In an ideal world, setting up a data storage system for your business would be easy. You’d look at your needs, identify the best solution, install it and away you’d go. In reality, that’s not how things tend to work.
To begin with, growth for companies isn’t linear. It happens in fits and starts, in peaks and troughs, and the technology installed over time to support the typical business tends to be an organic mix of what’s available and what’s affordable.
The end result is that the average company has bits of this and bits of that in terms of infrastructure, with IT teams tasked with managing this mish-mash as best they can.
“Things grow organically and different types of data become prevalent over different periods. That’s just the nature of things, and it’s particularly true of the last decade, where we’ve seen the rise of file storage which has been exponential,” said John O’Donoghue, solution consultant for the data centre computer group, Dell Technologies Ireland.
O’Donoghue explained that there are three kinds of storage – file, block and object storage – and each has different use cases.
“Traditionally people have stored their file data in NAS storage units – network attached storage – then generally have their production data on SAN, or storage area network, storage,” said O’Donoghue.
“There are also some people who combine these, using native file storage and block storage on the same unit, which is evident in our Unity and PowerStore range. But at the same time there are also people who want to store vast quantities of file storage and who might not need the same level of performance.”
This makes sense – some kinds of data need to be available as quickly as possible to the system, while other kinds sometimes aren’t looked at very often, or even at all. At the same time, storage technologies cost money, and the fastest and most modern of them cost a lot more than the slow and old ones.
The resulting technique of putting in-demand data on fast expensive storage and not so in-demand data on older, cheaper and slower storage is known as data tiering. The lowest tier of data is where companies store data that they don’t need to access very often.
“It’s data that you might be required to maintain but don’t actually need to use very often. Companies usually put that on the cheapest storage, and they usually compress it. The odds of you having to go looking for that data are quite low, but you still need to have it to hand for compliance reasons,” said O’Donoghue.
“Above that you might have a tier of data that also isn’t accessed that often, but that you do need to have accessible for performance reasons, rather than archived off. Then on the highest tier of storage you keep the data you need to be able to access fast and regularly – this tier is about maximising performance.”
This allows companies to run their business applications on the highest level tier where superior performance will be seen. Typically this involves storing data on solid state drives (SSDs) as opposed to older spinning disk type storage or even tape drives.
“There’s no one-size-fits-all approach in storage. What we tend to find is that our PowerStore units, which is our mid-range storage designed for most of the market, does just that, it fits most of the market,” O’Donoghue said.
“For companies with more robust needs, we have a PowerMax product, and then for those companies dealing with petabytes of data, we have our Isilon range. That handles vast quantities of data, typically what are called ‘data lakes’.”
For Sandra Dunne, sales manager for Logicalis, a key issue in deciding what kind of storage to go for and advising clients is knowing what it will be used for.
“When our customers are choosing a form of data storage, regardless of whether they are a small, medium or large company, we advise them to look first at what they require for their environment. What levels of data protection do they need and how portable or applicable is that data – where does it need to be accessed from?” she said.
The reason is that some companies operate data centres and have lots of edge locations, and others have small regional offices that need to access data remotely but not at speed. This will impact the kind of storage they need and data manageability and portability are big issues, along with acceptability from the customer perspective.
“You need to be able to put your data somewhere. People talk about SANS, NAS, Hybrid, software-defined and so on, and it can all get quite overwhelming. So instead of starting with solutions, we ask the question, ‘What do you want to do with that data?’” said Dunne,
“How available does it need to be and how protected does it need to be? The technology flows from the answers to those questions, and those are some important considerations.”
On top of these, companies usually want to know that the data storage solution they pick is as flexible and as cost efficient as possible.
“All customers have a requirement to store their data but each customer’s specific needs will be a little bit different. The key question is always: what do you want to do with that data? Today, most organisations for example want to run analytics and other tools that allow them to predict different aspects of their business, depending on their industry and so on,” said Dunne.
This means that they need a way to efficiently make data available to staff and partners wherever they happen to be. At the same time, this data needs to be protected.
“Even if you’re a small company, you still have the same basic requirements. If you have data it generally has value to your business and therefore you need to protect it,” she said. “The same rules apply regardless of scale. The goal is to drive efficiencies out of whatever method you use. You need to make it as agile and flexible as possible.”
Choosing the right kind of storage depends on the perspective of the company, according to Wayne de Vos, technical consultant at Paradyn.
“The trend today is for scalable storage, something that is adaptable to any need. Hyper converged is one option for scalable storage for example, but it all depends on what you want it to do. Cost is always a factor too. Small companies have a budget they want to work within and large enterprise companies do too – it’s always a factor in any design,” he said.
A big part of storage is making sure that whatever container you use to store the data, it can be protected properly. The recent coronavirus pandemic exposed weaknesses in many companies’ business continuity plans.
“We had clients who had virtual private network (VPN) access to their business systems so they could remotely access their stored data, but it doesn’t cater for the volume of data that needs to be put through it now. Going forward, these clients need to be made aware that their systems might technically have the capability to do something like facilitating remote access, but that doesn’t mean they have the capacity to handle the volume that might be needed,” said de Vos.
“Right now, everyone is working from home and many VPNs are falling over, and the result is that people can’t work. So that needs to be designed for. The coronavirus exposed the fact that these systems weren’t as robust as they thought.”
Paradyn offers storage as a solution using VMWare via public and private cloud with Azure and Amazon Web Services. From a data storage point of view, de Vos said it has to be asked, are remote access systems secure enough to meet demands?
“Are they secure enough and are they compliant? I had a couple of clients who wanted to allow their corporate data to be accessed by staff remotely and we had to say to them that before they did that, they needed to make sure there were no weak links in the chain,” said de Vos.
“The virus has made many companies realise for the first time that their staff can actually work from home, not just in an emergency but going forward as an increasingly normal part of their working life. But that’s not as simple as just telling everyone to take their laptop home. You have to have the support systems in place to make sure it just works.”
There are many crossovers in terms of the features of good data storage and good data security, according to Michael Conway of Renaissance. To begin with, good examples of both feature confidentiality, integrity and availability.
“If you look back at these three qualities, they are hugely important in data storage. Storage has to make data available at whatever tier is necessary, and that’s something that people sometimes forget. They buy storage based on price, without thinking through the circumstances in which they’re likely to need that data,” he said.
“But the only reason to store data is if you intend on accessing it at some point. If it’s long-term storage, then you need to make sure your storage technology will stand the test of time. If it’s short-term storage, then you need to make sure your technology can deliver that data as fast as possible.”
Conway suggests that many business people focus too much on the cost per gigabyte or cost per terabyte of storage, when instead they should be looking at the total cost of ownership of the data over its lifetime.
“Firstly, ask yourself why you are storing it. If you don’t need to store it then get rid of it – but if you’re getting rid of it, do that properly. Don’t throw a hard disk or server into a skip and presume the job is done. You have to erase the data properly,” he said.
“If you format a disk yourself using standard tools, then usually that just deletes the index of the data on that disk. The data isn’t gone, and someone who knows how can reconstruct the index and gain access to the data. That’s happened with second-hand laptops sold online, and data has been compromised that way.”
There are a few ways to physically erase a disk, such as degaussing it with a magnet, but merely clicking ‘delete’ in software doesn’t achieve the same thing.
“This is also true if you have stored data in a data centre, and that’s important to think through. You might be spinning up virtual machines all the time on a remote disk that you never actually see or physically touch, but that data is still there. There are technologies that can remotely wipe disks and do it to the standards of the Department of Defence,” said Conway.
Paid services the way to go for SMEs
Many small companies operate their data storage on cheap prosumer-style hosting plans, despite the drawbacks. By spending a little more, you can gain access to a lot more functionality according to Michele Neylon, chief executive of Blacknight Solutions, which works predominantly with small and medium-sized businesses.
“You can’t run a company on a two- or three-euro-a-month hosting plan. It’s just a bad idea. You need to match your needs to your budget, not the other way around,” he said.
When charged with helping clients improve their storage solutions, Neylon starts by looking closely at the company’s needs and the number of staff it has.
“We then get them to switch over to using services like Office 365, which comes with storage in the form of OneDrive; you get access to Microsoft Teams and a bunch of other tools as well. If they need their staff to be able to log in to specific servers remotely and access files, they might need a specialised server working with private cloud,” he said.
“The trick is getting them to understand that technology is there to help you, but it’s a tool and you need to have the right tools for the job.”
Often smaller companies are overwhelmed by the range of storage options out there. But many won’t be suitable for them.
“Buzzwords and jargon don’t help and often technology literally comes in and out of fashion, and that doesn’t help either. Just because something is fashionable and a lot of people are talking or writing about it doesn’t mean it’s suitable for your business,” said Neylon.