Why are so many organisations migrating from public cloud infrastructure back to an on-premises/hybrid approach?

Understand the driving force and impact of the great ‘reverse migration’.

With the ever-increasing year-on-year cost of managing IT infrastructure in the public cloud and growing concerns over privacy, data security, and meeting SDGs obligations in a global environment, many organisations are redefining their long-term cloud strategy. Taking steps towards what we call a hybrid cloudification journey.

Here, we present a collaborative view on the emerging trend of reverse migration from the cloud and whether it’s worth considering when deciding long-term cloud goals for your organisation.

The driving force behind reverse migration

In the last decade, the IT industry has been busy moving infrastructure and managed services to public cloud providers, which promised overall low cost of ownership and flexibility at scale. However, things have changed over recent years. The industry is now exploring avenues back to hybrid/on-premises solutions from the cloud in a reverse migration trend that is gaining acceptance and adoption by more organisations. This trend is expected to grow in the mid to long-term as more organisations adopt this strategy to mitigate against overspending on operating expense (OpEx), security/privacy concerns, and availability issues. The ballooning cloud cost has led to the emergence of FinOps (a portmanteau of finance and DevOps) and GreenOps frameworks and principles, both contributing to better cloud usage by maximising efficiency to lower cost and reduce carbon footprints.

FinOps aims to maximise business value through a cutting-edge approach focused on cultural practice, which promotes visibility, best practices, and accountability on cloud management by aligning various functional teams.

GreenOps is the practice of reducing a cloud environment’s carbon footprint through the efficient use of cloud resources, energy consumption, and the use of renewable energy to promote sustainability.

According to Fortinet, most organisations (78%) prefer a hybrid or multi-cloud approach, integrating several deployments into one operating environment. A considerable portion of this (43%) use a hybrid of cloud and on-premises infrastructure.

There are five key drivers of this reverse migration:

1. Cost Management: The long-term costs of public cloud services can be higher than anticipated, especially with unpredictable expenses for high-end computing resources like Graphic Processing Units (GPUs). This has led to a growing awareness of the financial implications and the negotiation of mid to long-term contracts for cloud services.
2. Performance and Latency: On-premises solutions often offer lower latency and better performance for certain applications, which is crucial for real-time processing and other performance-sensitive tasks.
3. Data Control and Security: Companies prefer to maintain full control over their data, which can be challenging in a public cloud environment. On-premises and hybrid solutions provide greater security and control over sensitive information.
4. Compliance and Regulatory Requirements: Some industries have strict compliance and regulatory requirements that are easier to meet with on-premises or hybrid solutions. Non-compliance can lead to legal and regulatory disputes due to the laws governing the country where the data is stored or processed.
5. Business Continuity and Compatibility: Ensuring data security, integrity, and compatibility with existing systems can be more straightforward with on-premises infrastructure.

Predicting and remediating the potential OpEx and CapEx Cost

An IDC report shows continued increase in expenditure on cloud infrastructure as prices soared in 2022 and 2023. Over the next five years, IDC forecasts a growth in compute and storage systems spending for cloud-native workloads to be almost double that of infrastructure supporting traditional workloads (12.2% vs 6.2% CAGR).

Cloud services can quickly become more expensive due to the availability and deployment of various applications and service offerings. Especially in the areas of innovation and development tools for machine learning and artificial intelligence, flexibility, scalability, and integrations with other applications and services.

As cloud technology matures and organisations scale, the operating expense in the medium to long-term may not be cost-effective and can be more expensive in comparison to a five-year write off period. This is due to the unpredictable additional costs that are inherent in running these high performance, highly available compute and storage resources for the core business. Cloud providers charge their users for hardware costs, although these costs are often bundled into the overall pricing for cloud services. This includes expenses for servers, storage, networking equipment, and maintenance. These charges can vary based on the type and number of resources used, such as the performance and accessibility of storage, or the computing power required for specific tasks.

As cloud service costs increase and contribute significantly to overall revenue, income and profit margins decline. With a hybrid cloud compute and storage infrastructure, cost is managed through an optimum human and cloud infrastructure. And resource dimensioning i.e., GPU and hardware utilisation, is planned and optimised in line with the organisations’ core business applications and services.

Complying with data security and governance challenges

Data storage on hybrid/on-premises compute and storage infrastructure allows companies to maintain full regulatory compliance, requiring that data must only be stored and processed within a specified location or region. Other data regulations explicitly mandate compliance with various security requirements.

Despite assurances from cloud service providers, organisations still feel concerned about the data security risks associated with managed cloud services. Some believe these risks will be minimised if they manage and control their IT infrastructure. An IBM survey shows 54% of organisations agree that the public cloud does not present adequate protection for their data, and 50% have repatriated their data from the public cloud due to security risks and concerns.

Hybrid/on-premises for workload optimisation

The global demand for GPUs is rapidly expanding, driven by recent innovations and data analytics computing models. These allow organisations to access high-performance computing resources for AI, machine learning, data analytics, and complex computing tasks that are transforming the technology landscape. As workloads for core business applications increase on-demand due to their elasticity, it translates to higher operating expenses.

The migration trend to hybrid/on-premises compute and storage has significantly increased the global shipments and deployments of more GPUs as organisations migrate their core workloads to their optimised, customised, and self-managed hybrid/on-premises IT infrastructure.

Research conducted by research and markets reveals the GPU market for hybrid/on-premises deployments is projected to reach $63.0 billion by 2028 at CAGR of 34.6% from 2023 to 2028.

Our verdict on a hybrid/on-premises approach

This reverse migration to hybrid/on-premises cloud computing is steadily gaining a wider adoption by critical sectors of our socio-economic landscape, especially in the government/public, financial, health and manufacturing sectors. The benefits of having a hybrid, self-managed, highly secured, optimised for the critical core business application, are that it’s cost-effective over the long term, meets regulatory and security compliance, and offers low latency.

The hybrid cloud approach offers better data security and control, and the ability to scale resources quickly. Providing optimal value through the efficient use of cloud resources in relation to the public cloud where control is limited, less secured because it’s a shared infrastructure, and offers an unpredictable cost to the organisation.

Return on investment (ROI) is critical to the sustainability of any organisation and a ballooning cost from the cloud service providers may become inimical to the organisation’s revenue in the mid to long term.

Capgemini has extensive experience in supporting clients through cloud migrations, including cloud-to-on-premises migrations. Drawing on the lessons learned from this vast experience and deep industry expertise, as well as our refined processes, optimised resources, best practices, and reusable tools and scripts, we can deliver faster, smoother, and more efficient migrations. Supporting you end-to-end throughout your transformation.

Get in touch to start the conversation today.