research Market Intelligence /marketintelligence/en/news-insights/research/data-center-interconnection-faces-cloud-native-competition content esgSubNav
In This List

Data center interconnection faces cloud-native competition


Next in Tech | Episode 171: Concerns About Fraud Drive AI Investment


MediaTalk | Season 2
EP 16 - Streamers' Quest for Profitability: Ads vs. Price Hikes


Next in Tech | Episode 170: Corporate Language Analysis in Pride Month

Case Study

A Sports Team Navigates Business Through Disruptive Times

Data center interconnection faces cloud-native competition


In networking, the technology bar is rising. Newer forms of networks as a service, particularly those that 451 Research refers to as cloud-native wide area networks, or WANs, offer connectivity in self-service form with quick turnaround so that networking becomes a fluid proposition rather than an exercise in nailing down circuits. Data center operators should be thinking about a similarly cloud-native approach to interconnection because this is the model that their services will increasingly be compared with.

SNL Image

Some data center providers have already taken interconnection in a cloud-native direction, notably Equinix Inc. with Equinix Fabric. Others such as Digital Realty Trust Inc. have partnered with cloud-native WAN players — in fact, this would be a reasonable way for other data center operators to play catch-up. The point is that networking is going cloud native, joining computing and storage in becoming more agile and invisible to the customer. We believe this is the path that data center interconnection needs to take.

SNL Image

Networking and the cloud-native future

In recent years, many data center providers noted that interconnection could have a role in the enterprise-cloud relationship. The entity that enterprise employees think of as "the network" has changed — rather than a local area network, or LAN, connecting to a server nearby, it is now a virtual private network, or VPN, or even a vanilla internet connection going to an unspecified cloud data center. This new, ad hoc enterprise network has some unsavory security implications, and it invites shakiness in performance and reliability. Private connectivity offers a solution, and leased data centers, with their density of cloud and telco clients, could supply that connectivity via cross-connects.

However, the past few years have also seen the rise of cloud-native WAN, offering on-demand connectivity with the ease of cloud operations. Some providers run their own fiber backbones, while others are over-the-top services that build virtual networks on others' fiber. Either way, what the enterprise sees is that network connections can be created within minutes, or even seconds, via a web portal, and they can be decommissioned just as quickly. As with cloud services, the billing is on a pay-as-you-use basis. These services are the descendants of the software-defined networking concepts that were explored in the early 2010s, and they represent the future direction of networking.

What makes cloud-native networking attractive is that it gives enterprises the thing they've wanted from networking for decades: simplicity. The ease of use promised by cloud-native WAN is a radical departure — as is the agility, where connectivity no longer requires a new contract followed by weeks or months of installation time. This is the new battleground for networking, and therefore for data center interconnection. Several major data center operators have already embraced cloud-native WAN to varying degrees, sometimes via partners. The others should consider roadmaps for bringing their interconnection services to cloud-native form, either organically or through partnerships.

Data gravity

The enterprise networking discussion often focuses on cloud connectivity, but another dimension is emerging in the realm of data. The industry has long been aware of the concept of "data gravity" — the tendency for applications and services to move closer to data. Meanwhile, the amount of data being gathered is always rising, and the growth of machine learning is making that data more valuable. Combine that trend with the concept of data gravity, and the "location" of the data becomes a serious long-term consideration. "Inside the cloud" might not be the best answer.

To see why, consider the practicalities of moving that data. A data lake in the petabyte range could take days to extract. Even moving fractions of that data can incur costs, as the public clouds charge egress fees for data that is exported onto other networks. These charges are an easy target for cloud players. New storage services from Cloudflare Inc.PacketFabric Inc. and Rackspace Technology Inc. all offer ways to avoid egress charges — and all three emphasize that they have networks in place to bring applications to that data. Some data center operators are likewise pitching colocation as an alternative — and again, this argument will be more competitive if the associated networking is cloud native.

Egress charges might sound like just an irritant, but for some enterprises, they have become intolerable. In research for our recently published Hybrid Cloud Storage Maturity Model 2021 report, 34% of enterprises said egress charges have affected their use of cloud storage. Of those, 27% said they have repatriated data on-premises as a result, and 37% moved data to a different service provider that doesn't charge for egress.

The point is that even if the enterprise is moving only a fraction of its data, these charges can add up. Data center operators can play off of this by advertising their facilities as an alternate storage hub, but this argument again comes back to networking because users and applications will need to connect into this hub. Data center operators making this argument would do well to also offer simplified networking.

This article was published by S&P Global Market Intelligence and not by S&P Global Ratings, which is a separately managed division of S&P Global.

Gain access to our full news & research coverage and the industry-specific data that informs our insights.
Request Demo