This article is based on a January 2020 analysis originally sourced from analyticsinsight.net, covering the major cloud computing trends expected to define that year.
Cloud Computing Entering 2020: A Maturing but Still-Evolving Market
Cloud computing has been a top-tier IT trend since Amazon Web Services launched its S3 object storage service in late 2006. Before that, the Application Service Provider (ASP) model attempted to deliver hosted software, but it was fragile and unreliable. It took years of iteration, investment, and hard lessons for cloud providers to build the robust, enterprise-grade platforms we rely on today.
By 2020, cloud was no longer an emerging technology — it was foundational infrastructure. Nearly every connected device interaction touched a cloud service in some way. Yet the market was still far from fully mature, and 2020 promised meaningful changes in how cloud was delivered, secured, and consumed.
Two themes dominated the cloud conversation heading into the year: security and customisation. The following sections break down the key trends that shaped the cloud landscape in 2020.
1. Hyperscale Infrastructure Partnerships
A significant shift was underway among SaaS vendors that had historically managed their own infrastructure. Rather than competing on raw infrastructure capability — a race that only a handful of hyperscalers could realistically win — many enterprise SaaS providers began consolidating onto major public cloud platforms.
A clear example cited at the time was Infor building its CloudSuite product on AWS. Similarly, Salesforce was migrating more of its backend workloads to AWS infrastructure, and Microsoft’s Office 365 runs on Azure. The logic is straightforward: if a SaaS vendor’s competitive advantage is in application design and domain expertise, spending engineering resources on undifferentiated infrastructure is a poor use of capital.
Early regional partnerships reinforced the trend. Workday partnered with AWS in Canada, and Salesforce aligned with Alibaba Cloud in China — both deals reflecting the reality that SaaS vendors need local infrastructure presence to meet data sovereignty and latency requirements without building it themselves.
For DevOps and platform engineers, this trend meant a growing expectation to operate on hyperscale clouds rather than bespoke or co-located infrastructure.
2. Cloud-Native Applications and Microservices
The push towards cloud-native application design accelerated in 2020. Rather than lifting and shifting monolithic applications to the cloud, development teams were increasingly expected to decompose applications into smaller, independently deployable services.
This microservices approach offered several concrete benefits:
- Independent scaling: Individual services can scale horizontally based on actual demand without scaling the entire application.
- Fault isolation: A failure in one service does not necessarily cascade across the entire system.
- Technology flexibility: Teams can choose the best runtime or language for each service rather than being locked into a single stack.
- Faster deployment cycles: Smaller, decoupled units are easier to test, deploy, and roll back.
Beyond internal decomposition, SaaS vendors were beginning to co-develop complementary service packages. Where two providers served related but distinct business functions, joint offerings gave enterprise customers pre-integrated solutions rather than bespoke point-to-point integrations.
3. Software-Defined Everything and the Hybrid Cloud
Hyperconvergence — collapsing compute, storage, and networking into a single hardware appliance — had been a dominant data centre trend for several years. By 2020, a new architectural concept was gaining traction: treating the physical location of hardware as increasingly irrelevant, with software defining how all resources were provisioned, connected, and managed.
The practical expression of this shift was the rise of hybrid cloud as a first-class architecture, not a transitional state. Organisations were designing workloads with the explicit assumption that compute could run on-premises, in a public cloud, or at the edge — and that software orchestration would abstract away those differences.
For DevOps teams, this translated into investment in tooling and platforms that supported consistent deployments across environments: Kubernetes federation, infrastructure-as-code that targeted multiple providers, and unified observability across hybrid footprints.
4. Cloud Security: From Checkbox to Core Capability
Cloud security was already a pressing concern before 2020, and a series of high-profile breaches — including incidents involving major cloud service providers — brought the issue into sharp focus. The lesson was clear: traditional perimeter security controls such as firewalls and basic access controls were insufficient in cloud-scale, multi-tenant environments.
The areas where cloud providers and their customers were expected to invest more heavily in 2020 included:
- Multi-factor authentication (MFA): Moving beyond single-factor credential access, especially for privileged accounts and APIs.
- Network segmentation: Enforcing strict boundaries between workloads and tenants, even within the same cloud account or VPC.
- Encryption in transit and at rest: Applying cryptographic controls to all data paths, not just external-facing ones.
- Threat intelligence and behavioural analysis: Using machine learning and telemetry to detect anomalous access patterns before a breach is confirmed.
- Role-based access control (RBAC): Segmenting data and control planes so that each user or service has only the minimum permissions required.
Even a single significant breach carried reputational damage far beyond the directly affected organisation — it raised questions about cloud trust across the entire market. Cloud security was therefore framed not just as risk management but as a commercial imperative for providers wanting to attract risk-averse enterprise customers.
5. High-Performance Computing (HPC) Moves to the Cloud
High-performance computing workloads — traditionally the domain of on-premises supercomputing clusters and specialised hardware — were beginning to migrate to the public cloud by 2020. For years, HPC in the cloud was considered impractical: the infrastructure requirements were enormous, the demand was bursty and unpredictable, and building out dedicated clusters for short job runs was cost-prohibitive.
That changed as AWS, Microsoft Azure, and Google Cloud all added HPC-specific services and instance families, enabling organisations to provision massive compute capacity on demand and decommission it when the job was finished. Use cases driving this shift included:
- Weather forecasting and climate modelling
- Financial risk analysis and Monte Carlo simulations
- Genomics research and drug discovery
- Precision agriculture (e.g., analysing sensor data to optimise planting decisions)
- Computational fluid dynamics in engineering
The economics of cloud-based HPC are compelling: rather than purchasing and depreciating expensive specialised hardware that sits idle between jobs, teams pay only for the compute time they use.
The original source cited a claim that HPC cloud usage was expected to increase by approximately 40% in 2020, and that around 36% of infrastructure decision-makers ran HPC workloads in the cloud in 2018. These figures were attributed to a Forrester report but could not be independently verified for this article.
Looking Back: What These Trends Tell Us
Reviewing these 2020 predictions from a post-hoc perspective, most held up well. Hyperscale partnerships between SaaS vendors and major cloud providers deepened significantly. Cloud-native and microservices architectures became the default for greenfield development. Security investment increased across the industry, driven by continued high-profile incidents. And HPC workloads did migrate to the cloud in meaningful numbers, accelerated further by the pandemic-era surge in remote scientific computing.
For DevOps practitioners, the overarching theme is consistent: the cloud is not a destination but an operating model. The teams that thrived were those that built for cloud-native patterns from the start, automated security controls into their delivery pipelines, and treated infrastructure as software — manageable, testable, and version-controlled.