If we look at our infrastructure, whether it’s a physical infrastructure in our server room or a cloud infrastructure, we’re always looking for savings. Especially now that we have been hit by the energy crisis and energy prices have gone up dramatically. Issues of savings are becoming more and more sensitive. Many people are hypocritical of reality, saying that physical infrastructure and cloud infrastructure are two different things and calculating costs based on just having servers or applications installed on them. And where are the electricity costs, cooling costs, redundancy costs, maintenance costs, data security, etc.? Of course, we can optimize hardware or energy purchases but we will quickly reach a ceiling for which we will no longer find savings. And this is where cloud infrastructure comes in and the question of whether this is the right moment to move to the cloud. If so, how to look for savings in this area?

A common statement is – the cloud is expensive – but what does it mean? If we look holistically at our systems running in the cloud we will find that all the additional aspects involved in running these systems (compared to physical infrastructure) outweigh the costs we will incur. Why?

Cloud service providers no longer buy servers, they produce them for themselves – dedicated highly specialized hardware for both server and matrix or network services. The offer of these services is unbeatable compared to the costs we would incur buying the hardware ourselves.

In the cloud we buy services exactly as we need them at the moment and we don’t have to worry about the number of servers, selection and purchase of arrays, network devices or all the other aspects related to the configuration and launch of the environment. Initial costs in terms of physical infrastructure are many times higher compared to the cloud and where, for example, the scalability of the solution?

Physical infrastructure is scaled for the day of deployment. Any change requires further purchases or modification of already existing resources. Cloud is a fully scalable service. We need increased resources – we launch more services on demand, our need for computing power decreases – we give up additional services.

Cloud service providers have been relying on renewable energy for years, becoming largely independent of local suppliers, resulting in lower costs.

Duplication of cloud infrastructure operation schemes in opened data centers means optimization of procedures related to their operation and thus lower operating costs.

At the cloud service provider, so to speak, as a standard, we get resilience without redundancy – a mechanism whose implementation in the local infrastructure would entail additional costs.

And finally, cloud services allow for a fast change management in the IT infrastructure.

The physical infrastructure in a classic server room or DataCenter is technologically closed to solutions like virtualization or containerization. Moving to the cloud gives the opportunity to use newer technologies such as IaaS, PaaS, SaaS, FaaS or ..aaS, the choice of which can significantly affect the final cost of our solution.

An interesting example of cost reduction is the IDaaS service, which allows, for example, a company implementing a new application solution to use a service to identify and authenticate users. Instead of wasting resources and time on writing such a module, it is easy to implement a proven cloud solution

If we decide to use cloud services, then their optimization can be done through proper RIGHTSIZING or AUTOSCALING.

The first mechanism means the correct choice of virtual machine class or processor architecture. The second mechanism is to increase on demand the number of virtual machines supporting our application solution or the resources of the virtual machines themselves, such as the number of virtual processors or the amount of memory.

Cost optimization in the cloud is also the need for constant MONITORING of environments. Using the tools provided by cloud service providers, we can monitor resources, their use, load, which can ultimately serve to improve, for example, application code. Additional mechanisms include alerting, e.g. in case of a sudden event in the system, or the use of tools such as Advisor.

Another aspect of optimizing the cost of cloud services is to analyze the implemented mechanisms of our systems ACCESSIBILITY. Depending on the High Availability or Disaster Recovery model used, we can make significant cost reductions. An example would be covering an insignificant application with an expensive backup plan. In this case, choosing a basic backup plan can reduce costs without significantly affecting the disaster recovery aspects of the application.

An important aspect of cost reduction in the cloud is the implementation of AUTOMATION mechanisms – indispensable for IaaS services. Among other things, it provides a guarantee of reproducibility of environment construction and avoidance of errors in case of its restoration, modification or duplication. DEV/TEST/PROD environments based on IaaS can be generated and removed on demand.

Another element of cost reduction is to analyze contracts with the cloud provider in terms of the possibility of e.g. CAPACITY RESERVATION, i.e. a commitment to allocate resources for a period of e.g. one or two years. This service can reduce costs by up to 60%, or SUSTAIN USE DISCOUNTS – the longer the server is in use, the higher the discount you will receive. In this case costs can drop to 30%.

Cost optimization for cloud services is an ongoing process. Analyzing costs at the implementation stage and reducing them does not mean that in a few months or a year the whole analysis process should not be repeated.

But why such a subversive statement – “Why does your datacenter cost 30% too much”? The authors of the Webinar have been providing analysis and cost optimization services to an international company for several months. During that time, they were able to implement savings of 38%.