​Accelerating Next-Gen Data and AI Services in the Cloud

Once an emerging trend, the cloud has evolved into an indispensable component of modern IT deployments, helping enterprises to streamline their operations, enhance scalability and drive innovation across industries.

As businesses of all sizes turn to the cloud, cloud computing has grown rapidly into a booming industry today. Indeed, the latest Gartner forecast noted that worldwide public cloud end-user spending is set to grow by 21.7% and reach nearly US$600 billion by 2023.

More than one cloud
Despite the indisputable role of the cloud, only a handful of businesses have adopted a cloud-only approach. Instead, as organisations gain competence and familiarity with the cloud, they have turned to hybrid cloud deployments to meld the scalability that it offers with the control and cost stability of on-premises systems.

Yet others have adopted a multi-cloud strategy, harnessing the strengths of various cloud platforms to create a unique IT stack or to complement their existing IT infrastructure. The latter is achieved by selectively leveraging the best-of-breed capabilities from multiple public cloud providers.

The preference for hybrid cloud or multi-cloud can be attributed to the inherent challenge of finding a single cloud provider to fully meet the diverse needs of an organisation's operations. Then there is also the concern about platform lock-in or mitigating the impact of black swan events that culminate in cloud outages.

But whether enterprises opt to distribute their computing workloads across on-premises deployments and the cloud or build a comprehensive, multi-cloud strategy to leverage multiple cloud providers, the result is a smorgasbord of inter-cloud connectivity.

Suddenly, businesses find themselves requiring new capabilities to control, secure and manage their data as it traverses the network across private and public cloud resources, data centres and the corporate network.

Supporting data, AI workloads
The importance of networking is not limited to traditional workloads but is crucial for cutting-edge AI and data analytics deployments, too. Much has been written in recent months about ChatGPT, which triggered a new wave of interest in AI globally.

As organisations scramble to implement AI strategies and train machine learning (ML) models, more data than ever will flow through computer networks. And the cloud is ideally suited for analytics and AI workloads.

• Data storage and management: Training ML models requires large datasets. The cloud makes it easy to access the digital storage and management capabilities needed to handle the data for training AI systems.
• Processing power: The cost of purchasing graphic processing units (GPUs) can be prohibitive. The cloud opens the door to rent access to GPUs as required.
• Integration and interoperability: Cloud services often come with a rich array of APIs and integration options to support the seamless transfer of data across cloud services or compatible third-party cloud platforms.
• Data ecosystem: The top data platforms already run in the cloud, including enterprise-focused data ecosystems such as SAP Datasphere and Databricks Lakehouse.

Depending on implementation strategies, data could flow between clouds, the corporate network to public cloud platforms, or between data centres to the cloud. Regardless of one's mix, data connectivity between these points needs to be actively managed.

Highway to the cloud
So how can we get better networking for dependable access to the cloud? The easiest way to bypass the limitations and the inconsistency of the public internet is to leverage private connectivity. But connecting private networks directly to major cloud service providers is unlikely to be cost-effective.

An alternative is to work with a trusted provider for cloud connectivity. Enterprises can hence leverage an existing MPLS or Premium Internet connection to the cloud provider of their choice – all the while enjoying greater security and reliability than they would over the public internet.

Apart from benefitting from a reliable, high-speed infrastructure designed for cloud access, businesses also enjoy enhanced flexibility. This allows them to swiftly expand their network connectivity in line with business growth and to accommodate new services.

StarHub Cloud Connect is designed to help enterprises bypass the public internet and seamlessly access their cloud-based applications and data assets from both corporate networks and data centres.

Shared Cloud Connect offers access to the cloud through a shared high-speed virtual private access network, while Dedicated Cloud Connect provides dedicated transmission between customer premises or data centres to leading cloud service providers.

In conclusion, optimised cloud connectivity is an integral component of modern IT strategies, giving enterprises the secure and flexible connectivity they need to empower their IT and AI initiatives.