Cloud Repatriation and the Impact of AI

Public cloud services have experienced significant growth, with spending increasing from $197 billion in 2018 to $561.1 billion in 2023. This model provides organisations with flexibility and initial cost savings when deploying applications.

However, as more companies migrate to the cloud, costs can rise sharply, prompting many to consider cloud repatriation—returning to on-premise infrastructure. The expenses associated with migrating applications back from the cloud, especially data egress and transfer fees, can substantially affect organisations that handle large volumes of data. Therefore, a careful evaluation of cloud options is essential.

The Benefits of Public Cloud 

The benefits of the public cloud environment are well-documented. Organisations can quickly provision new resources, pay only for what they use, and leverage the expertise and economies of scale of cloud providers. 

Public cloud environments are an attractive solution for organisations that need capacity to grow and require rapid deployment as they scale up. 

Moreover, speed-to-market is accelerated when using the public cloud as the physical space has already been deployed by the public cloud providers. This enables organisations to easily take advantage of this infrastructure for their cloud deployment.

Public cloud services have allowed organisations to become more agile, innovative, and competitive.

Public Cloud Challenges 

While the public cloud has delivered significant benefits, it may not be the most cost-effective option for organisations as their requirements grow.

Public cloud is like a hotel, where you rent a room or as many rooms as you need on a short-term basis. This is very cost-effective for a short stay and you don’t need to own the building. But if your stay is lengthy or you need a lot of rooms, a hotel quickly becomes very costly. In this scenario, a public cloud solution may actually be more expensive.

As a result, there is a trade-off between flexibility and low capex, but higher operational costs. For organisations with large volumes of data, the costs associated with data storage, data egress and transfers, can make it more cost effective to use a private cloud solution.

The shift towards public cloud technology has led to the effective hybrid cloud model. In this arrangement, stable workloads are best managed in private clouds, often housed in colocation facilities, while dynamic requirements operate in the public cloud. This approach highlights the importance of moving stable application workloads back to more cost-effective private infrastructure, driving efficiency and cost savings. Embracing hybrid cloud solutions is essential for sustainable growth in today’s digital landscape.

Private Cloud Infrastructure as an Alternative 

A private cloud is a type of cloud computing model that is dedicated to a single organisation. It shares many of the characteristics of public cloud, but the infrastructure is usually fixed and either owned or leased by the company that uses it.

It can be hosted on-premises or in a third-party data centre and provides businesses with greater control over their data and resources. Private clouds address security concerns with enhanced security and privacy compared to public clouds, as they are not shared with other users. They are the ideal environment for the repatriation of public cloud workloads, because they are both cloud environments, which makes them readily compatible.

The Tipping Point for Cloud Repatriation 

Reverse cloud migration is on the rise. Moving applications back to private cloud, from public cloud infrastructure, is known as cloud repatriation and is driven by several factors. Some of the core reasons driving the cloud repatriation trend include:

Cost Savings

As organisations gain a better understanding of their cloud usage and costs, they can target 'easy to migrate workloads' to run in a private cloud or an on-premise environment. This careful planning helps to reduce costs, whilst retaining the benefits of public cloud. Performing a thorough cost-benefit analysis can identify significant operational cost savings.

Scale

Corporate workloads have been significantly growing since cloud dramatically changed the data centre landscape. This has been accelerated by the AI boom, as the typical scale of AI deployments are much larger than conventional IT systems. As this trend continues, private cloud infrastructure becomes a more attractive option for companies. This is a core driver for cloud repatriation.

Rate of Change

AI applications are on a huge growth curve. So, when companies are in their early stages of development, cloud is a flexible and attractive cost-saving approach. However, once the application or service takes off, cloud costs can quickly mount up.

Vendor Lock-In

While cloud platforms boast complete flexibility, there is an unexpected lock-in effect that comes when a company migrates all its workloads to the cloud, especially when this is done at scale.

The prospect of cloud repatriation can be so large and complex that companies may not have the resources or skills to bring workloads back into owned infrastructure.

Additionally, as they attempt to repatriate their data, expenses from egress fees, charged by providers for moving data out of their cloud, can be considerable.

As these factors converge, organisations are reaching a tipping point, where the benefits of cloud repatriation outweigh the original advantages that they realised from public cloud. This shift in strategy is becoming increasingly common and all indications are that this will continue to gain momentum in the coming years. Organisations should evaluate the following factors if choosing to repatriate applications:

• Cost Analysis

• Access to Capital

• Length of Requirement

• Data Centre Skills

• Finding and Leasing Colocation

• Security and Compliance in Cloud Repatriation

The explosive growth of AI Applications 

The rapid development and proliferation of AI applications has led to a boom in the technology industry, which has had an immediate effect on data centres. They provide the necessary infrastructure for storing and processing large amounts of data, which is crucial for AI models. 

With the growth of AI applications, the demand for data centres has also significantly increased, accelerating the expansion of existing facilities and driving the construction of new ones. 

AI adoption has been so rapid that almost everything about it is new, including the companies that are developing the technology. These new businesses typically don’t have data centres, while some of their early development needs can be met by public cloud, the need for GPU-enabled computing resources is driving them to seek GPU-as-a-service offerings. These offer access to scalable, on-demand high-power computing resources while minimising capital expenditure.

The Rise of GPU-as-a-Service 

GPUs, or Graphics Processing Units, were originally designed for rendering graphics in video games and other graphics-intensive applications. However, their parallel processing capabilities make them highly efficient for training and running deep learning models for AI.

With their ability to handle large amounts of data simultaneously, GPUs have significantly accelerated the training time of AI models compared to traditional CPUs.

The use of GPU technology is shaking up the cloud market as new providers are launching GPU-as-a-service offerings to compete with conventional public cloud. As these offerings mature, they should be considered alongside private cloud solutions.

There are close similarities between today’s public cloud and GPU-as-a-service offerings, both offer infrastructure that is capitalised by the operator and offered on a pay-as-you-go basis. So, it is highly likely that the same economics that are causing enterprises to repatriate from public to private cloud will also prevail for the next generation of AI companies, especially since the scale of AI deployments are so large.

Conclusions 

The rise of public cloud has transformed the IT industry, but the tipping point for repatriating cloud applications is becoming increasingly evident as costs rise and workloads increase.

By repatriating applications to private cloud environments, organisations can regain control over their IT infrastructure, optimise costs, and create a more powerful application environment. While public cloud will continue to play a significant role in the IT landscape, the growing trend of cloud repatriation is likely to accelerate. Businesses are trying to strike the right balance between the benefits of cloud computing and the need for greater control and customisation.

Companies developing AI applications at scale should seriously consider building their own infrastructure as soon as it is commercially feasible, to avoid being caught with the same cloud repatriation challenges as their enterprise counterparts.

Right-Size your IT Infrastructure

To learn more about the benefits of repatriating applications from the public cloud and the role of private cloud and colocation solutions, contact our team of experts today.

We can help you assess your IT infrastructure, identify the best-fit hosting environment, and guide you through the repatriation process, to ensure a successful migration. 

If you'd like more information on how we can support your data storage, please get in touch.