When Cloud Services Equal Profits and When They Don't: Typical Cases
Maxim Berezin: "Of course, using a cloud makes sense if your infrastructure is obsolete and your budget cannot afford buying new equipment, or if your company wants to shed the burden and reduce capital expenditures."
Clouds have acquired a reputation as a mythical versatile tool helping any company save money significantly. There is a grain of truth in every myth, though. Today, companies leverage cloud services whatever their business scope and industry specifics. CROC’s cloud, for example, hosts multiple retailers, banks, developers of online services and aggregators, manufacturers, pharmaceutical and media companies, and advertising agencies. By its very nature, a cloud infrastructure can fit into any ecosystem of technical solutions used by every particular organization or company. However, here we have another important question: is it reasonable to migrate every single service and system to cloud?
We have analyzed how business of large customers with a long market experience work and discovered that companies mostly have static load comprising, for example, various internal reporting systems or back office infrastructure, that account for around 60% of all systems companies use. Legacy systems — especially numerous in banks — account for about 15% of all systems. These are usually die-hard apps with obsolete architecture entrenched to the point where company’s management doesn’t bother to get rid of them. Finally, services with dynamically changing load account for 25% of the total.
These are the ones worthy of cloud migration, as it is economically reasonable. Prior to starting every project, we perform an express examination of customer’s infrastructure, calculate the TCO of cloud migration, and give our recommendations to a customer. Our experience shows that profits from cloud migration of systems with dynamically changing load (such as online stores, processing systems, development environments, etc.) vary from 15 to 30-40%. Savings can be achieved through flexible resource management—by enabling cloud resources when they are most needed (during peak loads of customer requests or load testing to launch new products) and disabling them at low load.
A large IT developer cost saving during product testing in CROC Cloud
Of course, using a cloud is reasonable if infrastructure is obsolete and procuring new expensive equipment is not affordable, or if a company wants to get rid of a ballast and reduce capital expenditure.
However, if you are searching for cost-effective solutions for the services and systems with static load — the sixty-percenters I mentioned above — then renting a dedicated infrastructure would be reasonable, since it can reduce TCO by 20% over the next three years compared to using your own IT resources. There is more to it: customer reduces the payroll budget and receives highly qualified support, thus avoiding the pain of upgrading infrastructure in the future. IaaS model works great for a large bank we work with. The dedicated equipment has been ensuring infrastructure’s disaster tolerance, backup, and system health monitoring, while saving 31% of IT cost over the five years, according to the bank’s finance department.
Another advantage of renting dedicated equipment is the ability to promptly build up capacity by engaging cloud resources to boost performance of enterprise systems. We know it for sure after the project for a large online home appliance store. The retailer uses mostly the dedicated equipment in our data center, but as soon as promo campaigns and sales kick off, it starts using our cloud resources. Firstly, this ensures that website is always up and running, otherwise it can fail when customers rush in. Secondly, the retailer manages services cost at its own convenience without spending a dime on redundant computing capacity.
A large online store manages infrastructure costs during promo campaigns
Summing it all up, clouds are best suited for new online services that companies develop to attract new customers, retain the existing ones, and stay ahead of the curve. As a rule, such services have a cloud-native architecture; therefore, cloud migration will be simple and painless, resulting in reduced financial risks during the launch. Though should any service fail to meet customer expectations, it would be doomed to shut down.