Credit: Dreamstime
As data, applications, and workloads continue to move to the cloud, demand for infrastructure-as-a-service (IaaS) networking is surging. The market for cloud-based IaaS networking will reach $19.4 billion in total global revenues this year, according to IDC, with a compound annual growth rate of 28% projected through 2026.
Increasing cloud-native application architectures, distributed workloads, and their respective integration needs are the biggest drivers of IaaS cloud networking adoption, says IDC analyst Taranvir Singh.
"Traditional network architectures, transports and operational models are no longer able to meet the growing requirements and objectives of enterprises’ modern networking needs," he says. "Networks need to be aligned with cloud principles."
IDC describes IaaS cloud networking as a foundational networking layer that allows large enterprises and technology providers to connect data centers, colocation environments, and cloud infrastructure.
With IaaS networking, the network infrastructure and services are scalable and available on-demand, provisioned and consumed just like any other cloud service. That makes this infrastructure more scalable and agile than traditional approaches to networking, according to IDC.
Direct cloud connects/interconnects is the largest segment of IaaS networking, accounting for more than half of all IaaS networking revenue. The four other major segments of the IaaS networking market are cloud WAN (transit), IaaS load balancing, IaaS service mesh, and cloud VPNs (to IaaS clouds), according to IDC.
Cloud WAN, which includes cloud middle-mile and core transit networks, is the fastest-growing segment of IaaS networking, with a forecasted five-year compound annual growth rate of 112%, says IDC. IaaS service meshes are also expected to see strong growth, with a forecasted five-year compound annual growth rate of 68%.
Customers with hybrid environments – both on-prem and in the cloud – are the primary users of IaaS networking services from Alkira, a cloud networking company founded in 2018 by the same team that was behind Viptela, which helped pioneer the SD-WAN market before it was acquired by Cisco.
"Right now, if you wanted to leverage different things from different clouds, you'd have to create a DIY cloud infrastructure," says William Collins, Alkira's principal cloud architect. Alkira uses the hyperscalers' cloud infrastructure, but abstracts all the complexity, he says. "It allows you to have the best-of-breed from all the different cloud providers. You can choose how you want to do the cloud connectivity."
Best practices for IaaS networking
The top five vendors in the IaaS networking space are the same as the top five hyperscalers: Amazon Web Services, Microsoft, Google, Alibaba, and Tencent. Together, they account for 86% of the total market spend. Amazon alone accounts for more than half.
For enterprises, it can make sense to get their cloud networking from their cloud providers, as it dramatically simplifies deployment and management. Companies that use multiple cloud providers may opt for a separate IaaS networking provider, or a third-party IaaS networking management layer.
However, some hyperscalers are stepping up. For example, the recently launched Google cross-cloud interconnect helps establish dedicated connectivity between Google’s cloud and other cloud providers, says IDC's Singh.
When selecting an IaaS networking vendor, Singh suggests that enterprises look at the breadth and depth of their IaaS networking services portfolios, their partner network, supported APIs, integrations, and egress costs.
A common mistake is choosing a cloud provider whose service offerings don't align well with an organisation's needs, he says. In particular, companies should be careful to ensure that the vendor offers all necessary APIs and integration tools.
Another common mistake is not including IaaS networking in the overall cloud strategy, he adds. To help address this issue, he suggests that enterprises set up cross-functional teams to set overall strategy and collaboration between business units.
Finally, as with any on-demand service, there's always a risk of paying too much if usage suddenly spikes.
"Developers, DevOps practitioners, and platform teams need to step up and assume larger roles in decisions related to network connectivity," says Singh. By working together, they can find the optimal balance between costs and latency requirements.
Cost is potentially the number one downside of using IaaS cloud networking, agrees Eric Helmer, CTO at Rimini Street, a technology service provider.
"It can get out of hand very quickly," he says. "The important thing is to not allow that wild, wild west where every department can fire up their own administrators and you have no idea what’s going on until the bills get bigger and bigger."
Enterprises need to work out a system for getting new business use cases approved and monitored, with alerts in case usage suddenly spikes unexpectedly.
"Once they're there, it's hard to rein them in," Helmer says. That's because nobody wants to accidentally shut down a business-critical function. "So governance is critical."
Another challenge of IaaS networking is understanding who's responsible for security. "People think just because they move over to an IaaS provider, that security is just taken care of," says Helmer. "And there are some layers of security that the provider provides, but it's not everything."
Companies need to understand what areas of security they are responsible for. That can include getting into the network, role-based security, and encryption. "You have to make sure you're not making any assumptions," he says. "There are a lot of attacks and breaches happening when people assume that the providers are handling those layers of security and they aren't."
Alongside security, enterprises should also keep an eye on compliance and privacy requirements.
"A lot of times, security and compliance changes when you move things externally," Helmer says. "So if you're looking to go into IaaS, make sure your governance and risk and compliance teams are involved as well."
There are also operational risks that go along with IaaS networking. For example, can the provider handle your peak usage? "When you test out an IaaS provider, do full load testing and performance testing in the worst-case scenarios, high-peak values, to make sure that what you get will work for you," Helmer says.
And, when those peak loads are predictable, companies can save money by only provisioning for what they need at the time they need it, he adds. For example, a company that has high traffic demands at the end of every quarter can plan ahead for extra capacity at that time, and lower capacity during the rest of the year.
"If you leave it running 24 hours a day, for the worst-case scenario, you'll find that the cloud does not have a cheaper model than doing it yourself," he says. "So take advantage of the elastic model."
Finally, enterprises should be prepared to hire new staff when they switch to IaaS networking.
"A lot of people think they’ll be able to free up some staff, and in many cases you can," he says. "But also, in many cases, you’ll need additional staff that have those specific IaaS skill sets because IaaS doesn’t run itself. You still have to monitor it and maintain it and manage it."
But some vendors do make management easier than others, he adds. When these services first appeared, the interfaces were often unnecessarily complex. "You had to be a rocket scientist to figure out how to architect a service and setup failovers and availability zones."
In particular, enterprises need to be able to get connections up and running quickly to their clouds, on-prem data centers, and SaaS platforms. "If it takes months and months and months, and setting up firewalls, and security, it takes away the advantage over doing it yourself," Helmer says.
Some providers also do a better job with transparency than others. Enterprises need to be able to easily get performance-related metrics and help with continuous rightsizing to manage costs, he says.
