Businesses are learning to love the cloud, but the move can still be disruptive. They must prepare thoroughly first – and then decide if it’s truly the right answer for their needs.
There are numerous recent examples of manufacturers and distribution firms discovering the benefits of the public cloud. Volkswagen has launched an ‘industrial cloud’, connecting data from 124 factories. SKF, a Swedish ball bearing company, uses a cloud connection to time when lubricants should be added to its products.
It’s easy to see the attraction. The cloud offers easily accessible data storage and computing power that can be quickly scaled up or scaled back. This offers potential cost savings and improved productivity.
However, the transition can be challenging, sparking a whole new set of worries for managers about costs, security and efficiency.
Many enterprises are still wedded to running their services through on-premises data centres, or renting servers in a managed data centre. Adoption of the cloud is certainly accelerating, but more than half of IT workloads are expected to remain on-premises in 2022, according to a survey by the Uptime Institute.
There are several factors to consider before making the move to cloud. To begin with, companies need to determine the optimum time to shift services away from on-premises servers to the cloud.
Ilja Summala, chief technology officer (CTO) at cloud services provider Nordcloud, says that once software becomes an essential part of a company’s business strategy, “this is the time that they should consider moving to the cloud if they have not already done so.” At this point, the enterprise will need the ability to instantly ramp up its software capabilities and digital solutions, which it can do with a cloud provider. This helps to reduce the cost of developing software in the cloud, although Summala notes that it “doesn’t always mean that the software is cheaper to run in the cloud, but it’s definitely cheaper to develop”.
Carla Arend, senior programme director for software at research company IDC, says the cloud’s flexibility is ideal for businesses looking to speed up their innovation and new product development. “If you want to create a prototype, test it and get users to interact with it, the cloud can provide the IT resources to do so almost instantly, and then you can shut those resources down once you’ve done your test,” she says. “You have much greater flexibility and a much lower cost profile in this experimentation phase. So it’s very good for rapid innovation cycles.”
Nonetheless, some firms have struggled to make a good fist of shifting to the cloud. If they fail to create a consistent strategy, they may struggle to get good value for money from the move.
“Wasted cloud spend is still a major issue for many enterprises,” says Rob Robinson, head of Telstra Purple EMEA, which provides technology services to enterprises. “To mitigate against spiralling costs, organisations must recognise that reducing cloud costs is not a one-time, tick-box task. It requires continuous assessment to identify precisely where overspend occurs.”
While the cloud’s speed, scale, innovation and productivity benefits boost business opportunities, tunnel vision can stop companies from achieving the best value. “CIOs and CTOs need to have a coherent and sustainable roadmap in place that captures continuous value from their multi-cloud investment. Without these, enterprises will be continually disappointed by their cloud results,” says Robinson.
Once an enterprise decides to create a cloud strategy, it should first migrate the applications that take the least effort: the “low-hanging fruit”, says Sascha Giese, head geek for technical product marketing at SolarWinds.
“Probably the most obvious one is a mail server – there’s little reason to keep a mail server on-premises these days,” says Giese. “Databases are next, as it’s relatively easy to migrate data from an on-premises database to a cloud-hosted, maybe even a cloud-managed, one.”
He adds that human resources, customer relationship management and data warehousing are all easy wins when it comes to transferring functions to the cloud.
Persuading supply chain-based industries such as manufacturing, distribution and building supply to shift to the cloud has been a work in progress since the middle of the last decade. There is considerable activity in this area. Volkswagen Group last year announced it was working with AWS to launch the industrial cloud, bringing together production data from 124 factories on a single digital platform. The aim is to use the data in real time to improve productivity by 30%, for instance by optimising the use of machinery and the flow of materials.
A long-term aim is to use the cloud platform to create an open marketplace for industrial applications. “On that kind of platform, everyone involved would be able to swap, acquire and use each other’s applications. It would be a place open to all companies in principle, from suppliers through technology partners to other car manufacturers,” says Nihar Patel, Volkswagen AG’s executive vice-president for new business development.
However, some of the most important processes in manufacturing systems are “literally bolted to the factory floor,” says Summala. Manufacturing execution systems (MES), which control production processes, are unlikely ever to make the jump to the cloud, he says. “It’s those MES systems that I suspect will stay inside the factories for a long time, because if there were a network breakage, it might mean that you could not manufacture anything. That’s not a risk worth taking, even if the cloud is cheaper and better.”
One benefit of moving manufacturing processes to the cloud lies in predictive maintenance of machinery, says Tobi Knaup, chief executive of D2iQ, which advises businesses on their journey to the cloud. For instance, a paper mill will have massive paper-producing machines costing millions of dollars that need to run continuously. If they break down at any point and parts need replacing, this creates lost revenue.
A D2iQ customer is currently deploying 80,000 sensors across its factories to ingest sensor data in real time. The data is stored in the cloud and runs through machine learning software to analyse unusual vibrations. The system can then predict which parts need replacing in the next maintenance cycle. This minimises disruptive breakdowns.
“The old-school way of avoiding unscheduled maintenance is having experts physically listen out for unusual vibrations in the machinery that suggests something might break,” says Knaup. “The new and far more efficient method is to put cloud-based sensors everywhere in the machines.”
Some think a cloud-first strategy – in which cloud solutions are looked at first when considering new or existing processes – is not always the best way ahead. Many large companies are happy running some applications in their data centres, as they have developed stable applications over time which are customised to their needs. They know the workload so they can easily procure and provision the right hardware and infrastructure.
“In these cases, businesses might find it attractive to do it themselves through a capital expenditure model, where they can purchase the hardware and write it down over three or five years. This financial model makes sense in certain situations,” says Arend.
Raj Sukumar is head of Europe at Persistent Systems, an Indian technology services company. He says there’s a “common misunderstanding that moving to the cloud means every single organisation moving all their data to the cloud”. Depending on individual business requirements, that is not always needed.
This view is echoed by Álvaro Verdeja, chief operating officer at Making Science, who says moving away from legacy systems should not be a single step but managed in stages. “Businesses should tailor their cloud implementation and consider the most effective strategy, whether a combination of cloud and on-premises, cloud as a sole solution or multi-cloud usage – to deliver optimised results.”
A carefully considered cloud strategy, then, can potentially provide huge benefits. But enterprises must have a clear idea of the most appropriate services to migrate and keep a watchful eye on the costs.