How cloud automation is changing IT: repair, replicate and correct
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
When Moonpig wanted to speed up the development and release of its online greetings cards, it turned to cloud automation.
Normally, a new feature such as “group cards” — which allows up to 20 people to sign one card for, say, a colleague’s wedding or retirement — would take several months to develop. But, by automating the creation and modification of cloud processes, Moonpig cut that to a few weeks.
These days, the London-based online card and gift retailer releases more than 340 new features and updates every month using cloud automation, compared with just 10 a month under its old system.
“The increased velocity that cloud automation provides means that we release [updates] many times a day,” says Peter Donlon, Moonpig’s chief technology officer.
This faster release of incremental changes “means we have significantly reduced the chance of any bad releases, negating the need to do things like rollbacks, increasing the productivity of our engineers”, he says.
This “contrasts with the old way of doing things whereby releases were made once a fortnight or less and came with a higher degree of risk”.
Originally, IT professionals relied on manual processes for provisioning, configuring and managing their cloud-computing infrastructure.
But these tasks can now be automatically carried out using cloud automation tools and software — allowing businesses to cut costs, free staff for more productive work, reduce errors and improve security.
“Many companies have already started moving parts of their business to the cloud,” says Richard Smith, senior vice-president for technology in Europe, the Middle East and Africa at Oracle, a US cloud computing group. “By automating much of this they will accelerate their digital transformation. It boosts efficiency and allows for better management of business operations. Historically, a new database could take weeks to have up and running, while an automated one can be spun up in seconds — without the need for a highly technical expert.”
Cloud automation can also assist with back-up, managing queries and interacting with customers, says Smith. “While it enables companies to grow their business, it helps them to get the most from data, including a deep insight into customers.”
He also notes that the tools can prevent delays in updating technology systems, which can improve security. “We know that the largest cause of security breaches stems from systems [that are] not up to date. Automated cloud technology ensures the latest fixes are installed and issues flagged as they arise.”
Pip White, UK and Ireland managing director at Google Cloud, says automation tools can be used in areas ranging from product development to the financial administration of cloud deployments. “Automation removes a lot of the manual labour associated with processes like billing management or creating cloud delivery pipelines,” she says.
However, not everyone sees cloud automation as solving every problem.
A word of caution is offered by Emma Roscow, lead for intelligent cloud infrastructure at consultancy Accenture UK. She says that while automation has its benefits, there are also significant challenges.
“Just because a process is automated doesn’t mean it can be left entirely unchecked,” Roscow points out. “Although automation reduces the chance of human errors, if there is a misconfiguration in the initial set-up, or a vulnerability creeps into the system, it could spiral into a security issue.” This means that cloud automation needs regular reviews, she adds.
That view is backed by Ajay Sabhlok, chief information officer at Rubrik, a US cloud data management company. “Though efficient in the long term, cloud automation tools require a lot of budget upfront,” he says. “These costs are usually in the 5-10 per cent range of the total cloud infrastructure spend.”
Sabhlok says security has to be paramount when writing automation scripts, as “human errors introduced . . . could make them vulnerable to outside threats where a hacker gains access to them and carries out a ransomware attack, steals sensitive data or corrupts it”.
More stories from this report
Some believe the next step in the technology will be “self-healing automation”, whereby artificial intelligence and machine-learning algorithms are applied to testing procedures.
“Self-healing automation is a game-changing concept,” reckons Mark Turner, director of cloud business at Claranet, a London-based IT services provider. He envisages IT services being able to “detect problems, and adapt, move, repair, replicate and correct themselves”.
But even if better cloud automation tools continue to eradicate manual computing tasks, they will still require human intervention, argues Alessandro Perilli, general manager of management strategy at Red Hat, a US open-source software specialist.
“When automation adoption projects fail, this is usually down to companies setting unrealistic expectations,” he says. “Automation best serves to multiply a company’s operational capabilities, helping glue together different parts of a business process, rather than replacing them 100 per cent. Automation is still a long way off from replacing human labour.”