While AI (artificial intelligence) is currently the raging trend in the technology sector, its vast potential can only be realized if a company’s data infrastructure is secure.
This is according to Arnie S. Alvarez, chief technology officer for the enterprise business group at Huawei Technologies Philippines, during the recently concluded VST ECS CxO Summit 2024 at the Shangri-la Resort in Boracay Island, Aklan.
Citing a study by consulting firm McKinsey, Alvarez said 50% of business organizations worldwide used AI in 2022 with $13 trillion expected to be the AI’s contribution to the global GDP by 2030.
“For enterprises, data is increasingly becoming a production factor. Driven by AI technology, it makes applications smarter and business more efficient,” Alvarez said.
In the era of AI where computing power, algorithms, and data converge to drive unparalleled intelligence, data storage assumes paramount importance, the tech executive said.
“As AI foundation models redefine enterprise applications, the need for scalable, sustainable, and flexible storage solutions becomes imperative,” he stressed.
Alvarez noted that large AI models are injecting new vitality into data and further unleashing its value. “Powered by AI, data is becoming the new oil and the key element for driving new productivity,” he said.
He said data infrastructure carries data, so the strength of a country’s data infrastructure will determine its ability to stay ahead in the AI era. Thus, he said the development of AI is largely dependent on data infrastructure.
“Today, the rapid construction of infrastructure for large AI models is entering a new phase, shifting from just stacking computing power to combining computing, storage, and transport power, which forms the basis for high-quality data mining and advanced storage,” Alvarez said.
The Huawei official said data infrastructure is becoming the cornerstone of large model development in three major ways.
“First, it stores the data needed to fuel large AI models, which are only as good as the data they are trained on. The quantity and quality of data determine the learning ability of large models.
“Second, data infrastructure solutions tailored to AI critically accelerate the training speed and computing power utilization of large models and enable the application of AI across a vast range of industries.
“Finally, data infrastructure plays a critical role as a line of defense for data. The importance of building AI-ready data infrastructure cannot be overstated. It is an essential step in the development of AI and a strategic imperative for countries and organizations seeking digital transformation and innovation,” he said.
Alvarez said that an AI-ready data infrastructure needs to have the following features:
- First, openness and interconnection. As data resources become increasingly diverse, data infrastructure will need to eliminate 37 information silos and enable seamless data mobility across departments, domains, and regions, to form an open and shared data ecosystem. There needs to be an efficient data exchange platform, unified data standards and interface specifications, and data resources’ flexible flow and value creation.
- Second, intelligence and agility. Data infrastructure for the AI era must be capable of intelligent processing: using automation tools and algorithm models to efficiently clean, integrate, and mine massive amounts of data to provide accurate and real-time data services for AI applications. In addition, it must be nimble and scalable enough to adapt to service requirement changes, and offer data services on demand and dynamically adjust them.
- Third, data resilience and regulatory compliance. Data infrastructure must provide data resilience and privacy protection while realizing the value of data. Data infrastructure must embed strict data access control mechanisms and use advanced encryption and anonymization technologies to prevent data leakage, tampering, and abuse. In addition, its design and operation must comply with local laws and regulations, and must include a sound data lifecycle management mechanism to ensure legal and regulatory compliance of data processing activities.
- Finally, sustainability and energy saving. With the explosive growth of data, the energy consumption of data infrastructure will have a massive environmental impact. Green data centers will be critical to a green future.
As the core of data centers, Alvarez said data infrastructure can achieve sustainable development by improving energy efficiency and reducing carbon emissions.
“This will require persistent technological innovation and management optimization. AI-ready data infrastructure is already a major research area in the industry and a direction that many countries are investing heavily in,” he said.
