The world may be a lot more connected today with 5G networks and subsea cables covering large swathes of the globe. Where they don’t reach, satellite connectivity may come in handy.
Yet, setting up a complex network that spans countries and continents to run mission-critical applications is still a challenge for many organisations.
In Asia-Pacific, for example, connectivity options can be uneven in quality and may often require upgrades and modernisation for new apps that demand low latency and high bandwidth, said Eric Wong, president for Asia-Pacific at Expereo, which offers global connectivity services to multinationals.
“Only a few companies believe their networks are fully ready to support future needs, and over a third acknowledge that their networks are unfit or will soon need upgrading or replacement,” he noted, referencing a recent study Expereo conducted.
“If anything, having networks that do not scale flexibly on demand, or deliver the performance required, will end up hindering those companies from implementing AI projects, and impact the business’ ability to grow in an increasingly cloud-based digital business environment,” he stressed, in this month’s Q&A.
NOTE: Responses have been edited for brevity and style.
Q: Driven by remote work during Covid, many businesses modernised their networks with SD-WAN (Software-Defined Wide Area Network) and other setups to deliver improved connectivity. What is driving their network strategy now?
A: As much as Covid-19 was a terrible thing for everyone, an unseen benefit was that it forced businesses to take a long hard look at digitalisation and transformation.
With the pandemic firmly in the rear view, the priorities of today have shifted. Instead of looking at connectivity to drive business continuity, businesses see connectivity as a means towards building a more agile network that is not just scalable, but also resilient.
Businesses today need networks and connectivity that can support the demands of modern technologies like AI and its related workloads.
More importantly, network performance is more critical than ever to support the needs of day-to-day business. Employees must be enabled to connect seamlessly to the applications and resources they need, regardless of location or device. This requires a reliable network that can deliver both high bandwidth and low latency.
In a recent study, we found that only a few companies believe their networks are fully ready to support future needs, and over a third acknowledge that their networks are unfit or will soon need upgrading or replacement.
If anything, having networks that do not scale flexibly on demand, or deliver the performance required, will end up hindering those companies from implementing AI projects, and impact the business’ ability to grow in an increasingly cloud-based digital business environment.
Q: In the past 18 months, AI has triggered more growth in a region that was already experiencing fast data centre growth. How does connectivity come into the picture, since much of the computing work is happening in the data centre itself?
A: While we cannot deny that the data centre is where AI workloads live, we cannot ignore the role that connectivity plays in enabling the broader AI ecosystem.
AI models require massive datasets to be trained effectively, and these datasets can reside in the cloud, sometimes across different data centres.
Moving all this data from one data centre to another requires high-bandwidth, low-latency connectivity. As such, having a robust and performant network is critical towards the training and deployment of AI models.
Research from Forrester predicts that by 2025, 60 per cent of organisations in the region will use AI language models tailored to local markets. These innovations demand high-speed, low-latency, and reliable connectivity, prompting significant investments in advanced networks and data centres.
Edge computing models also mean that in some cases, AI processing does not just live within centralised data centres. Edge devices such as those used in IoT applications require reliable connectivity to communicate with central servers and the cloud effectively and efficiently.
This is particularly important as applications such as autonomous vehicles and smart buildings increasingly come into play.
Q: Where in Asia-Pacific, in particular, are the “bottlenecks” for connectivity, when it comes to delivering emerging business and consumer applications?
A: In Asia-Pacific, there are varying levels of infrastructure development and digital maturity, across different countries. While major investments are being made in digital infrastructure, there are some bottlenecks and roadblocks that persist in emerging markets and underserved populations.
Providing basic Internet access can be a challenge, never mind being able to ensure the availability of the high-bandwidth, low-latency connectivity required for emerging technologies. Factors can include the lack of, or aging, connectivity infrastructure.
Vietnam, for example, has five major existing cables which link the country to the global Internet. These cables are aging, with several reported outages over the past couple of years, although Vietnam has announced plans to deploy two new undersea Internet cables by the end of 2025.
In some areas, the lack of adequate infrastructure in rural regions also restricts access to essential services like e-commerce or telehealth solutions for consumers residing outside major cities.
On top of that, geographical challenges must be taken into consideration, such as the potential impact from earthquakes, or the number of hops required to connect islands.
Additionally, the region’s diversity presents unique challenges in terms of regulatory compliance and access to skilled professionals. Navigating the different regulations while ensuring adherence to local laws can be a complex process.
Moreover, the availability of skilled professionals to build, manage, and maintain these advanced networks is crucial but continues to be an issue in markets like Asia-Pacific.
For enterprises to solve these issues, they should establish strategic partnerships with technology providers that can help optimise network design for resilience, as well as ensuring uptime and performance.
Partnerships like these also enable insights into market dynamics, local consumer behaviour, and regulatory environments. This approach can facilitate smoother market entry and reduce operational risks.
Q: Are there tailored connectivity options that certain industries should consider, given their unique circumstances?
A: Connectivity cannot be a one-size-fits all solution and should be tailored to suit the specific requirements of a particular industry and use case.
For instance, industries like healthcare are now deploying telehealth applications, require high-bandwidth and low-latency connectivity to ensure that patients get the timely medical advice that they need. Taking it one step further, critical applications such as remote surgery must be able to function reliably and without interruption.
Industries that operate in remote areas where it is extremely cost-inefficient to deploy last mile connectivity infrastructure can benefit from technologies like low-earth orbit satellites.
This can help bridge connectivity gaps and minimise the impact of terrestrial infrastructure limitations, ensuring business connectivity even in challenging and remote environments.
Businesses must take a strategic approach to connectivity. Businesses should assess their specific needs and challenges, considering factors such as bandwidth requirements, latency tolerance, security needs, and geographic location.
