The Alibaba Cloud-commissioned survey by NielsenIQ, The Next-Generation Cloud Strategy in Asia, reveals continued interest in the use of cloud services among respondents in the region. The survey also revealed changing cloud strategies among respondents in the study with 39% having switched to hybrid cloud citing improved operational efficiency (71%) as a key benefit. In addition to security (70%), businesses tend to choose hybrid cloud for customised cloud services.
Drivers of change and the status quo
Which comes first: maturing data literacy or data democratisation? Perhaps it matters less in the context of organisations finally being able to ethically monetise data while observing privacy and the protection of data.
Perhaps what is more practical is the integration of non-aligned technologies around business outcomes.
Why hybrid works for Asia
Han Chon, managing director for SEATHK at Nutanix says organisations in Asia have caught on to the idea that hybrid cloud is the way to go for addressing their specific business requirements. The COVID-19 pandemic created a lot of uncertainty, and businesses learned the hard way that having a backup plan and enabling remote work is crucial. Now that we are recovering and moving forward, organisations are realising that cloud is an operating model and not a destination.
He cited Nutanix’s fifth annual Enterprise Cloud Index report wherein 67% of surveyed respondents from Singapore are said to have multiple IT infrastructures, including a mix of private and public clouds, multiple public clouds, or a combination of on-premises and hosted datacentres.
"Whether it's utilising edge computing or leveraging the scalability of public clouds, organisations are looking for that sweet spot to deliver the best possible service levels for their unique business requirements," he added. He claimed that although the interpretation of hybrid cloud may vary slightly depending on each organisation's specific needs, there is a clear consensus that it is the preferred approach to tackle their unique business challenges.
Forrester VP and research director, Charlie Dai, observes that the interpretation of “hybrid cloud” varies across geographies, industries and even organisations.
"Most firms would consider hybrid cloud as the hybrid deployment in both public cloud and private cloud environments, while some firms would add edge deployment into the overall picture; some may consider multi-cloud as part of the hybrid cloud; and in mainland China, most firms would use “private cloud” instead of “on-premises” to describe their adoption of cloud in their own data centres as part of the hybrid deployment."
Charlie Dai
Data democratisation's impact on infrastructure strategies
The democratisation of AI will accelerate digital transformation within the enterprise and lead to business efficiencies, even in a tight fiscal environment," says Chon. He asserts that this will be true especially with the democratisation of generative AI and its associated compliance issues (with public models).
"It is also my view that generative AI or foundational models (e.g., large language models like GPT or LLMs) will become more ubiquitous in the enterprise. The new demand will be for compact, private, and explainable LLMs," he posited.
He further opined that generative AI would need accelerated computing, such as GPUs and other special-purpose hardware, in addition to new CPU designs with support for neural computation (e.g., Apple’s M1/2).
To train these models, he asserted that multiple servers (with accelerated computing) communicating on specially tuned network fabric will be needed. "Model training and fine-tuning also need special purpose databases (e.g., vector databases), data pipelines, and unified storage, along with lifecycle management and governance," Chon added.
Dai says the democratisation of AI will significantly drive the infrastructure needs for computing, storage, and networking; however, this will also bring challenges of management complexity towards cloud practitioners.
The most optimum solution
Asked what are the operational conditions for which a hybrid cloud IT is the most optimum solution, Chon believes that the diverse requirements of different applications play a significant role. He acknowledges that certain applications need to be or perform better when hosted on-premises, while some applications may benefit from the flexibility and scalability offered by public clouds.
"By adopting a hybrid cloud approach, organisations can strike the right balance between control and performance, tailoring their infrastructure to meet the unique needs of each application," he continued. "Another operational condition that favours hybrid cloud adoption is the increasing complexity of modern applications."
Hon posits that with the growing popularity of cloud-native applications and the exponential growth of data, having robust IT is crucial.
"Hybrid cloud environments enable organisations to effectively support these complex applications and efficiently manage increasing amounts of data. By leveraging the scalability of public clouds alongside private or on-premises infrastructure, organisations can ensure data security and compliance while meeting their application requirements."
Han Chon
Chiming in, Dai says the public cloud has the advantage of self-service for minimised CAPEX, elasticity for dynamic workloads support, and the synergy of a wide range of technology and application services for a faster innovation pace.
"On the other hand, private cloud/on-premises deployment allows regulatory compliance and is easier for legacy integration, with better support for data gravity and latency-sensitive applications. The more needs in both areas that business operations require, a better fit that a hybrid cloud approach will be," he added.
The automation equation in operations
Chon says automation is the foundation for managing complexity (from apps to infra) and deriving efficiencies in an enterprise. It is a key tenet of a hybrid cloud strategy. "In my view, AI (and generative AI) will need to be a part of this hybrid cloud strategy to accelerate digitisation. There are a few dimensions to embedding AI in the hybrid cloud strategy," he added.
According to Forrester principal analyst, Carlos Casanova, the artificial intelligence for IT operations (AIOps) platform market is moving faster than can really be imagined, both in terms of vendors as well as the capabilities that make up AIOps solutions. "This was driven by a continued increase of unmanageable data volumes and continued desire for business insights," he continued.
Dai believed that intelligence powered by AI is becoming increasingly critical for automation – the automation is shifting from a rule/policy-based approach into a more predictive and cognitive approach that powers AIOps, digital process automation, and RPA.
He believed that AI is changing the engagement model of most technology and application domain, such as queries using natural languages into distributed databases and cloud data warehouses and data lakes, content generation/generative AI using computer vision and speech analytics services powered by foundation models, as well as next-gen development powered by TuringBots.
"And all the training and inferencing of these models will need a hybrid cloud strategy," he concluded.
