As cloud computing rapidly proliferates enterprise IT and organizations migrate more of their traditional workloads and on-premise data in the cloud, the cloud will remain one of the fastest-growing segments of IT spend.
According to Gartner, by 2024, more than 45% of IT spending on system infrastructure, infrastructure software, application software, and business process outsourcing will move from traditional IT infrastructures to the cloud.
Enterprises are struggling to scale their infrastructure to the insatiable demands. They are adding up costs to hardware procurements, issue software updates, secure infrastructure, train staff, etc. On top of that, the ongoing pandemic is holding available human resources, limiting data center facilities, and shrinking hardware supply chains. Cloud computing brings much-needed relief.
No doubt, Flexera State of the Cloud report indicates 61 percent of organizations plan to focus on cloud migration this year. However, against what marketing pages of AWS and Microsoft Azure make you believe, cloud migration is more than moving VMs and databases in the cloud. You have to understand app dependencies, access technical feasibility, select the best instances to name a few. And then there are post migrations woes. These challenges translate to added cost and not all of them are conspicuous until very later. To estimate the real cost of cloud migration for your business, you need to take a thoughtful, disciplined approach. The first step is to evaluate your current infrastructure.
The real cost of current infrastructure
If you don’t know how much your organization is paying for all those servers, software licenses, maintenance contracts, extended warranties, networking equipment, security contracts, then a visit to the accounting department will ease your day.
Also, you can gather data on network bandwidth, storage, and database capacity, etc. In a typical SME setup with a five-year hardware upgrade cycle, the first-year cost can be $40000-$50000 considering these variables:
In addition to the first year's costs, there are recurring costs. You may not buy new servers every year but must maintain them. The same goes for all the pieces of hardware and software running your IT infrastructure and resources managing it. Your recurring cost should be anywhere between 10-20% of your first-year cost.
However, the real cost of your current IT implementation is more than direct, operating, and administrative costs. Did you know you lose money every time there is downtime? When a power system in British Airway’s data center failed, the airline lost roughly $68 million in fare refund alone, not to mention 2.8% dip in their stock prices. If your on-premise setup has a 98% downtime, then you’re giving up 14 hours of productivity hours every month. Most cloud vendors guarantee a 99.9% availability and compensate for any additional downtime based on your contract. According to a Forbes report, for an hour of downtime, you could be losing thousands of dollars. However, for most SMBs, you could be losing anywhere from $3000 to $5000 per month in downtimes.
So the cost of total ownership of your on-premise setup for three years is approx. $75000 to $85000.
The real cloud infrastructure cost
So the long story short: once upon a time, it was next to impossible to estimate cloud infrastructure cost up until the actual migration. There was little competition and cloud computing was all about AWS and little of Google Cloud and Microsoft Azure here and there. Now we have every major tech firm in the cloud business and the competition is at an all-time high.
The pricing of major cloud vendors is still a little too complicated, but they now provide an approachable price estimation tool. Every major cloud vendor, along with AWS, offers some form of Total Cost of Ownership (TCO) Calculator, including Google Cloud and Microsoft Azure.
Mostly, it is a simple tool that asks you a few basic hardware questions like your required bandwidth, RAM, processor core, storage, etc. If the tool feels a little arbitrary, then there is an advanced tool too that asks enough questions to cover your entire infrastructure, but the estimations are more accurate.
So I ran estimations on TCO tool for AWS, GCE and Azure respectively for a typical SME setup consisting 32 vCPUs and 128 GB RAM.
Cloud offerings from Microsoft, Amazon and Microsoft are same in pricing although Google provides better value over Azure and AWS.
Moreover, all these public clouds have vast resources of case studies, pricing sheets, comparative analysis to back your estimation. These resources can act up as a template.
Nevertheless, migrating to the cloud results in up to 80% less cost.
The real cost of moving data to the cloud
If you’re moving a database into, for example, AWS, then you may have to create an on-demand instance. AWS bills you for on-demand instances and charges anywhere from $0.018 to $3.30 per hour based on your instance type and size. For instance, an on-demand t3.large instance goes for 0.146/hour while a r5.4xlarge goes $1.65/ hour. For storage, C4, R4, and R5 have a quota of 100GB while for T2 and T3, it is 50GB. Beyond the quota, AWS charges $0.115/GB.
Though data transfer is free within AWS in the same region, AWS charges for data transfer outside the region and AWS. Then, the data transfer to t3.large goes for $0.083/hour. For c5.9xlarge, it is $1.53/hour.
When it comes to GCE, Google bills you for Data Storage, network, and operations at the rate of $0.020 per GB for standard storage, $0.12 per GB for data outside the region and GCE, and $0.05 per 10,000 operations. Microsoft Azure has a similar billing method for data transfers.
The real cost of application data synchronization
Perhaps, the application you’re moving is a customer-centric application with thousands of simultaneous users at any time of day. The data may not synchronize once you deploy to the cloud from legacy systems. A lack of data integrity can break the app leading to downtimes and bad customer experience. When migrating data, to ensure synchronization, you have to opt for Multi-AZ deployments, which double the cost of data migration. t3.large that costs 0.146/hour in a single AZ deployment costs $0.292 per hour in multi-AZ deployment.
The real cost of hiring a consultant
So many hidden costs, are you still going to keep the migration stress to yourself? Well, a consultant can ease your migration stress to a large extend. A consultant will work with your respective teams to bring the most inclusive estimate without an asterisk mark. Even with all those cost estimation tools from all sorts of cloud vendors, a consultant will give you the most honest opinion according to your time, budget, and resources.