Top IT predictions in APAC in 2019

Distributed AI architectures to emerge

The growing use of AI will increase data usage exponentially. As part of Singapore’s smart nation initiative, the government has planned to invest up to S$150m from the National Research Foundation on AI over five years through the AI Singapore programme.

While first-generation AI architectures have historically been centralised, Equinix predicts that enterprises will enter the realm of distributed AI architectures, where AI model building and model inferencing will take place at the edge, physically closer to the origin source of the data.

To access more external data sources for accurate predictions, enterprises will turn to secure data transaction marketplaces. They will also strive to leverage AI innovation in multiple public clouds without getting locked into a single cloud, further decentralising AI architectures.

Multi-cloud, hybrid and edge as a continuum

While the shift to cloud is happening, migrating data to the cloud should also be done with care. Too much centralisation of data in one place may risk lock-in, with associated back-end costs. It may also remove the ability to be flexible in meeting policy and regulatory requirements, said Dan Sommer, senior director and global market intelligence lead at Qlik.

Beyond data protection, managing data in the cloud has its own set of rules, and if not done right the cost, complexity, and risk can bring down the house, Sommer said, noting that the shift from on-premise and legacy datacentres should be done at a pace organisations are comfortable with.

“The ability to centrally calibrate and distribute [data and workloads] to multiple clouds, as well as hybrid on-premise and cloud continuum, are good ways to hedge bets,” said Sommer, noting that “a post-modern platform should be able to handle distributed data, workloads and usage across multi-cloud, hybrid and edge as a continuum”.

Getting ready for 5G

5G, the next-generation mobile connectivity technology, will open up possibilities of digital transformation, especially in an internet of things (IoT) and AI connected world.

Noting that APAC is on track to become the world’s largest 5G region by 2025, Equinix expects mobile operators to make massive investments in 2019 to revamp existing cellular infrastructure and build new edge infrastructure.

But in attaining the high radio density required for 5G, operators will need to look for ways to optimise costs, such as through the use of open-source commodity networking hardware and virtualisation of the wireless networking stack.

According to Ericsson, total mobile data traffic in Southeast Asia and Oceania is expected to grow by eleven-fold by 2022, and early deployments of 5G will translate into an estimated 28 million 5G subscriptions by then.

Data management takes centrestage

In 2019, the focus will be on uncovering the true value of data, driving a renewed quest to embed telemetry, as well as to collect and enrich data, according to Dimension Data. This will lead to a change in information architectures.

For example, instead of putting data in one place, organisations will need to design information architectures that provide a single view of all their data, according to Qlik's Sommer.

“Historically, it has taken a lot of effort to make that happen,” Sommer said. “Cumbersome efforts to put all the data in one place – such as the all-encompassing data warehouse or lake – didn’t reach the goal, and that’s happening again now in the cloud.”

Sommer also noted that two trends are set to shift the situation in 2019, making it possible to get a single view of all data while keeping it where it resides.

First, there are different suppliers coming together to standardise data models. This will mean cloud-based data sources will have more consistent formats.

Second, and more importantly, is the emergence of enterprise data catalogues. Accessible in a hub, data catalogues make it possible to audit the entire distributed data estate, delivering a shop-for-data marketplace experience. The more users share, collaborate and use the hub, the more valuable it becomes to the business, he said.

The continuing march of microservices

One of the biggest unsung megatrends today, said Qlik’s Sommer, is microservices and Kubernetes. What used to be monolithic is now distributed, essentially enabling a new way to scale workloads. Sommer said this will have the same quantum-leap effect as scaling the hardware, and scaling the infrastructure before it.

“It’s a way to access and process data locally, and at the edge, reaching places that modern BI [business intelligence] platforms of the past couldn’t reach,” he said.

“In the span of a year, it’s gone from emerging to becoming a hygiene factor, where application development teams at leading enterprises are now orchestrating container-based applications, and demanding production Kubernetes environments.”

More enterprises become “digitally determined”

At its recent Futurescape event in Singapore, IDC APAC’s group vice-president Sandra Ng noted that at least 55% of organisations in the region will be digitally determined by 2020, transforming markets through new business models and digital services.

According to IDC, digitally determined organisations are those that can demonstrate the ability to plan and run digital transformation initiatives through ambition, grit, discipline, commitment and hard work.

“To be one of the digitally determined, APAC organisations require more than tenacity,” said Ng, adding that such organisations require a blueprint that comprises a single enterprise strategy, resoluteness to make required organisational and cultural changes, and a long investment strategy based on the principle that digital is inherently valuable to business.