Eclipse Digital - Fotolia


Understand top public cloud repatriation use cases

Cloud technology can pose billing, management and compliance issues. Here are five reasons why repatriating cloud workloads back on premises might be a better option.

Despite countless cloud provider case studies and user adoption success stories, the public cloud picture isn't always rosy, and some businesses are turning to cloud repatriation. Cost, security and business practices are among the reasons why organizations are switching back to an on-premises data center.

Read up on five reasons businesses often end up considering and undertaking cloud repatriation.

Operating costs

The cloud's potential to lower per-hour costs for services, such as computing instances and storage, is attractive to businesses. Compared to buying, installing, operating and maintaining those resources in-house, the public cloud often appears to be a bargain.

However, cloud providers have extensive a la carte menus of per-use, hourly and monthly costs. These recurring cloud costs can be difficult to predict and often extend beyond the core services of compute, storage and outbound data transfer.

Most enterprise workloads require support from other applications, such as databases and multiple compute and storage instances, that involve additional costs. Associated services, such as load balancers, monitoring and management tools, network connectivity and security, are also needed. Even the cost of cloud repatriation can be surprising.

A single service, such as Amazon Simple Notification Service, has no upfront costs and is pay-as-you-go, but users are billed based on several different factors: the number of notifications published and delivered, additional API calls for managing topics, and subscriptions. Pricing also varies by endpoint type.

An application's total monthly bill can sometimes result in sticker shock, especially for ones with erratic or unpredictable workloads. When compared against the cost of running such a workload in-house, the long-term cost of a major cloud deployment can become unattractive. In some cases, businesses find it's cheaper over the long-term to locally operate a workload.

Cloud-based workload performance

Performance is another factor that must be weighed when considering cloud repatriation. Performance indicators include resource configuration and use, plus networking concerns, such as bandwidth utilization and latency. When a workload's performance in the public cloud is unacceptable, unpredictable or doesn't reliably meet service-level agreements (SLAs), admins should evaluate other options.

Some organizations may opt to increase their control over the maintenance and resilience of vital workloads. This usually means repatriating the most mission-critical workloads to the local data center.

A workload can also have performance requirements that a public cloud can't readily accommodate. For example, latency-sensitive workloads may encounter performance issues in the public cloud.

Public cloud infrastructure isn't tailored to a specific workload, but rather is provisioned and replicated through an established service menu. If an organization wants to run a specific workload on a public cloud deployment, there's no guarantee the infrastructure can handle the workload's latency requirements.

Another consideration is a workload's need for the cloud's elasticity. If an organization puts a workload in the cloud to take advantage of scalability but then finds the workload doesn't require dynamic scaling, it may be less expensive to repatriate the workload back on premises.

Security and compliance concerns

Cloud resources are protected through security measures, such as identity and access management, data encryption, monitoring and logging. These measures require careful configuration and attention to detail to ensure adequate security and detect and respond to breaches.

Enterprises must have detailed knowledge of a cloud provider's security services and how each service is used. Mistakes or oversights can affect countless compute instances, services, log tools and firewall protections.

Managing the security demands of multiple large cloud applications can be time-consuming and error-prone. As a result, some organizations choose to undertake cloud repatriation to a local data center, where security tools and postures are better understood and controlled.

Businesses also face increasing regulatory pressure. Government regulations dictate how and where organizations store data, access methods to workloads and data, and workload resilience. Public cloud providers can offer the essential services to comply with major regulations, such as GDPR and Payment Card Industry Data Security Standard, but those services aren't a guarantee of full compliance.

Regulatory changes can render continued cloud use impractical. Thus, some businesses opt to move a sensitive application back into the local data center to ensure complete oversight and control over the workload and any associated infrastructure.

Workload and resource availability

Cloud providers make substantial investments in infrastructure and technology to ensure resource availability. But human errors, systemic failures, network problems and undetected hacks can trigger cloud service outages.

Such problems aren't exclusive to the public cloud, but when cloud infrastructure goes down, it affects countless workloads. Remediation from cloud providers for such outages can be a credit for lost service time, or additional bandwidth options, depending on the SLA. The penalty for businesses is more severe with lost revenue, customer dissatisfaction, violation of the SLAs of the business and possible compliance issues.

Some organizations may opt to increase their control over the maintenance and resilience of vital workloads. This usually means repatriating the most mission-critical workloads to the local data center, where in-house staff can implement a redundant local infrastructure capable of surviving hardware, software and external events.

Changing business needs

A business can simply conclude that certain workloads are no longer suitable for the cloud. This can happen when a workload's usage patterns change and don't require as much compute power or scalability.

The organization may also decide to leave a given market, making associated workloads and data unnecessary. Cloud repatriation may be a first step toward retiring the application.

Other business changes might include financial reporting and investment strategies that emphasize capital investment and make the public cloud's recurring operational expenses less attractive.

Mergers and acquisitions also change cloud requirements. Consider a situation where a business that doesn't use the public cloud buys another business that uses public cloud technology. The purchased business typically realigns its assets to conform to new ownership requirements, which may make cloud repatriation the best option.

Dig Deeper on Data center ops, monitoring and management

Cloud Computing
and ESG