it comes to storage, that couple of years is being extended even further up to seven years. “Although,” he said, “there were cases that I’ve heard of where storage products are being used ten years after they were bought.”
Juan Orlandini of solutions integrator Insight said that there were ways around this problem with modern storage systems, including those from HPE, as they often allow for in-place upgrades. “This shifts the focus from the entire system’s lifespan to that of its components, like controllers, shelves, and even drives. These individual components can indeed last ten years or more as you rotate and upgrade them”
The demand for data storage is growing in parallel with data creation, but the needs around type of data storage have branched away. This whitepaper looks at challenges current legacy storage systems present and which solutions are available to meet modern market demands. Specifically, it considers the latest storage needs around structured data, and how this should be managed in a cloud-centric business landscape.
Orlandini added that companies must be smarter when it comes to purchasing decisions. “The key is to invest in a system that supports such incremental and non-disruptive upgrades. This approach means you’re not ‘integrating’ an old, static system as much as you are evolving your current system over time.
For Orlandini, there are several issues that companies face when implementing a modern storage system, highlighting, in particular the capability of modern products. “We see companies underestimating the value of a multi-protocol solution – one that can handle block, object, and file storage from a single management platform – and that is a significant oversight.”
The second issue, he said, is that companies are not accurate in planning for the future. “Companies often either underestimate their growth, which means they might outgrow their initial capacity and face a massive overhaul instead of an incremental upgrade – or they overestimate their needs and have unused capacity ‘wasting’ capital.”
He added: “With environments increasingly becoming hybrid, not having systems that allow data interaction across both private and public cloud locations is a crucial mistake.”
The use of AI and the rise of machine learning has also influenced the way that the systems work. “AI is growing in the impact that’s it having,” said Lock. “Everything is contact sensitive here, making sure that the right data is in the right platform with the correct levels of security.”
There’s also the small matter of accessing data going back many years, for purposes of data analytics and AI training. Archiving has become much more important and companies must look at how readily that data is available. “People are taking 360-degree decisions and need to access historical data,” he said.
Orlandini agrees that the advent of AI will have a major effect, but it won’t necessarily mean that companies need to shelve their existing legacy stack. “The full impact is still materializing, but the initial thought that every company would need entirely new storage for AI workloads is probably not the case for most.”
He also agrees with Lock that companies must look closely at what their current systems offer.
“The majority of businesses will be ingesting their existing data into AI systems for new kinds of processing,” said Orlandini. “This likely means storage will need to support a couple of distinct workload patterns: a highly read-intensive, high-throughput pattern for the initial data ingest, classification, and vectoring, and then, as these AI technologies mature, a lot of random I/O with high IOPS for ongoing operations”
It’s early days for the incorporation of AI into organizations. “However, it’s important to note that we are still working through this as an industry,” added Orlandini. Companies must be able to respond quickly to changing demands.
This is where he sees the strengths of companies like HPE, as it provides the flexibility that organizations need to address these modern challenges. “HPE offer the ability to scale out both horizontally and vertically, support multi-protocol environments for block, object, and file storage, and are built for hybrid cloud integration. Essentially, HPE systems are built to handle pretty much any kind of workload you need to support, and that includes the evolving demands of AI.”
