Objects in your rear-view mirror may be closer than they appear. For datacentres, that driving warning might also apply to risk forecasting – with Brexit, Covid-19, supply chain shortages and other global uncertainties blurring perspectives.
Adam Bradshaw, commercial director at colocation and disaster recovery (DR) provider ServerChoice, agrees that it is difficult to pinpoint the Brexit effect – but “it’s very easy to find excuses”, with typically risk-averse datacentre operators taking a longer view than most.
Operations managers typically pre-prepare for resupply or governance issues, for instance spreading risk and diversifying supply well in advance. That said, security of supply has been “paramount” of late, says Bradshaw.
“What happens if we need to stop burning diesel, or make sure that, at the very least, we have resupply on-site with the types stipulated within agreements with our supplier?” he adds.
ServerChoice regularly tests agreements – including conducting dry runs with fuel suppliers to ensure they can be called on “at any given time”, and can be on-site and ready within a given period. Datacentres need to ensure diverse customer and supplier bases, even across multiple exchanges, says Bradshaw.
“It’s not just an overarching agreement where we say we’re going to be online 24/7,” he says. “It’s that we have tested every component in the supply chain.
“Customers tend to be far better informed now than two years ago. I think they are more sceptical of ambiguity – is their data actually in Dublin, or somewhere else?”
Adequate engineering skillsets are available 24/7 and more flexible contract terms might suit customers that may be struggling with risk to their own businesses as a result of the UK becoming a third country, says Bradshaw.
He downplays sustainability pressures, however, noting that cost reductions from efficiency can dovetail with some greening activity while saving money. But there is potential there for increased cost in coming years, he concedes.
Brexit, naturally, mostly affects the UK and how it works with the European Union (EU) – and it complicates the risk picture.
Global risks further blur DC view
Risks affecting datacentres globally include the flatlining of power usage effectiveness (PUE) and demands for greater sustainability, rack densities, server utilisation, skills shortages and supply chain problems, according to the Uptime Institute’s global survey of datacentre operators.
Chris Brown, chief technical officer at Uptime, notes that improving PUE may entail a compelling economic reason to build datacentres or switch to liquid cooling, unless energy price rises accelerate the sustainability imperative.
“What datacentres have today is sufficient for cooling what they have and the price per kilowatt-hour [kWh] is not high enough that a small change in PUE will have an effect,” he says. “Older datacentres that can’t be retrofitted or upgraded continue to hold on.”
Increased compute load requirements are not yet driving huge changes to server technologies. However, manufacturers are expecting to need more heat-sensitive, generative components in every box, which means rack inlet temperatures must come down.
This is likely to entail more compressor-based cooling – which a lot of PUE improvements have been based on so far, says Brown.
Also, an ongoing datacentre staffing and skillset shortage could make it harder to manage and mitigate outages – which, although less frequent, are having heavier impacts, partly because of increased digitisation.
The only way to eliminate that risk is to become “a lot less reliant” on single datacentres, spreading the compute load across multiple platforms so that systems can transfer “without skipping a beat”, says Brown.
Meanwhile, hyperscalers are increasingly likely to cause disruption by setting up their own supply chains, with costs rising around IT training and certification as well as logistics and supply.
“In future, any reduction in service quality will be felt – so prices could rise and severity and visibility could well go up,” says Brown.
David Friend, chief executive of cloud storage supplier Wasabi, suggests that the risk picture means more organisations are favouring cloud, especially public cloud.
Most “notable” data theft resulting in business shutdowns and intrusion is on-premise, he says. After all, Amazon, Google, Microsoft and so on employ “a legion of PhD security experts and white-hat hackers”, something “few corporate datacentres have the scale or incentive for”, says Friend.
“Also, everybody strives for 100% uptime – but, you know, nobody ever literally achieves it,” he adds. “Everybody has a few minutes of outage here and there.”
Toby Pearson, chief executive of data consultancy DTSquared, agrees that the risk picture for datacentres has changed over the last two years, involving multiple causes and effects – and he does see clients moving to cloud.
“But they need to understand what that data is,” he says. “The data is maybe dirty, and it costs money to process. They say ‘can you take 100% of my data and turn it into the 12% I need to store?’”
We need to talk about data
Even cloud data is still somewhere, says Pearson, and managing it represents risk down the track for business customers of datacentres, causing “significant drag” on business innovation and growth.
“The real problem is the absolutely enormous mountain of data they have to move,” he says. “They ask ‘how can I use AI [artificial intelligence] or ML [machine learning]?’. To be honest, you have really neglected your data – and data ethics – for the last 30 years.”
Michael Cade, senior global technologist at backup supplier Veeam Software, says customers also repatriate data on-premise, especially if a “suck it and see” cloud attempt did not prove cost-effective.
“Some are learning from experience,” says Cade. “I don’t think we hear enough about that side of it. We hear ‘I go to the cloud and get all these great services’. There are people out there bringing it back as well.”
Cade agrees with Pearson that managing hoards of unstructured, unclassified, dirty data represents a problem for customers and represents risk to datacentre operators too, especially with increased remote working and the prevalence of bring-your-own-device (BYOD).
“Wherever it is, it’s running in some sort of software, backed by commodity hardware,” he says. “It’s very easy to go and delete the wrong file, for example.”
Costly paperwork versus increase in colo
Terry Storrar, UK managing director of infrastructure-as-a-service (IaaS) company Leaseweb, suggests the Brexit effect may remain obscured, even “swept under the carpet slightly”, beyond the “real struggle” to collect duties and other new paperwork-related tasks that add cost and risk for all types of business.
“I don’t think we’ve seen the full impact of Brexit yet,” he says, adding that so far he has seen an increase in colocation, reflecting continued expansion in datacentre utilisation. “But the scale of the pandemic has certainly changed people’s appetite for risk.”
To counter rising labour costs, Storrar suggests working smarter, leveraging automation and machine learning, where possible, to free staff up for core tasks and innovation.
“I don’t think we’ve seen the full impact of Brexit yet” Terry Storrar, Leasweb
For Simon Chappell, CEO at DR managed services provider Assured Data Protection, although data sovereignty has been the question on customers’ lips, backup aspects have been less discussed until rising cyber attacks pushed DR up CTO agendas.
“A lot of our customers had primary workloads within the EU, and DR plans in places that would become outside the EU, requiring a significant scramble,” he says.
Datacentre operators in nearer EU countries such as Ireland and the Netherlands have been “winners” as customers sought to separate and add secondary locations during the year with regionalised subscriptions and auditable statements of work, he says.
“But more and more people want to get rid of datacentre operations – and it’s much easier than it ever was to start up new places for workloads.”
Should datacentres move deeper on cloud?
Danny Nicholson, principal cloud specialist at cloud services provider Claranet, agrees that Brexit has changed the supply chain picture on lead times and pricing, at least in ways that encourage cloud and hosted offerings.
Datacentre operators should increasingly figure out how to take a slice of that revenue, he says. “Physical is no longer the de-facto state. And 99 times out of 100, customers don’ know what is in their datacentre. I would challenge some of the risks: the biggest risk is that you fall behind your peers.”
Bill Bould, general manager for Europe at datacentre-focused Fusion Risk Management, notes that risk appetite is driven by customer expectations and service-level agreements (SLAs) alongside governance and regulatory requirements.
While Brexit has meant end-to-end supply chain and logistics reviews and changes in approach to ensure inventory, there may not have been shifts in customer base or risk responsibilities, says Bould.
This, of cours,e entails a tighter focus on information security – whether or not that results in availability issues.
Peter Groucutt, managing director of Databarracks, a DR services provider with its own datacentres, agrees that data sovereignty remains top of mind for many, with nuance in the overall downtime risk picture difficult to assign to a single cause.
There is a move towards the cloud, with the large public-cloud providers increasingly able to confirm local storage that can meet compliance needs, he says.
“With few exceptions, we don’t write much business now where it’s data coming into our datacentre or colo,” says Groucutt. “Public cloud was probably going to be the direction of travel at some point anyway.”
Source is ComputerWeekly.com
The datacentre sector could become one of the fastest-growing industries in the UK, but doing so will require greater collaboration between...
Dit bericht verscheen eerder bij maincubes
Ein zukunftsweisendes Business, eine faszinierende Arbeitswelt. Deine Themen?Willkommen im Team von maincubes!Die Welt wird digital – und mit ihr...
Source is ComputerWeekly.com
An operating system (OS) is the program that, after being initially loaded into the computer by a boot program, manages all...
Source is ComputerWeekly.com
Amazon Web Services (AWS) is experiencing a “significant re-acceleration” in its growth, its parent company’s CEO has declared, after the public...
Source is ComputerWeekly.com
Meta, the parent company of Facebook, Instagram and WhatsApp, has reported a 19% increase in revenue to $40.6bn for the third...
