Futuriom Wraps Up Cloud Infra Trends for 2024!

As always, it was an interesting year in cloud and AI infrastructure, as 2024 brought many twists and turns in key markets. Whether it was cloud service providers spending hundreds of billions of dollars on AI infrastructure, Kubernetes rising to the fore as the de facto cloud computing environment, or cloud security headaches, there were plenty of things to follow.

At Futuriom, we try to keep our focus on core infrastructure services and investment. The leading buzzword of the cloud infrastructure market was, of course, AI, but AI is a general term and has many nuances. It includes the market for building infrastructure for large language models (LLMs), a market for private AI infrastructure and retrieval-augmented generation (RAG), as well as AI for operations, or AIOps. We’ve been following all of these evolving markets, and we have expanded our coverage significantly to track the impact that AI and next-generation applications are having on infrastructure.

In 2024, we launched our first Networking Infrastructure for AI report, which highlights the need for faster, more scalable networks to support the unique demands of AI on both the training and inference sides. The implication for the networking industry is that architectures are required to support the needs of AI clusters, which connect many GPUs together. This new market has become a focal point for many networking players, as Ethernet emerges as a technology rivaling InfiniBand, which has been a technology traditionally driving back-end networks for high-performance computing.

One of the takeaways of 2024 is that Ethernet is rapidly evolving to become the AI networking technology that can deliver the benefits of scale and standardization. Large public companies driving innovation in the networking market include Arista Networks, Cisco, F5, HPE, Juniper Networks, and NVIDIA. Executing in this market will be key to the success of the forthcoming HPE-Juniper linkup, which we'll be watching in 2025. We also expect many startups to contribute innovation.

Cloud Costs Come to the Fore

After years of huge growth, the major platform-as-a-service (PaaS) cloud providers, such as Amazon, Google, and Microsoft, are now the subject of another discussion: their costs. The dialogue about cloud cost management (CCM) is increasing. Based on what end users are telling us in surveys and conversations, they are really struggling to get a handle on managing these costs.

The primary challenge in CCM and FinOps is complexity. Cloud services are neither cheap nor simple; in fact, in many cases they turn out to be expensive and complex. Not only are there many platforms (AWS, Microsoft Azure, Google Cloud, and IBM Cloud, just to name a few), but the areas that need to be managed are also numerous—showback and chargeback; reserve instances and commitment management; workload management; Kubernetes right-sizing; storage management; networking costs, etc. As practitioners increasingly look at hybrid cloud architectures, they are also examining how they can consolidate cloud spend, whether that’s by implementing better cloud visibility and automation tools or consolidating platform functions to automate service and data consumption.

In the second year of our CCM and FinOps report, we investigated the major considerations that enterprises decide on when crafting their cloud adoption strategies. We also looked at the macro drivers of cloud cost complexity and how to establish a FinOps governance model to help enlist stakeholders across business groups, tech teams, and the finance department to address cloud cost growth holistically. If you are new to the FinOps world, we invite you to download and read this report as a primer on how FinOps is evolving to meet new challenges.

As generative AI-fueled innovation moves from evaluation and proof-of-concept stages, enterprises must now take a much closer look at the costs of running public cloud infrastructure while comparing it to alternatives, such as on-premises and hybrid cloud architectures.

Psyching Out Cisco: Most Popular Posts 2024

One of our interesting discoveries of 2024 is that readers are as obsessed with Cisco as we are. We continue to hear from Cisco employees past and present who grumble about the company's bureaucracy and future.

Our series of articles on what Cisco needs to fix as well as its strategic needs for the future generated nearly 100,000 pageviews! In an interesting twist, after our first article was published, Cisco's analyst relations team reached out to us for the first time in seven years. We suggested a meeting, but we haven't heard back (that was months ago)!

Other hot topics included AI datacenters, Futuriom 50 startup news, and the HPE-Juniper deal. Check out the stats of our most popular posts below (traffic stats are as of December 30th, 2024, and rounded to the nearest 100):

Here's What Cisco Needs to Fix (36,800 pageviews)

Is It Time for Robbins to Move on From Cisco? (28,400 pageviews)

Cisco Shrinks Campus and Cuts Staff, Ending Era (22,500 pageviews)

The 2024 Futuriom 50 Is Here! (11,200 pageviews)

2024 Trend Preview: AI Infrastructure Startups (10,500 pageviews)

Does an HPE-Juniper Deal Make Any Sense? (6,300 pageviews)

Will the Cisco Reorg Work? (6,000 pageviews)

How Big is the Datacenter AI Chip Market? (4,600 pageviews)

HPE's Future Hangs on Stalled Juniper Deal (3,700 pageviews)

DriveNets Boosts Acacia Optical Integration (3,500 pageviews)

Why Cisco Is Buying Isovalent (2,800 pageviews)

Inside Cisco's $1B AI Play (2,600 pageviews)

NVIDIA Has Joined the Ultra Ethernet Consortium (2,500 pageviews)

Why IBM Wants HashiCorp (2,300 pageviews)

Security Is Cloudier Than Ever

By far the noisiest and busiest market we cover is cybersecurity, filled with its wide range of acronyms like SASE, CASB, DLP, CSPM, and other alphabetized technologies.

The cybersecurity market is the subject of much attention and investment, but the high-profile attacks keep coming. This year's highlight attack was Salt Typhoon, which targeted the networks of the largest North American operators, AT&T and Verizon.

One of the themes we have been tracking includes shift-left security, in which security operations become more tightly coupled with developer pipelines, code, and application programming interfaces (APIs). This was one of the drivers of F5's acquisition of Wib Security, a Futuriom 50 company, in February. This is an intriguing buy by F5, and it should serve it well by providing additional API-level security in its suite of networking security products and services.

Another trend in the market is security platform unification, which means packing more cyber-acronyms into products while at the same time delivering better integration so that end-users don't have to toggle among dozens of products.

In April at the big RSA Conference in San Francisco, we caught up with new Aryaka CEO Shailesh Shukla, a long-time industry executive (Cisco, Instart Logic, Juniper, Google), who has injected new energy into Aryaka, which has recently built a leadership position for itself with Unified SASE as-as-a-service. Separately, Versa continues to differentiate itself as a Unified SASE player that can service on-premises or cloud architecture with a single operating system.

There was some other interesting M&A in the cloud security market in 2024. Futuriom 50 member Lacework was sold to Fortinet in August 2024. Despite being one of the industry’s rising cybersecurity stars, Lacework didn’t live up to the expectations that its one-time $8.3 billion valuation raised in the market. Fortinet acquired its cloud-native application protection platform (CNAPP) at the garage-sale price of $152.3 million in cash, which showed how sometimes things don’t live up to the hype.

Aryaka Networks, Cato Networks, and Versa Networks are three network security companies on our Futuriom 50 list that are being closely watched as potential IPOs.

AIOps for Cloud Automation

There's networking for AI, but there’s also AI for networking. These technologies include AIOps, which uses AI tools to automate configuration and management of infrastructure; and observability, a large bucket of technologies that gathers data and insights from cloud infrastructure that can be used to drive automation.

In another new report for 2024, our Observability, AIOps, and NetDevOps report described how these technologies will need to be combined to deliver true cloud automation. To support a hyper-distributed infrastructure and data environment, the next-gen hybrid cloud infrastructure will require more automation than ever before. It’s no longer possible to manually connect, monitor, and operate cloud infrastructure. The scale of data, connectivity, and storage is too vast. As organizations seek more flexible and affordable cloud platforms for managing data and applications in both on-premises and public cloud environments, the demand for cloud automation technologies is greater than ever before.

Cloud automation spans needs across both enterprises and telecom infrastructure. At last year’s Mobile World Congress in Barcelona, we learned that telcos are pursuing AIOps technologies to drive down costs and boost productivity.

Rising Interest in Platform Engineering, MCN

In November of 2024, we launched yet another new report, this one on the emerging Platform Engineering and IaC market. This report focuses on how the technologies that comprise platform engineering, IaC, and internal developer platforms (IDPs) are being integrated to provide a more robust management and orchestration system for hybrid environments.

Both platform engineering and IaC service important needs for large enterprises trying to get their heads around a wide range of cloud infrastructure technologies and drive more automation, compliance, and security into their environments. Expect this to be a rising area of interest. It’s still early days, but we expect a lot of activity in this market, with a wide range of interesting companies to watch, including Hashicorp, Humanitec, Pulumi, Spacelift, and System Initiative.

We attended both Kubecon conferences (March in Paris and November in Salt Lake City), which serve as the focal point of this market. We also spent months talking to end users, investigating public research, and getting briefed about the platform engineering environment. A big takeaway from the Kubecon visits is that Kubernetes has now established itself as the growing architecture for just about any cloud compute environment, including AI. Each Kubecon we've attended has grown in size since the last, and we look forward to visiting Kubecon in 2025.

Speaking of distributed infrastructure, our Multicloud Networking and NaaS report hit its fourth year, highlighting the needs for dynamic networking infrastructure that can tie together multicloud and hybrid cloud architectures. We see demand for these multicloud challenges climbing as the data demands of AI increase, often crossing infrastructure and silo boundaries. This is a robust market with many players, including networking and colocation incumbents such as Cisco, Equinix, and F5, as well as a wide range of startups such as Alkira, Aviatrix, Itential, and Prosimo.

With that, we sign off on 2024. Have a Happy New Year! You can check for updates on our 2025 Calendar here.