Cost of Cloud Migration: The Hidden Factors

Jul 12, 20244 min read

If you think migrating away from the cloud to some self-managed servers will bring you profit, then you are not even getting started, you are moving backwards. In my decade plus DevOps career, I did 8 cloud migrations across Retail, Banking & Financial Services, only 2 were on-premise and it was not for costs. It was for security reasons. And these were 5 years back.

That 30% reduction you see in your opex is just a temporary fix. It is going to double when your business scales & it won't automatically manage it when you are sleeping or away for vacation.

I was up working on an on-premise k8s cluster issue until 3 am in the morning before leaving for the airport at 4 am for my first ever family vacation. This would not have happened if it was in the cloud.

Many think on-premise is going to reduce the costs, that's true only if the business is good enough to be running a simple monolith for the next 10 years. But if you are going to be at par with your competitors you cannot do it by yourself.

Here are some examples of what you end up doing for the rest of your career on premise:

Factors impacting cloud migration:

1. Growing Data Size:

As organisations witness an exponential increase in data volume, migrating to the cloud becomes an attractive option. However, rapid data growth can pose logistical hurdles, especially when dealing with terabytes of information across various sources like databases, object storage, VM images, and multimedia files. The sheer magnitude of data makes traditional migration methods impractical, necessitating scalable solutions.

In one such migration projects, when we started building the parallel infrastructure, it was in some hundreds of GBs , and by the time we reached the day of migration it had already become a few TBs which was practically impossible to copy over standard internet connection. Not to mention these were from a variety of data sources like databases, object storage, vm images, videos and what not.

2. Storage Limitations:

Traditional hardware storage area networks (SANs) lack the flexibility of dynamic on-demand disk allocation, posing challenges in resource optimisation and cost management.

Efficient utilisation of storage resources is critical for minimising expenses and maximising performance in both cloud and on-premise environments.

Dynamic on demand disk allocation was impossible since it was all on hardware storage area networks.

3. Infrastructure Automation:

Inadequate support for Infrastructure as Code (IaC) tools in bare-metal virtualisation platforms hinders automation efforts, impacting Service Level Agreement (SLA) compliance and Recovery Time Objective/Recovery Point Objective (RTO/RPO) benchmarks.

Automation is essential for optimising resource utilisation and enhancing operational efficiency, necessitating robust toolsets.

4. Architectural Disparities:

Variations in cloud providers' physical network topology, administration interfaces, and API access control present compatibility challenges. Designing a cohesive architecture that ensures secure connectivity across disparate cloud environments requires meticulous planning and resource allocation. Achieving seamless integration while maintaining data integrity and security is paramount.

There were some significant differences between the cloud & on-premise especially in their physical network topology, cloud administration and api access control.

Due to this we had to design an architecture which will be logically the same in terms of secure and seamless connectivity for services in both clouds.

5. System Diversity:

Operating systems play a crucial role in infrastructure deployment, with organisations often utilising a mix of Windows and Linux environments. This diversity complicates migration strategies, as compatibility issues and software dependencies must be addressed to ensure smooth transition and operational continuity.

There can be operations which need some specific softwares operating in specific systems. That's why there were a mix of systems with both windows and linux operating systems.

7. Containerisation and Real-Time Replication:

Implementing fully automated containerisation platforms on bare-metal hypervisors and ensuring real-time data replication between separate datacenters require robust infrastructure and network capabilities. These operations are essential for enhancing scalability, resilience, and data availability, albeit at additional cost considerations.

8. Support & Maintenance:

And last but not the least limited operations support for on-premise systems. Everything from hardware procurement to designing the infrastructure to renewing the licenses & warranties while maintaining an SLA of 99.99% falls upon self.

There will be no one cloud support tickets to blame for.

Summary:

Sometimes cloud services might have some hidden service charges, these are mostly cases where cloud providers(like GCP) do not host the software themselves, instead partner with third parties to manage it for them. So it is important to discuss the pricing model in detail with your cloud vendor and get a quotation only when the pricing is clear.

Many startups recently have moved to self managed on premise datacenters from the cloud just for this one factor. While on-premise solutions offer greater control and security, they come with their own set of disadvantages, especially for organisations accustomed to the scalability and flexibility of the cloud.

Migrating from cloud to on-premise entails significant upfront investments in hardware, infrastructure setup, and maintenance. Additionally, on-premise deployments lack the inherent scalability of cloud environments, making it challenging to accommodate sudden spikes in workload or data volume.

Moreover, managing on-premise infrastructure requires specialised IT expertise and resources, which may strain internal teams and result in higher operational costs in the long run.

Furthermore, on-premise solutions often entail longer deployment times and increased complexity in adapting to evolving technology landscapes compared to their cloud counterparts.

Therefore, while on-premise deployments offer advantages in control and data sovereignty, organisations must carefully weigh these benefits against the associated costs and operational overhead before committing to migration.

If this post was helpful, please share it and follow Kamalika Majumder for more.