Zdenk Kluka - stock.adobe.com

Gartner: Most firms not tapping cloud’s full potential

Organisations have not fully leveraged the full potential of cloud and must navigate emerging trends such as artificial intelligence, multicloud complexity and security perceptions to stay competitive

Cloud has become a business necessity, but “there is still a tremendous amount of headroom left for most organisations”, said Joe Rogus, advisory director at Gartner, during his keynote on the future of cloud at the recent Gartner IT Infrastructure, Operations and Cloud Strategies conference in Sydney.

Rogus warned that businesses are at risk without a solid strategy for moving up the cloud value chain. Most organisations currently see cloud as a technology disruptor, but he urged them to progress through using it as a capability enabler, an innovation facilitator and ultimately a business disruptor – a state Gartner predicts will be common by 2029.

Gartner’s view is that the cloud market – currently seen as comprising infrastructure, platforms and software-as-a-service (SaaS) – is morphing into just two components: SaaS and engineering technology platforms. He added that “it’s becoming more and more difficult to distinguish between infrastructure and platform services”, while SaaS suppliers are increasingly offering platforms for creating organisation-specific applications.

This shift is partly driven by economic factors, with business departments procuring SaaS, developers using platforms for applications, and infrastructure teams running infrastructure-as-a-service (IaaS) – often independently, preventing the business from extracting the full value of cloud.

Rogus suggested that effectively combining business technology platform with engineering platform capabilities, and enabling a marketplace of vendors, partners and internally built software, will allow organisations to derive greater business value from the cloud, especially when packaged for different user personas.

At the conference, Rogus also identified three current trends in the cloud market: cloud dissatisfaction, increasing demand for artificial intelligence (AI) and machine learning (ML) capabilities, as well as multicloud and cross-cloud operations.

Cloud dissatisfaction is real, Rogus said, with organisations repatriating workloads from the cloud. While sometimes driven by unanticipated costs, such moves often stem from inadequate planning in cloud implementations, ongoing operations optimisation and governance.

“There is a direct correlation between poor outcomes and improper execution of these upfront activities,” he said, suggesting that successfully addressing these areas will likely result in cloud satisfaction.

Gartner believes that over the next four years, about 80% of cloud repatriations will continue to be linked to one-off scenarios for specific projects or use cases, mirroring the current situation. However, this could be disrupted by geopolitical and cloud sovereignty issues, or commercial disruptions such as recent changes to VMware.

“When repatriation does take place, it probably will not just be a simple shift back to your existing on-premises virtualisation platform,” Rogus said. “Lift-and-shift has long been frowned on by many of our analysts because it is well known for not delivering on the expected cost savings results. But with these changes happening at VMware, we find that this perspective is shifting. We believe that by 2029, over 20% of organisations will have rethought their stance on lifting and shifting their migrations.”

While AI and ML activity is largely driven by hyperscalers, Rogus cautioned organisations against equating AI solely with generative AI. He noted that no AI works without data, and most enterprise data remains on-premise, not migrating to public cloud. This means organisations need to consider a hybrid scenario, and that implies datacentre modernisation.

Also, the growing use of AI and ML will lead to the percentage of cloud workloads related to AI growing from around 10% to 50% by 2029, “and that’s in addition to the fact that the total pie of cloud workloads is, of course, going to increase as well”, said Rogus. So, organisations should start thinking about energy requirements and cost implications of increased AI and ML use while there’s still time.

Multicloud adoption is widespread, often for valid reasons. “Most organisations are selecting a primary cloud provider and then selecting a different [additional] provider that helped them with a specific use case that their primary provider couldn’t support,” said Rogus. Additionally, post-acquisition, immediate migration of an acquired company’s workload to the preferred cloud platform isn’t always practical. 

Although 81% of enterprises are already multicloud, Rogus highlighted drawbacks. For example, achieving resilience requires them to write applications for multicloud environments while seeking lower costs is complicated by the need for skilled staff in each environment, along with reduced negotiating power with fewer workloads per cloud. “So, for these reasons, we expect that by 2029, more than half of enterprises who take a multicloud approach will not see their expected results back,” he said.

Turning to the common requirements of cloud, Rogus highlighted industry solutions, digital sovereignty and sustainability.

“We track more than 60 different vendors that provide industry cloud platforms” that combine software with PaaS and IaaS to deliver composable building blocks, said Rogus. Gartner predicts that over half of all enterprises will use these platforms to accelerate development, simplify compliance and avoid data silos by 2029.

Rogus noted that suppliers in this space typically originate from two directions: hyperscalers, who start with technology and suit tech-leading enterprises; and industry-focused suppliers, who address industry challenges by applying relevant technologies.

Digital sovereignty, he added, is an important consideration involving three distinct principles around data, operations and technology.

Data sovereignty concerns compliance with home country laws and regulations, irrespective of data storage location. Operational sovereignty relates to transparency and control over a provider’s operations, often favouring trusted domestic partners over those under foreign government control. Technological sovereignty refers to an organisation’s ability to ensure its own technology independence.

“So, the key takeaway here is to not jump to conclusions around sovereignty. Consider all three of these different principles. And then make sure that you understand what exactly your requirements are, so you can pick the right solution or the right mix of solutions,” Rogus said.

Sustainability remains a top priority, according to Gartner. Rogus noted that in some regions, power generation capacity is insufficient for additional datacentres, and AI workloads “have power requirements unlike anything we’ve ever seen before”. Consequently, Gartner predicts that over half of global enterprises will prioritise sustainability in their procurement processes by 2029.

Finally, Rogus urged businesses not to overlook security, observing a shift in attitudes. A decade ago, cloud was often deemed less secure than on-premise environments, however, “there has been an evolution here from cloud security being a concern to now actually being a standard”, he said, adding: “From an architectural point of view, cloud is in many ways better than what you would already find existing in a lot of organisations.”

Gartner suggests that cloud computing will set the security standard by 2029.

Read more about cloud and AI in APAC

Read more on Cloud computing services