Organizations adopt this cloud strategy to “avoid being locked in the infrastructure, complementary services and price model of a specific provider”, and to achieve the adequate combination of cost, return, access to innovations in software development, etc. Sometimes they also pick providers who have geographically close data centers, or who distribute their applications according to certain legal requirements, Data Government requirements, sensitive data and regulations in force.
In a study, 86% of respondents defined their organization strategy as multicloud. Estimates indicate that 2021 will be multiclouds’ year, since the COVID-19 global pandemic has strengthened the critical need for business agility. It is also expected that by 2022 more than 90% of companies worldwide depend on a combination of local/specific private clouds, public multiclouds and inherited platforms to meet their infrastructure needs.
Portability in the spotlight
Modern cloud-ready applications are created from containers and microservices. This makes them adequate for their implementation in private and/or public clouds from different providers. Anyway, although applications can be designed to be “portable” –that is, to migrate through different cloud platforms without changes, – truth is that the main providers expect their platforms to tend to hold clients. For that purpose, they develop differential functions and services. Because of this, for organizations to find balance between portability and complete functionality use of a platform (which may involve a possible blockage) is pure art. Why? Because a portable application with a “common minimum denominator” may not take full advantage of a cloud provider.
What actually happens is that few applications move once they have been implemented into production and adopted by the company. Most multicloud strategies focus more “on the advantages of acquisition, on functionality and on risk mitigation, than on portability”.
The need to plan and why consider Multicloud
Multicloud strategy has many advantages. However, it should be pointed out that it adds complexity to management. Therefore, it must be planned.
Indeed, to avoid cloud service proliferation from going out of control, organizations must adopt policies that determine the circumstances under which certain workloads (applications, repositories, databases, etc.) must be deployed in certain execution places (stores, public cloud, private, cloud, etc.) This will make the growth of the constantly evolving Multicloud IT architecture more manageable.
Nevertheless, where does the need to have several cloud providers stem from?
Considering the fact that the business is a key factor in decision-making, there are certain aspects to be taken into account when adopting this multicloud strategy.
- The cloud solution does not meet the needs: Although choosing a cloud may be based on geographic closeness or latency, not all of that cloud’s services meet the business needs.
- Redundancy: The need not to build all applications in a single cloud provider, because it may fail. The big questioned here are SaaS services.
- The cloud provider does not offer what I need: Generally, some cloud providers are better as to certain characteristics than others.
- The IT department/other areas use the cloud like a shadow IT: Each company and department have their own systems, but many times the solutions they use are outside the scope of IT government. This tends to happen when the IT area is not fast enough to meet the business’s demands.
In reality, most multicloud scheme policies are based on the value or risk associated with each workload. For example, it is quite frequent for high-value or high-risk workload to be executed in the own data center or in a private cloud service -in fact, an investigation found that 90% of decision-makers consider local infrastructure to remain a critical part of their hybrid cloud strategies-. Besides, more generic commercial processes are frequently deployed in a SaaS service or in a public cloud.
Commonly, different cloud services operate independently from one another; however, many companies are starting to create hybrid cloud systems -that is, schemes that combine private cloud (local or hosted in third-parties’ facilities) and public cloud infrastructures-, which are interoperable and allow the management of processes in several execution places.
In short: ideally, companies should be able to execute their applications and workloads in a cloud environment which they consider most convenient from a cost, return or functionality point of view. But in practice, the situation may be different, since the larger number of platforms from different providers, the more complexity and management and administration challenges. Therefore, each company must evaluate whether the improvements they may get from a multicloud scheme, “pay” for the extra administrative load that this model may involve.
Did your company opt for a multicloud scheme?