Work video calls . . . home-schooling platforms . . . entertainment streaming services . . . Millions have turned to them amid the Covid-19 pandemic. And they have turned two decades of growing demand for remote cloud computing solutions into a sudden boom.
Cloud computing — putting data and applications into large data centres, accessible via the internet, rather than storing them on devices — has made all of these technologies possible, and available to nearly all consumers.
But some corporate users are still discovering that migration to the cloud can be a bumpy ride.
For companies, the main attraction of the cloud is being able to add large amounts of computing power at the click of a button, while paying only for what they use. Their spending on public cloud services will grow 18 per cent in 2021 to $305bn, according to a forecast by research company Gartner.
For all this growth, though, some businesses have been underwhelmed after moving software once stored on their premises to data centres run by cloud providers. Only 37 per cent of companies questioned by Accenture, a consultancy, said they were getting the full value they expected from cloud technology.
Problems identified in the research included a shortage of the necessary skills on the part of the cloud provider’s business clients, and difficulties moving old systems to the cloud.
Cloud projects are among the most complex types of corporate IT projects because the technology often affects most parts of a business.
Even so, consultants often find that customers have unrealistic expectations of the savings and underestimate the costs of modifying IT systems to run in the cloud, and of training staff.
“Calculating a financial business case for cloud is exceptionally hard [but] the calculators on cloud provider websites make it seem simple,” observes Mark Brinda, a partner and expert in cloud computing at Bain & Company, a consultancy.
When budgeting for a cloud project, organisations can add 30 per cent or more to a cloud project’s known costs for inevitable unforeseen expenses, he says.
Costs can also creep up if companies are not flexible, says Scott Chancellor, chief product and technology officer at Apptio, which advises companies on cloud projects. For example, a company might pay for a temporary increase in computing resources to help it deal with a spike in web traffic, but then forget to scale back when demand dips.
“It’s easy for costs to easily spiral out of control and it’s exceedingly time consuming to manage and track these costs without the appropriate [software] tools,” Chancellor says.
To optimise software for the cloud, some clients of cloud companies build applications differently as “microservices”. These are functions within an application — such as checking a customer’s name from a database or a customer’s bank balance — that can run independently.
Carlsberg, the brewer, customised some of its IT when it moved 700 servers and 350 applications to Microsoft’s cloud platform. Sarah Haywood, chief information officer at the Danish company, says moving to the cloud was an opportunity to “bite the bullet and go to a digital infrastructure”.
The move paid for itself within two years through lower IT costs and reduced IT downtime, Haywood says.
But the ease, or difficulty, of a migration will partly depend on the type of cloud service required. The simplest is “infrastructure as a service”, in which a customer rents computing capacity supplied and managed online by the cloud provider. The customer runs its own software, while the cloud provider runs the data centre and servers.
However, this “lift and shift” option may create limited benefits for a customer because only the location of the IT has changed, says Carla Arend, head of European cloud research at research company IDC.
“Without changing the fundamental design of the application [customers] will not get the expected cost savings [from cloud technology],” she says.
Some companies are building new software applications designed for the cloud or overhauling the structure, or “architecture” of existing software to make them run better in the cloud. Doing so can take years and adds to the cost of a cloud project. However, the savings and improvements in performance can be worth it.
Capital One, a US bank, uses software “microservices” and Amazon’s cloud services to monitor large volumes of customers’ credit card transactions in real time for potential fraud. A digital “intelligent assistant”, can flag a potential fraud if, for example, the tip added to a restaurant bill card transaction is suspiciously large. The bank can then alert the customer.
Organisations planning a cloud project should be prepared for a fundamental review of the systems first, says Will Grannis, managing director of Google’s CTO Office, a unit at the technology company focused on collaboration with its largest customers. He says companies have to identify those applications that are vital to their future.
“It’s a bit like a really thorough spring clean for a company's applications and IT systems,” he says. “Doing it before a migration to the cloud can help companies clarify their IT priorities and strategy.”
Given that many cloud projects are part of bigger digital transformation projects, moving data online can be the easiest part. Tougher steps include training staff, adapting IT systems and using technologies such as artificial intelligence and data analytics to improve a company’s performance.
“The real work starts after the migration,” says Carlsberg’s Haywood. “You have to reimagine your [technology] architecture, manage and develop technology in an entirely different way — and equip your IT team with new skills.”
Get alerts on Cloud computing when a new story is published