Addressing the Four Critical Pillars of Enterprise Connectivity in 2025

PinIt

For enterprises embracing multi-cloud strategies or deploying AI for mission-critical tasks, connectivity is no longer just a utility; it’s a strategic asset that determines their ability to innovate and compete.

With artificial intelligence (AI) booming and quantum computing now firmly on the horizon, we’re rapidly approaching an era in which a company’s technological ambition is likely to outgrow its network infrastructure. According to Gartner, expenditure on cloud infrastructure and cloud applications surged by 20% in 2024 and is expected to climb even higher in 2025. Yet, despite this investment, IDC reveals that almost a third (29%) of organizations in Europe say networks aren’t ready, with a staggering 64% claiming that network latency and performance represent a “moderate” or “significant” challenge to their operations.

The Internet has arguably been the greatest economic catalyst of modern times, but it was originally conceived as a “best effort” medium. When the idea for a federated “network of networks” was coined back in the 1970s, the use cases on which our society is built today were scarcely being imagined, let alone cloud computing, the Internet of Things (IoT), or the wholesale adoption of AI and Large Language Models (LLMs). The Internet’s primary purpose was – and in many cases, remains to be – moving data from point A to point B with minimal concern for the speed, security, or reliability of the journey. While this design has been sufficient for general use, enterprises that rely on latency-sensitive applications such as AI inference and cloud computing or data-intensive workloads such as AI model training are now encountering significant limitations. Unacceptable latency, unpredictable pathways, a lack of bandwidth scalability, and a lack of control over data flows are turning connectivity into a chokepoint rather than an enabler, jeopardizing the potential of advanced digital solutions to drive productivity and innovation.

It’s time to rethink connectivity in 2025. This article explores the four critical pillars of enterprise connectivity—private, public, cloud, and cloud-to-cloud. Each pillar serves distinct needs, from interconnecting internal systems to supporting complex cloud-based workloads. By optimizing these layers and addressing challenges like cost, resilience, security, and latency, enterprises can build the robust digital foundations needed to unlock their ambition in the coming years.

See also: Revolutionizing Connectivity: Navigating the Future of Multi-Access Edge Computing

The Four Pillars of Enterprise Connectivity

For enterprises embracing multi-cloud strategies or deploying AI for mission-critical tasks, connectivity is no longer just a utility; it’s a strategic asset that determines their ability to innovate and compete. This connectivity needs to be designed in a way that ensures the fundamentals of bandwidth, latency, resilience, and controllability – all while keeping costs in check. Let’s review the four current pillars of connectivity:

Private connectivity: The connection of locations within the company, providing access to internal resources for a workforce potentially spread around the world. Enterprises today operate in a global environment, often with multiple isolated network environments in various countries on different continents. The challenge here is to ensure resilient, low latency, high bandwidth, and flexible access between locations.

Public connectivity: This encompasses the connection of the company network to vendor, partner, and customer networks, as well as to Internet service providers serving their remote workforce. This needs to cover local Internet breakouts and access to services (XaaS) in all locations where the company has staff.

Cloud connectivity: To secure business-critical applications and data flows, it is essential to have dedicated connections to clouds in order to control bandwidth, latency, and resiliency. For the best performance of cloud technology, a company should ensure low latency access by using local cloud regions at every company location and ensuring direct connectivity to the cloud service provider’s own connectivity solution.

Cloud-to-cloud connectivity: Connections between clouds should bypass the public Internet to secure the same benefits as for connections to clouds. This is done by directly connecting the private connectivity solutions of the various providers in use. Here, there is the added challenge that different cloud service providers use different types of infrastructure, and these were never designed to be interoperable.

Most organizations will use a combination of the above depending on their industry and requirements. The question then becomes, how do they create a holistic, frictionless network environment that affords them the right level of control, performance, and resilience?

Overcoming “Best Effort” IP Transit

The days of “plug and play” internet connectivity are far behind us. A forward-thinking, globe-spanning enterprise simply cannot afford to trust their connectivity to chance or rely too heavily on third parties or basic IP transit. There are several ways to approach this, each with their own advantages and limitations.

A common starting point is connecting locations via an Internet Service Provider (ISP), which often relies on third-party networks (IP transit) to transport data. However, this method—based on the Internet’s layer 3 routing—is inherently flawed. The lack of control over data pathways introduces latency, territoriality issues, and potential security risks, with little control and virtually no ability to optimize. These limitations highlight the need for enterprises to explore alternative solutions that go beyond the “best effort” nature of IP transit outlined earlier.

At the foundational layer 2 of the Internet model, dense wavelength-division multiplexing (DWDM) offers improved bandwidth by transmitting multiple data streams over fiber. While typically the domain of wholesale network providers, some enterprises lease dark fiber to run their own high-capacity layer 2 networks. Alternatively, multiprotocol label switching (MPLS) on layer 2.5 provides secure, dedicated connections. However, owning and managing an MPLS is often too costly and inflexible for most enterprises, making it an impractical choice for scaling global operations.

More recently, software-defined wide area networks (SD-WAN) and Secure Access Service Edge (SASE) have gained traction as flexible and cost-efficient layer 3 solutions. These technologies excel in low-bandwidth use cases and offer security through encrypted overlays on existing infrastructure. That said, they struggle with heavy workloads requiring bandwidth beyond 10 Gbps, limiting their application for larger offices or data-intensive environments. For most enterprises, achieving a robust network will require combining these technologies with tailored connectivity solutions to meet diverse operational needs.

Finding the Connectivity Sweet Spot

Many enterprises find that combining multiple technologies and network service providers delivers the best connectivity for diverse use cases. For private connectivity, outsourcing to a Network-as-a-Service (NaaS) provider offers the advantages of a resilient, fully redundant design without the complexity and cost of managing a proprietary MPLS network. This approach can be paired with SD-WAN or SASE for smaller locations and workloads, ensuring cost-effective and flexible connectivity. For public connectivity, direct peering at Internet Exchanges (IXs) is a proven way to reduce latency and improve performance compared to IP transit. Additionally, an IX that offers API capability has the added advantage of enabling flexible scalability of the connections to other networks as required. To maximize resilience, the IXs used should be data center and carrier neutral, and connections should involve geo-redundancy across a distributed data center infrastructure in each market where the company is active – a model that avoids vendor lock-in and enhances resiliency.

Cloud and cloud-to-cloud connectivity also benefit from direct, private connections to cloud providers – examples include AWS Direct Connect or Microsoft ExpressRoute. By utilizing cloud exchange capabilities at IXs, enterprises can maintain low-latency access to critical cloud regions while using remote connectivity to other regions for less time-sensitive workloads, such as AI training or backups. Cloud-to-cloud communication is best managed with highly scalable virtual cloud routers, which enable direct, secure, and low-latency connections between providers while ensuring interoperability across different cloud platforms. These solutions simplify operations by eliminating the need for additional hardware, making them an ideal choice for scaling multi-cloud environments.

Multinational enterprises are increasingly relying on a mix of Internet and Cloud Exchange platforms as well as Network-as-a-Service (NaaS) providers to optimize their global IT connectivity. By prioritizing data sovereignty, security, and high-performance interconnection with external networks, organizations can create tailored solutions to meet diverse needs such as product development, cost reduction, or supporting a remote workforce.

The Boston Consulting Group estimates that commercial data sharing could unlock value equivalent to 2.5% of global GDP. Effective data exchange would not only address challenges in 2025 but also preemptively solve challenges likely to emerge in the years ahead.

Dr. Thomas King

About Dr. Thomas King

Dr. Thomas King has been Chief Technology Officer (CTO) at DE-CIX since 2018, and a Member of the DE-CIX Group AG Board since 2022. Before this, King was Chief Innovation Officer (CIO) at DE-CIX, starting in 2016. He has been instrumental in his role at keeping DE-CIX at the forefront of technological development of Internet Exchanges, establishing DE-CIX as a neutral Cloud Exchange, pushing the boundaries of what is possible in terms of high-bandwidth access technology and security solutions for IX platforms, and trailblazing the automation of IX services with the implementation of patch robots, the development of the DE-CIX API, and overseeing the DE-CIX self-service customer portal. Thomas King has also overseen the technical implementation of the international expansion in markets spanning from North America to Europe, the Middle East, India, Southeast Asia and most recently Africa.

Leave a Reply

Your email address will not be published. Required fields are marked *