The 2-Z of 2025 predictions: D is for data management

D is for data management

Cross-border considerations

Source: RackSpace. Banerjee.

"With the Asia Pacific region experiencing diverse data residency laws, optimising cross-border data infrastructure is key. CIOs should strategically consider hybrid cloud and edge computing solutions that allow compliance with local data residency laws while providing scalability for AI applications," said Hemanta Banerjee, VP of Public Cloud Data Services, Rackspace Technology.

"Work with cloud providers that offer localised data centres in the region and edge computing solutions to minimise latency. Partner with multicloud vendors that understand the regulatory landscape and support data localisation needs in countries with strict data requirements," he recommended.

Banerjee further said that cybersecurity is paramount, especially for cross-border data transactions. "CIOs should adopt Zero Trust models that enforce stringent access controls across regions, protecting data as it moves across the region and ensuring compliance with data residency laws," he commented.

"Deploy region-specific threat intelligence solutions that address the unique cybersecurity threats in different regions. Focus on encryption and secure access management to protect data at each cross-border exchange."

Interoperability

Qlik forecasts that a common language for data would become popular in 2025. "The idea of a single data lakehouse that brings together the best of data lakes and warehouses while supporting multiple use cases has been around for a while," the company acknowledged in its discussion of 2025 AI trends. 

"However, a lack of interoperability has restricted this vision to theory. Deploying open table formats, particularly Iceberg, is emerging as a modular format every vendor will accept, as it enables companies to organise their data in any storage, avoiding vendor lock-in. This, in turn, could help reduce costs, boost velocity, and improve governance.

"But more importantly, it brings interoperability that supports a single view of data. It creates a common language for the entire industry, which will help organisations demonstrate authenticity."

Literacy

Ensuring the workforce is ready for AI will be important in 2025, Banerjee added. "The region’s diverse workforce benefits from data literacy initiatives tailored to different cultural and educational backgrounds. CIOs should support data literacy training and upskilling programmes to foster AI readiness across the organisation, with an emphasis on data interpretation, ethical considerations, and regulatory compliance," he said.

Banerjee suggested developing regionally-tailored, multilingual data literacy programmes that incorporate contextualised data ethics. "Collaborate with universities and technical institutes to build a talent pipeline that supports the region’s evolving data and AI needs," he recommended.

Privacy

"The region is home to stringent data privacy and sustainability regulations, such as Personal Data Protection Act (PDPA), and the upcoming AI guidelines. CIOs must ensure compliance by enhancing data governance practices, particularly around data sourcing, usage transparency, and ethical AI practices. Building compliant data pipelines and audit trails can ensure adherence and help organisations avoid penalties," said Banerjee.

His recommendation: invest in regulatory technology (regtech) solutions that automate compliance and offer real-time monitoring of data practices. "Establish dedicated compliance teams to stay updated on regional regulatory shifts, including cross-border data sharing requirements," he advised. 

Source: Adobe. Sharma.

Adobe found that 64% of consumers in APJ are worried about how much data brands hold about them, and 64% would be more open to granting permission to their data if brands were more transparent.

"As organisations increasingly rely on advanced technologies, the need for responsible innovation has never been greater. This means embedding ethical considerations into AI adoption, ensuring data privacy and security, and developing transparent practices that build consumer trust," said Shashank Sharma, Senior Director, Digital Experience, Korea and SEA, Adobe. 

"Responsible innovation also requires thoughtful leadership—senior executives must champion AI strategies that align with organisational values and customer expectations while empowering teams with the training and governance needed for effective implementation. Companies that adopt a proactive approach to responsible innovation will not only avoid potential pitfalls but also position themselves as trusted leaders in their industries, fostering sustainable growth and customer loyalty." 

The growing demand for data privacy and user control will fuel a surge in the adoption of decentralised identities, iProov predicted. "Empowered by the ability to selectively disclose only the necessary information, individuals will embrace this technology to seamlessly and securely prove their identity or attributes without revealing their entire personal profile. This rise in the use of decentralised digital credentials will empower individuals to confidently navigate the digital world while safeguarding their privacy. It will also create new opportunities for businesses and public bodies to build trust and offer personalised services without running the risk of compromising user data," said the company.

Security 

Andy Ng, VP and MD for Asia South and Pacific Region, Veritas Technologies, has highlighted a trend called data laundering for 2025.

"As generative AI (gen AI) revolutionises industries across Singapore, its insatiable demand for large volumes of high-quality training data is creating new opportunities for cybercriminals," he explained.

"A concerning trend known as data laundering is emerging, where stolen databases are repackaged and sold to legitimate organisations eager to fuel their AI models. The urgency to adopt gen AI and gain a competitive edge has created a perfect storm, making businesses more likely to purchase external data sources without thoroughly verifying their origins."

"This rapid adoption raises significant concerns. According to a recent study, 58% of IT leaders warning that gen AI could expose their networks to cyberattacks, and 81% expressing doubts about their network infrastructure’s capability to manage its data demands," Ng continued.

"As we move into 2025, it is imperative for organisations to stay vigilant against critical security vulnerabilities linked to unverified data or laundered data, or risk being susceptible to breaches and potential liabilities."

Data minimisation is emerging as a critical differentiator and marketable commodity in privacy-conscious markets, he said. "With the increasing focus on synthetic data generation, as highlighted by Singapore's Personal Data Protection Commission, businesses are exploring ways to achieve their objectives while reducing reliance on sensitive datasets," Ng from Veritas said.

"Practices like pseudonymisation, adding noise, and extracting only relevant attributes showcase how minimising data collection can enhance both privacy and innovation. By adopting these measures, organisations not only mitigate risks of breaches but also claim the moral high ground, differentiating themselves as champions of data privacy and sustainability to resonate with consumers and stakeholders alike." 

Pure Storage, on the other hand, focused on data protection for 2025 "as organisations come to terms with being attacked no longer a question of 'if' but 'when'. Several factors are driving this shift in strategy: cybercriminal capabilities being enhanced by AI; increased national legislation; and more stringent compliance requirements from regulatory authorities," the company said. 

"Having a data protection strategy gives organisations a means of resuming business operations quickly in the event of an attack." 

Sovereignty

Source: Aicadium. Poon.

"Transitioning from data hoarding to a more strategic approach to data management—understanding what data is valuable and how to best secure it in which location—will enhance operational efficiency and reduce risks related to unmonitored data," said Phoebe Poon, VP, Product Management, Aicadium. 

"A critical aspect involves strategically aligning workloads with the most suitable platforms, ensuring optimal performance, regulatory compliance, and cost efficiency. For example, on-premises environments may be required for processing highly sensitive or regulated data. Processing data on-premises—storing and managing data within an organisation's physical infrastructure—offers several advantages, including control over data security, reduced latency, cost management, data sovereignty, performance optimisation, legacy system integration, and more." 

Matthew Hardman, CTO, APAC, Hitachi Vantara listed regulations such as Singapore’s Model AI Governance Framework and Indonesia’s data sovereignty laws and said: "Combining AI-driven management tools with Kubernetes for container orchestration, enterprises in the region can deploy applications dynamically while maintaining data sovereignty—a critical need given the evolving compliance landscape for countries in APAC." 

APAC stands for the Asia-Pacific region.

Architectures

Things will have to change at the architecture level, said Couchbase. Data architectures will be redesigned to support AI integration and ensure transparency, said Mohan Varthakavi, VP, AI and Edge, Couchbase. "As AI becomes more integrated into applications, data architectures will be fundamentally redesigned to support AI workloads. Companies will implement new data architectures that go beyond simple record storage to capture the 'intelligence history' and thought processes of AI systems. They will need to simplify complex architectures, including consolidation of platforms, and eliminate data silos to create trustworthy data," he said.

"These evolved architectures will incorporate robust security measures for both data and AI communications. They will prioritise transparency and governance, enabling organisations to track how their data was used in AI training, monitor the decision-making processes of AI systems, and maintain detailed records of AI-generated insights and their underlying reasoning." 

Source: Couchbase. Duddi.

Hardware will dictate how the software is architected, explained Gopi Duddi, SVP, Engineering at Couchbase. "After half a century of CPU-based database design, the massive parallelism offered by GPUs is forcing a complete rethinking of how databases process and manage data," he said. GPUs, or graphics processing units, are the processors that power AI as opposed to CPUs, or central processing units, which power traditional computing.

"Traditional databases were built around CPU architecture, gradually evolving from single to multicore processing. These systems learned to chunk data into smaller bits for parallel processing across multiple CPU cores. However, GPUs — capable of running thousands of parallel threads simultaneously — challenge many of these long-standing architectural assumptions," Duddi elaborated.

"The potential for GPU-powered databases is staggering: operations that traditionally required complex CPU-based parallel processing could be executed across thousands of GPU threads simultaneously, potentially delivering ChatGPT-like performance for database operations."

However, challenges remain, Duddi said. "GPUs don't offer the same reliability as CPUs, and reimagining fifty plus years of database innovation for GPU architecture isn’t an easy feat. Yet the potential performance gains may make this transition inevitable in the coming years. Organisations developing cloud database infrastructure must prepare for a hybrid future where GPU acceleration becomes as crucial for database operations as it is for AI workloads. This shift
will likely reshape cloud offerings, requiring providers to balance traditional CPU-based services with new GPU-accelerated database solutions." 

Source: Hitachi Vantara.
Hardman.

Hardman touched on the need to rethink traditional storage architectures. "Object storage solutions accessible via industry-standard protocols provide scalable, cost-effective platforms for managing large-scale data compared to traditional block storage systems," he said. 

Remus Lim, Senior VP, Asia Pacific and Japan, Cloudera, said that as hybrid environments grow, corporate data footprints will span on-premises, mainframes, public cloud, and the edge. "Businesses need the capability to bring gen AI models to wherever the data resides, and seamlessly move data and workloads across the business, to derive valuable insights and address organisational needs," he noted. 

"As businesses turn to running AI models and applications privately, whether on premises or in public clouds, there will be a greater emphasis on hybrid data management platforms that integrate both on-premises and cloud data sources
for greater flexibility and wider access to diverse datasets while maintaining control, security and governance over model endpoints and operations."

The edge is also important, said TeamViewer, particularly for AI. "By bringing AI capabilities directly to edge devices, organisations will be able to process and analyse in-session data in real-time, creating highly contextual and personalised experiences," said Mei Dent, Chief Product & Technology Officer, TeamViewer. 

"Current remote connectivity solutions are largely reactive, addressing problems after they occur. However, the combination of edge AI and in-session data analysis will enable a more sophisticated approach. Organisations should begin mapping their in-session data opportunities and developing strategies for edge AI deployment. 

"Success will require understanding the unique contexts of different users and use cases, as well as investing in edge computing infrastructure that can support AI workloads. Companies should focus on building preventative capabilities that leverage real-time insights while maintaining user privacy through local processing."

Explore

Read the full set of 2025 predictions from 2-Z at https://www.techtradeasia.com/2025/01/the-techtrade-asia-2024-roundup-2025.html

Hashtag: #2025Predictions