Expert opinion

The Economic Feasibility of Clouds: Factors Not to Be Missed

9 minutes
#iaas  #haas 

Maxim Berezin

Despite the Russian cloud market growing by more than 20% annually and cloud services successfully entering almost all economic sectors, users still have different opinions on whether it pays to opt for HaaS or outsourcing.

Some believe that it is by all means more beneficial to have on-premise infrastructure and in-house resources ("self comes first"), while others embrace clouds as panacea to reduce equipment costs and find qualified IT administrators. As is often the case, the truth lies somewhere in between. Cloud adds value subject to scrutinizing cloud services, cloud consumption model and specifics of business apps that the customer strives to transfer to the cloud. In this post, the economic feasibility of clouds is cleared up by TAdviser together with CROC Cloud Services.

Russian companies use the following cloud consumption models:
  • Hardware as Service (HaaS)
  • Infrastructure as Service (IaaS)
  • Platform as Service (PaaS)
  • Software as a Service (SaaS)

Maxim Berezin, CROC Cloud Services:

"Actually, Russian corporate customers use IaaS and certain PaaS elements, but IaaS predominates in general. SaaS-related cases are rather rare; our experience shows that corporate customers opt for customization, each having a dozen of critical technical and organizational requirements that put an end to SaaS."

The choice of a certain approach is due to technical limitations and economic feasibility rather than customer's personal preferences.

Maxim Berezin, CROC Cloud Services:

"The audit of IT infrastructures and IT services that we conduct for our customers enables us to look at cloud migration feasibility from different angles and finally make a balanced decision. Large companies rarely opt for either all-in-the-cloud or all-on-premise approach. The use case is often as follows: 25% of systems hosted in the cloud, 40% (HaaS) and 35% (on-premise or colocation). In addition, customers often combine two or three public clouds (multicloud), use IaaS for some of the services, while running the others on premise. We have completed such projects and do know the best deployment options for such services."

Creating a cloud strategy is considered quite useful if not crucial. How to do it, is yet another (somehow still disputable) question. But in any case, there are some must-do’s of this strategy impossible to avoid, like deciding whether it pays to deploy certain IT services in the cloud or not and how to assess the result in economic terms.

In particular, you have to realize the following:
  • What do you need the cloud for: developing new IT systems or managing the data for further analysis? In these two cases, the business effect will be less noticeable.

"The transfer of the development and test environments to the cloud can provide both direct and indirect economic effect. The cloud can be managed automatically via API, which helps reduce the infrastructure team workload," commented Maxim Berezin.

  • Does the underlying server infrastructure experience relatively stable load or certain fluctuations (both scheduled and unforeseen)? Predicting such fluctuations more accurately often calls for historical data and mathematical statistics methods. However, truly accurate prediction is a rare case yet.

"We definitely recommend the cloud for hosting front services with high floating load (e.g. e-commerce, portals, loyalty programs, etc.)," said the CROC Cloud Services expert. "Here, opting for Pay-as-You-Go model enables cost savings, while operating on-premise infrastructure makes you overpurchase capacity."

  • Are you already certain about the IT service development prospects or not? Rapid launch and decommissioning of many new products, services and practices are easy to organize, even from the technical perspective, and it doesn’t mean poor management. On the contrary, it is often part of the business model.

"Cloud is a quite reasonable option to host new IT services with unclear future," Maxim Berezin believes. "They may either skyrocket by 30 times like 'Samokat' service during the pandemic or disappear after less than six months together with the business idea that gave birth to it. This is a standard case, yet exposing company to certain financial losses, with the cloud being here a good way to avoid resource planning mistakes."

  • Is a service, on the contrary, of standard nature, present in automation landscape of many (or even most of) corporate customers? For migration purposes, it is important to single out such services as well.

"There are so-called basic infrastructure services, e.g. mail, portal, 1С, VDI, which we recommend hosting in the cloud," CROC Cloud Services expert says.

  • Should the enterprise feel like not hosting a particular group of services in the cloud, whatever the reasons, it is better to define such a group in advance if possible.

Another quite important point is to define how mature your enterprise is in terms of automation with regard to application automation, app development technologies and corporate information resource management.

Particularly, application automation calls for:
  • distributed data processing technologies
  • service-oriented architecture, microservices, and containerization
Corporate development requires:
  • DevOps culture and its components (CI/CD, etc.)
Data management needs:
  • data warehouses (DW)
  • centralized relational databases (usually a common thing in any infrastructure), distributed files storages, NoSQL databases, and data lakes.

Collectively, almost all of these technologies are part of a popular Native Cloud concept. Surely, traditional software infrastructure also moves to the cloud rather successfully in most cases, but its cloud-native elements (already or soon-to-be adopted) are always easier to migrate to public cloud and then to improve, administer and scale in this environment.

Also, keep in mind hardware infrastructure of what type initially predominated in customer's IT landscape. Overall, there are two common options here:
  • scale-out architecture when many considerably cheap servers are easily united into a unified computing or data storage cluster, using cutting-edge software technology.
  • scale-up architecture based on a certain number of expensive high-performance scale-up servers from leading global vendors. Such servers are in place when a company has business-critical IT systems and databases running for a long time.

Cloud architecture is always scale-out and therefore, given technical specifics, cloud migration seems to be cost-effective by default.

Maxim Berezin, CROC Cloud Services:

"Large organizations often use heavy DBMS or application servers with scale-up capabilities only. Firstly, such systems require 256 cores and 4096 GB RAM and thus may simply not start in the cloud. Secondly, in case of, say, Oracle solutions, licensing is rather tricky when it comes to physical nodes in virtualization environments. You’ll have to purchase 2-4 times more DBMS licenses compared to physical equipment, and the entire cloud-related cost savings will vanish at once. Finally, it’s quite odd to use cloud for hosting virtual machines that occupy the entire physical node as they would perform better on dedicated physical equipment."

Taking a bunch of professionals’ opinions, you can gain an averaged, yet quite unbiased, understanding of which solutions are recommended for hosting in the cloud and which are not. However, searching for any universal business effect measurement method will be fruitless.

Maxim Berezin, CROC Cloud Services:

"Detailed IT system hosting quote requires an audit. We check the current load of IT services, equipment health (age, failure statistics), development plans, dynamic and static load, and RTO/RPO defining the quality of backup processes. We calculate the cost of ownership for each IT service and show in what case (cloud, HaaS or on premise) it is the least. The customer sees high-level comparison of various approaches in terms of TCO given the discount rate and, with cold statistics at hand, can make a balanced decision."

When selecting a contractor (cloud provider), pay attention to the following:
  • Maturity of the cloud market
  • Service provider's maturity and therefore its ability to provide all the methodology and, if necessary, clarifications before the customer makes a detailed decision in terms of business and technology. Here, we mostly mean the audit of existing infrastructure and an opportunity to draw unbiased conclusions on whether it pays to transfer a particular IT service to the cloud or not. Understanding of how long the customer is going to use the cloud also matters as economic benefits of the cloud (compared to on-premise) become more vivid over time.

"A cloud provider upgrades its data center infrastructure from time to time. Suppose, in the 2021 budget, the provider has allocated RUB 8M for overhaul repairs of a diesel rotary UPS. Big sum for a particular year, but not quite so in long term. Presumably, the cost for the customer will increase, but not much, just by 10-15%. However, it’s a trifle compared to customer’s CapEx when maintaining on-premise infrastructure," Maxim Berezin believes.

  • Broad range of IT service outsourcing options offered by the provider, not necessarily cloud-related ones, like building a customer's private cloud on the provider's platform or leasing data center physical servers. Sometimes systems are to migrate from the cloud back to traditional infrastructure. If this is the case, it will be good for the customer to have a choice after all. Of course, the customer can go back to its own servers, as well as build its private cloud or lease a physical infrastructure from the service provider. The last option is viable for the services that seem to run in the cloud ineffectively.

Maxim Berezin, CROC Cloud Services:

"If the infrastructure is large (1,000+ users), we recommend using HaaS, which is, in fact, the customer's private cloud based on a commercial data center. As for IT services with static or predictable load, the costs will decrease by 15-20% in 3-5 years compared to the public cloud while HaaS will be even more cost-effective or at least comparable to on-premise. In addition, the customer can simply rent hardware in an external data center".

In the public cloud, the customer pays for the very service it uses during a certain period of time, with cloud providers' billing systems ensuring flexible cost management.

"In CROC Cloud, the customer has 24/7 access to the billing system to control the expenditure," Maxim Berezin said. "The customer can set alerts triggered in the case of excess expenditure, so that the superior rapidly gets informed and takes measures. We use a fair hourly billing. The costs can be reduced by 60% by cutting off a part of the infrastructure during the non-working hours (nighttime, days off). If your cloud provider can do this, the cloud will be difficult to compete with in terms of cost effectiveness."

26 may 2021
How to Move to New Data Center and Avoid Losing Data?
Cloud services gain in popularity every year. In our estimates, some 70% of large Russian companies have used the cloud or other services based on commercial data centers, at least once. This consumption model has always been good for rapid data transfer from local to the provider’s infrastructure, enabling similarly fast service termination if it is not needed anymore.
4 minutes
24 may 2019
CROC's Cloud Bridging Staffjet HR Service and Russian Market
Staffjet is a recently developed digital service of ANCOR Holding, an international staffing company, and is now available across Russia from CROC’s Cloud, ensuring that the automated head hunting and recruiting system runs failure-free 24х7.
2 minutes
26 february 2019
CROC’s Cloud Improves Grain Holding’s Operation
Grain Holding, Russia’s largest flour and bread maker, recently moved its entire infrastructure to CROC’s cloud. The project involved transferring approximately 40 systems, from infrastructure services to applications, ensuring uninterrupted processing of 1,700 tons of grain and production of 250 tons of bakery products per day, as well as timely delivery of goods to 11 million consumers.
2 minutes
31 january 2019
Foreign Logistics Company Leverages Growth Driver in Russia
SK Operations, a Supply Chain & Logistics service provider and member of international investor SK Group, has transferred business data of their Russia based operations to CROC’s Cloud for consolidation and compliance purposes.
2 minutes
6 september 2018
KDL Labs Enjoy Business Continuity via CROC's Cloud
KDL, a clinical diagnostic laboratory network covering 40+ Russian cities, has moved its mission-critical app Laboratory Information System (LIS) to CROC’s Cloud from another virtual data center in order to ensure continuous LIS availability and hence reliable storage and fast processing of test results, which are often key to accurate and prompt disease diagnosis. CROC performed the migration on an extremely tight schedule, with almost no impact on KDL clients. The company now accesses computing resources through a combination of HaaS and CROC’s Cloud services.
3 minutes