The fast‑moving cloud computing industry is already showing several major trends that, according to CRN, are likely to define the cloud market in 2021.
Cloud computing is evolving so rapidly that it often feels impossible to keep up or predict where things are going. However, some powerful trends that are expected to dominate the market next year are already attracting a great deal of attention—even though the underlying technologies and the vendors behind them are still changing and may be disrupted at any time.
The traditional “islands” between different types of cloud infrastructure (public cloud, hosted private cloud, and on‑premises data centers) are gradually disappearing. This increases workload portability and makes data flows more mobile.
Many of these trends are the result of the industry entering a phase of standardization and improved interoperability, which is typically a sign that a technology sector is maturing.
Standardization has been driven largely by the open‑source movement, which has allowed the focus to shift “up the stack” and created new roles in the partner ecosystem to support application‑level workflows—from enabling AI and high‑performance computing (HPC), to delivering innovative SaaSOps and application development services.
By 2019, as enterprises increasingly deployed workloads across multiple infrastructure‑as‑a‑service providers, we had clearly entered the multi‑cloud era—so much so that “multi‑cloud” became something of a cliché.
But as applications become more portable, compute capacity more readily available on demand, and data integration platforms simplify connectivity, and as vendors form cross‑platform alliances, the current multi‑cloud trend may soon start to look more like an “all‑cloud” trend.
In practice, the largest enterprises may soon become customers of all major hyperscalers and several niche providers, allowing them to take advantage of increasingly differentiated services and specialized protocols—while reducing the risk of vendor lock‑in.
Hearst Corporation, which owns more than 360 companies worldwide, is a typical example. The New‑York‑based media, information, and services company has undergone a digital transformation across AWS, Microsoft Azure, and Google Cloud. This multi‑cloud strategy gives Hearst’s developers and business units a strong competitive position in all of its key markets.
Enterprises are choosing the Kubernetes platforms that best meet their specific operational and functional needs. This might be a prescriptive solution such as Red Hat OpenShift, a distribution from Pivotal, standalone offerings from Docker or Rancher Labs, or managed Kubernetes services like Google GKE, Microsoft AKS, and AWS EKS.
Kubernetes often becomes the foundation that allows organizations to scale applications across different cloud infrastructures, thus fulfilling the promise of multi‑cloud.
As a result, Kubernetes not only breaks down cloud silos but also creates some unusual market dynamics.
Providers of cloud infrastructure software are increasingly decoupled from the operators who actually run the data centers and own the physical server racks. This has led to products and combinations that would have been hard to imagine a few years ago.
For example, Google’s Anthos can run on AWS or Microsoft Azure almost as easily as it runs on Google Cloud Platform. VMware’s upcoming Tanzu portfolio is also designed to span on‑premises and all major hyperscalers.
In this multi‑cloud world, it’s not just customers’ workloads that span multiple clouds—cloud providers themselves are increasingly expanding onto one another’s turf.
Kubernetes has been a huge success, but the cross‑cloud unification it promises is being challenged by the growing ecosystem of tools built around it.
The Cloud Native Computing Foundation (CNCF) maintains tight control over the core Kubernetes project while incubating a range of complementary technologies to round out the platform. The CNCF landscape and project roadmap describe a path for enterprises to adopt containers more broadly based on real‑world use cases.
One key technology on that roadmap for mixed‑environment deployments is the Istio service mesh. Other open‑source projects are also becoming critical components, from Prometheus for monitoring containerized applications, to the ELK stack (Elasticsearch, Logstash, Kibana) for logging and troubleshooting, to Harbor as a production‑grade container registry, and Jaeger as an emerging de‑facto standard for distributed tracing across microservices.
While major cloud providers continue to support upstream Kubernetes, they are increasingly promoting competing technologies for many of the surrounding components on the CNCF roadmap. For example, AWS encourages customers to use CloudWatch instead of Prometheus for monitoring, and Google does something similar with Stackdriver.
Fragmentation in the Kubernetes ecosystem can undermine progress toward cross‑cloud unification. As more enterprises adopt cloud‑native platforms, tensions around these competing stacks may intensify.
Every cloud infrastructure vendor—whether a public cloud provider or an on‑premises hardware/software supplier—needs a strong Kubernetes strategy to remain competitive in today’s market.
Time and again, acquisitions have proven to be an effective way for traditional vendors to inject new life into their Kubernetes portfolios or build entirely new offerings.
The most notable example is IBM’s USD 34 billion acquisition of Red Hat. With OpenShift and the earlier acquisition of container pioneer CoreOS, IBM assembled a powerful Kubernetes solution.
In 2017, Microsoft acquired Deis after first committing strongly to its ecosystem, expanding its Kubernetes tooling. NetApp entered the market in a similar way by acquiring StackPointCloud.
Perhaps the best deal was VMware’s acquisition of Heptio, whose technology laid the foundation for Tanzu—a portfolio expected to underpin Dell Technologies’ cross‑cloud vision in the coming years.
Meanwhile, there are more container‑focused startups than ever (Docker itself is one), so we can reasonably expect a new wave of Kubernetes‑related M&A from major cloud providers in the year ahead.
When choosing a cloud provider, many enterprises want a comprehensive set of native security capabilities rather than stitching together numerous third‑party tools.
Vendors that lack the capacity to build a complete modern security architecture in‑house are turning to acquisitions. That is why security M&A has been such a prominent theme in the past year.
VMware has strengthened its security story by acquiring endpoint protection specialist Carbon Black and, later, Intrinsic (Intrinal) to bolster its application‑security capabilities.
Earlier this year, Microsoft acquired BlueTalon to enhance its position in data security and governance. Google folded its sister company Chronicle (under Alphabet) into Google Cloud, although some reports have suggested the integration has been challenging.
Broadcom completed its acquisition of Symantec’s enterprise security business. Cisco acquired Duo Security, and AT&T acquired AlienVault and used it as the foundation for a new cybersecurity division.
Security is such a complex and ever‑evolving area that there always seem to be more gaps to fill. In 2021, the M&A trend is almost certain to continue—and likely accelerate. There is no shortage of startups that can help cloud providers build out a full suite of native security features.
The previously one‑way migration of workloads toward public cloud is starting to look more like a two‑way street.
Containers and other technologies that improve application portability make it easier for enterprises to move workloads back to private infrastructure. Many organizations are becoming more familiar with the unique strengths of each environment and are learning to leverage them.
This does not mean that the pace of migration to public cloud will slow or reverse. But as some customers realize that certain workloads can achieve better cost efficiency, performance, or security in software‑defined data centers, workloads will move more freely in both directions.
The resurgence of private cloud will further fuel the already hot market for hybrid‑cloud solutions.
Today, nearly every SaaS, IT operations, analytics, and business intelligence (BI) product claims to incorporate machine learning in one way or another—whether it is truly needed or not.
Beyond the very real benefits that automation and insight can deliver, “AI” is also a powerful marketing term; adding it to product brochures rarely hurts.
From chatbots and inference engines to predictive analytics, AI can be plugged into almost any cloud software product. Some machine‑learning‑based features are genuinely useful; others are little more than hype. But by next year, it will be hard to find a product that does not brand itself as “intelligent”.
As SaaS adoption explodes, more specialized platforms have emerged to manage the migration, operations, and spending associated with these cloud‑based applications.
To keep up, the discipline of SaaSOps—operations focused on managing SaaS applications—is becoming firmly established as a new role for IT professionals, often working in tandem with offerings from cloud access security broker (CASB) vendors.
In addition to the well‑known suites from Salesforce and other major SaaS providers, vendors such as BetterCloud, CloudManager, and Blissfully now offer end‑to‑end management solutions for Microsoft Office 365, Google Workspace (G Suite), and more.
As Allen Falcon, CEO of Cumulus Global, puts it: “The term SaaSOps is still somewhat vague, but it can include security, license management, spend management, discovery/mitigation of shadow IT, and managing account creation and deprovisioning across platforms and business domains.”
In many ways, Kubernetes has brought order to what was once a chaotic cloud infrastructure market. It has effectively become the standard way to deploy infrastructure for cloud‑native workloads.
According to Dan Kohn, former executive director of the CNCF, this shift opens the door for competition to move up the stack and focus on improving application delivery.
We can expect to see major open‑source projects emerge around “application development” practices and services designed to package and deliver applications more effectively.
To support this vision, the CNCF has created a new Special Interest Group (SIG) for application delivery.
High‑performance computing (HPC) workloads are often sporadic and run in batches. In such use cases, the elasticity of public cloud—scale up and down on demand—looks very attractive.
Yet despite these apparent benefits, the HPC market has long remained primarily on‑premises. Only in the last few years have HPC users begun moving workloads to the cloud on a regular basis to ease capacity bottlenecks.
Skepticism about cloud in the HPC community largely stems from the nature of the work: these systems often support new product designs, proprietary data models, and advanced simulations. Scientists, engineers, and product developers have historically been very cautious about letting such critical workloads run off‑site.
But the situation is changing. The expensive multi‑core systems required for advanced computation will increasingly move to public cloud providers.
Over time, the desire to avoid constant capex upgrades in capital‑intensive data centers, combined with the ability to provision massive compute resources on demand, will prove irresistible—even if the cost of running HPC in the cloud can be somewhat unpredictable.
Technical support
Scan the code to join the technical support WeChat group
Official account
Scan the code to follow and get the latest updates