Is It The Time Of A Cloud Exit? Finding Savings And Repatriation Strategies

In recent months, we have been surrounded by a new boom in the IT community, and Cloud repatriation, in a narrower sense, refers to moving corporate IT from a multi-national cloud, such as AWS, Azure, or Google Cloud Platform (GCP), back to an on-premises infrastructure. However, cloud repatriation is more often thought of as an alternative solution that is not built on the services of a single cloud provider.

In addition, there is no universal solution when placing IT infrastructure in the cloud. Each organisation needs to mix its own cocktail of cloud services, and anyone who missed the mixology lessons when planning a move to the cloud was left wondering. Perhaps that is why public statements about a big cloud exodus are causing so much controversy in the ensuing discussions. Yet the reasons for leaving the cloud can be quite objective, and it is not always faulty input analysis or naive trust in marketing platitudes that is to blame.

The Limits Of Global Clouds And Why Companies Are Leaving

Public hyper-scaler clouds promise flexibility, ease of management, and cost savings on in-house hardware.

However, in practice, it is becoming evident that cloud services are binding in certain ways — some point to non-transparent pricing and others to the need for qualified platform administrators. So, let us now take a closer look at the most problematic points.

• Contentious Flexibility: The scalability options with cloud giants are essentially endless. Setting up thirty new servers anywhere in the world simultaneously is something you could never achieve with your own infrastructure. However, when adjusting parameters, you must conform to the provider’s precisely defined products. Another issue is the strong dependency on a single provider, known as vendor lock-in. Do you have everything in AWS and want to move part of your operations elsewhere? If you do, then you have a problem because the whole process will be very complicated and time-consuming.
• Expertise In Provider Services: Cloud services undoubtedly offer more convenient management than physical servers in a data centre. Everything can be arranged with just a few clicks from one place. However, effective optimisation and configuration of complex services require advanced knowledge of individual products. Of course, you can outsource service management and utilise support directly from the provider – it probably will not surprise you that you will also pay a corresponding premium for it.
• Unexpected Costs: We have already described the problematic billing of major providers in the article Pricing AWS: How to Save on Managing Services. In the most common scenario, the first invoice is especially surprising. Initially, no one knows how much the charges for resource usage, request processing, or data transfers will amount to and whether some items can be adjusted and services consolidated. However, other costs must be considered in the future, and even professional optimisation of the hyper-scaler’s price list will not help.
• Security Risks And Legislation: Inadequately or incorrectly implemented security measures increase the risk of data loss. According to HashiCorp, up to 50% of organisations are concerned about data leakage or theft associated with cloud operations. Interestingly, 44% of companies consider the lack of skills as the biggest internal threat, and this shortage becomes even more serious in the context of increasingly stringent cybersecurity requirements. Not only legislative rules but also pressure on companies and institutions to ensure the quality and security of their IT infrastructures is rising. In many cases, partial cloud exit becomes the only solution.

Furthermore, all the issues mentioned above are often interconnected. For example, inadequate security can lead to data leaks, resulting in increased costs for detecting and fixing threats, which cannot be done without trained people.

However, one way to prevent some threats and problems is to prepare your application to run in the cloud (cloud-native).

Will Cloud-Native Architecture Help?

IT infrastructure can be moved to the cloud in line with the lift-and-shift concept, meaning as it is. However, such a system is not ready for the cloud environment and will likely face several issues. In addition to those mentioned above, instability or limited scaling capabilities are added.

On the other hand, cloud-native applications are modified or designed and developed to be cloud-compatible. Their fault-tolerant architecture, created with security rules in mind, enables easy scaling of the number of application instances and, thanks to containerisation, allows automated creation, testing, and deployment of new versions.

However, utilising modern practices and technologies in cloud-native approaches requires corresponding expertise. In addition, the demands for more sophisticated management tools and monitoring will consume what is saved on serverless operations. The more specialised services a company uses from a provider, the more challenging it is for them to migrate some systems, leading to more intense vendor lock-in. A cloud-native architecture can, therefore, only help eliminate some of the challenges.

How To Get Out Of It: Five Cloud Repatriation Scenarios

Scenario 1: Distribution Of Systems To MultiCloud

Problem: Depending on the services of a single provider, the organisation faces the risk of downtime and lack of availability of its systems. Such problems could cause it to lose revenue and reputation.

What MultiCloud Solves: By spreading services across multiple providers, the organisation can reduce the risk of outages and strengthen the stability of its IT infrastructure with a redundant and distributed architecture.

Who It Will Help: MultiCloud will primarily benefit large companies with high demands for availability and resilience to outages, aiming to minimise the risk of financial and reputational damage.

Scenario 2: Combination Of Shared And Private Solutions (Hybrid Cloud)

Problem: The organisation has specific data security requirements, particularly in access control and compliance with legal obligations. Managing this sensitive data in the cloud is complex as it requires a high level of expertise, which also translates into costs. Underestimating this area increases the risk of regulatory breaches.

What A Hybrid Cloud Will Solve: Hybrid solutions allow organisations to operate part of their infrastructure on physical hardware, helping them meet legislative requirements and better secure and monitor sensitive data. This approach enables the organisation to reduce costs associated with managing sensitive data in the cloud without losing flexibility and other benefits of cloud services for other systems.

Who It Will Help: A hybrid cloud is ideal for companies and institutions with high or specific data security requirements that also need flexibility for future growth and handling unpredictable data demands.

Scenario 3: Smaller Provider’s Public Cloud

Problem: For some companies, operating in large clouds may be associated with strict technical requirements. Standardised solutions and expensive support become financially unsustainable in the long term.

What A Local Public Cloud Solves: Moving to a local cloud provider allows companies to reduce costs and get better support. With tailored solutions, they can respond well to the current needs of their IT infrastructure and keep up with the market.

Who It Will Help: Local public cloud is especially suitable for small and medium-sized enterprises that require a personalised approach and cost-effective solutions within a limited budget.

Scenario 4: Operations Solely In A Private Cloud

Problem: The organisation deals with highly sensitive data, and standardised solutions in a global cloud do not meet the requirements for their security. Implementing adequate protection would significantly increase service costs and the recruitment of a specialised team to maintain security standards.

What The Private Cloud Solves: Operating solely in a private cloud will allow an organisation to achieve the required data and application security at a lower cost. At the same time, it eliminates the complexity of the services provided, making it easier to manage the entire infrastructure.

Who It Will Help: This scenario brings the greatest benefits to financial institutions such as banks or investment firms that have to meet the financial sector’s high data protection and regulatory requirements. However, the private cloud can also be used by healthcare organisations and other entities that place a premium on data security and privacy while maintaining high availability of their systems.

Scenario 5: Complete Return To Physical Servers

Problem: The company uses specific production processes and procedures. The cloud environment is not adjusted to the company’s requirements – limited access to data negatively affects production processes and logistics. In addition, standardised products from large providers do not allow full optimisation of production processes, which are, therefore, prone to failures.

What Physical Infrastructure Will Solve: Returning to traditional operations on in-house hardware will enable the company to gain full control over the infrastructure and reduce costs associated with cloud services. Performance requirements are stable and do not exhibit unusual fluctuations, allowing the company to afford a longer setup time for potential new physical servers.

Who It Will Help: This option suits manufacturing companies and enterprises with high reliability, speed, and process accuracy demands.

On-premises vs. Colocation in Data Centre Operations

If you decide on hybrid or private cloud or a return to physical hardware, the question remains: where will the infrastructure run? There are two options: directly in your own server room (on-premises) or with a professional data center provider.

On-premises operations – the physical placement of servers, storage, and network devices under your own roof – provide maximum control over infrastructure and data. Operating your own data center involves ensuring optimal conditions that meet the requirements of the number of servers and support the longevity of these devices. In such cases, a company cannot do without qualified staff to manage the server room.

Colocation in a data center allows companies to place servers in an environment with high standards for data protection and service stability. Colocation eliminates the costs and burdens associated with operating your own data center. Additionally, through partnerships with operators, it’s possible to obtain more convenient conditions for purchasing hardware or network connectivity.

How To Plan A Cloud Exit Strategy

Analysing existing services and potential solutions will help you choose the right scenario for your business.

However, incorrect analysis may result in unnecessary costs or additional subsequent migrations in the future. Deficiencies in analyses can also lead to cloud exit; for example, if the entire infrastructure had been moved to the cloud in the past when it would have been more appropriate to keep some systems on-premises.

So, what are the key stages of a cloud repatriation strategy?

1. Analysis: Includes an assessment of the services currently in use, including available features, quality levels, performance, and operational support in relation to the current and future needs of the specific organisation. The entire system needs to be reviewed in terms of data needs, security requirements, and overall cost.
2. Identify Gaps And Needs: Is the most pressing problem inadequate performance, limited flexibility, security margins, or high costs? What does the IT infrastructure need, and in what areas are errors occurring and recurring? Do you need to respond quickly to changing demand and unexpected performance fluctuations? Which systems require flexibility, and which ones work with a consistent set of requirements?
3. Evaluation And Comparison Of Alternatives: Based on the identification of needs, select possible scenarios, reflecting on the advantages and disadvantages of each. Focus on acquisition and operating costs, and especially potential risks. Then, develop a strategy to mitigate the impact of these risks.
4. Schedule The Migration: After the previous three steps, you should have a relatively clear idea of the best scenario. Now, it is time to draw up a timeline and the steps the organisation and individual departments need to take. Prepare a plan to ensure service continuity and data security during the migration. Everything should take place without significantly impacting customers, internal processes, and employees’ work. You can read about how to plan a migration in the article Strategic Tips and Common Mistakes in Data Migration.
5. Consideration Of Legal And Contractual Obligations: Cloud service providers will have a set period for the termination of services, typically 30-90 days. In some cases, there may also be a precise procedure for migration from a given provider, which may include the use of specific tools or protocols, as well as the adherence to security standards. Withdrawal from the contract before expiry may also incur a fee.
6. Testing And Verification: Before initiating migration, it is advisable to simulate and verify the functionality of the new systems. Specific scenarios of the live operation should also be tested to uncover performance and security gaps.

Tailor Your Cloud Stay To The Requirements Of Your IT Infrastructure

Despite the growing popularity of the term cloud repatriation, it is evident that a mass return to on-premises infrastructure is not happening and is not expected. There are also several ways to address some of the drawbacks of operating in the global cloud, and it depends on the specific needs of each organisation. Choosing the right solution is best guided by an analysis of the current state, and a smooth migration process is ensured by detailed work planning and thorough testing of the new environment. With all the steps of a cloud exit strategy, we at MasterDC can assist you.