Data Center Construction vs. DCaaS: What to Choose and Why?

Why DCaaS?

Data center infrastructure underpins all global digitalization projects and service roll-outs, but its construction and operation are very expensive. Typically, a medium-sized facility for several hundred racks costs at least two billion rubles, with the construction taking about three years. Moreover, data centers for corporate use bring zero return on investment, as companies, unlike commercial data center providers, do not derive profits from such sites. This is with the exception of telecom carriers that can diversify their business by offering data center based services.

Over time, the cost of construction and operation only goes up. Such data centers are mainly built using foreign equipment, which went up in price by 20% over the past year due to the increase in the exchange rate. Crisis trends negatively affect almost all economy sectors, with companies having less and less resources to upgrade their infrastructures. Therefore, leasing data center equipment and using related services (such as clouds) are becoming the most cost-effective ways towards infrastructure upgrade or its fault tolerance improvement. In the latter case, businesses use commercial data center capacity to build backup data centers as an addition to existing corporate racks. Even banks, usually reluctant to use outsourcing services, do so.

Our customers outsource expertise mainly due to the clear lack of qualified specialists, especially in the regions. We witnessed ourselves how difficult it is for data center owners to find and retain qualified personnel. Moreover, they need to continuously invest in staff professional knowledge, which may cost a fortune. That is why more and more customers choose outsourcing to ensure the required level of infrastructure support while spending less on human resources. Thanks to a more balanced administrator workload, a provider’s man-hour usually costs less than that of a customer. Thus, according to Stockmann’s CIO, using its own resources to maintain computing infrastructure would cost the retailer dramatically more than going for cloud services, with support being already included in the price.

Indeed, many companies opt for Infrastructure as a Service (IaaS) to avoid heavy capital expenditure on equipment procurement and continuous upgrades. For several years now, we have been seeing companies focusing on their core business and outsourcing all non-core functions. This is especially true of foreign companies which do not need a data center or IT staff to open their representative offices in Russia, with a few exceptions described below.

“It takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!” (Lewis Carroll). Now, all businesses that initiate digital transformation are guided by this idea. They cannot promptly roll out and customize new products without computing resources (preferably outsourced). Indeed, service launch usually takes one month or less in a cloud and a little longer in case of HaaS, which is way faster than equipping an in-house server room. Therefore, companies increasingly often move their IT development activities to third-party clouds. What’s more, outsourcing involves way less red tape. For some projects, our customers from the finance and even public sectors went for colocation in a third-party data center as such option was easier to approve internally.

Why build your own data center

Most data centers are located in European Russia, primarily in Moscow and St. Petersburg. The provinces from the Far East to the Urals still lack commercial data centers, with new sites only starting to emerge there, while metropolitan providers located 300 km or further away from users cannot ensure effective service delivery and fast data transfer. In some cases, customer employees have to personally visit a site, set up the equipment, and control processes. That is why large regional customers, such as manufacturing, energy, and oil and gas enterprises, opt for building their own data centers.

In addition, the IIoT development has been driving more and more manufacturing enterprises to adopt so-called fog and edge computing. These technologies process data from multiple sensors and other devices at source. In such cases, data transmission to a cloud is not always reasonable due to overload of local communication channels, never being stable enough in harsh Russian climate.

Perhaps, stringent data processing standards adopted at finance, manufacturing, and state companies represent the biggest roadblock on the way to commercial data center development. The transfer of such data to third parties is prohibited or limited. Thus, manufacturing corporations and such either build private clouds based on on-premise architecture or move out only low-value services that do not contain commercial information or employee and customer personal data.

Sometimes businesses doubt whether they made the right choice, data center investments are justified, or their current data center provider is trustworthy. An independent market assessment will answer these and many more questions. Even a quick analysis of existing processes and infrastructure can help identify problems and decide between the two: own or outside resources. Such consulting often becomes the cornerstone of the long-term IT development strategy.