The Public Cloud has long ago become an integral part of the IT infrastructure for most businesses. For some companies that might just be email services, for some – a much wider range of Public Cloud services. And some new ventures are born in the Cloud and operate solely using resources in one or more public Cloud providers.
Over the past few years even the most conservative organizations have adopted the public Cloud at least for some of the services. Of course, there are still some specific scenarios when entities are strictly limited in where particular data must be stored, but that is rarely the case for all of the infrastructure. And when legal obligations and contracts are not in play, there are few reasons not to use public Cloud services.
Management overhead and automation capabilities
While knowing exactly where your organization’s data is and almost having the ability to touch it gives a sense of being in control, the reality of facilities and hardware management as well as capacity planning kicks in right after the initial euphoria has passed. A lot of companies are doing quite well with capacity management and expansion planning, but with their own infrastructure a company is simply bound to overspend on new hardware if it wants to be prepared for unexpected workload increases. And capacity planning becomes even less predictable, if unexpected scenarios strike (like the supply chain crisis that we’re seeing now in 2021).
But having a steady facility and physical server infrastructure is just part of the game. While Mike and Anthony might be the friendliest and most dedicated IT engineers in your company, who can build an environment for the infrastructure on virtual machines in just a couple of days right after being tasked with it, a solution that revolves around human beings for infrastructure provisioning each time simply doesn’t scale. It might be OK for a fairly small organization that is not in the business of delivering software, but in a lot of cases less than end-to-end automation is just too little.
The largest public Cloud providers have spent more than a decade making their services easier to use and consume. And not by making new and shiny user interfaces, but polishing their APIs, developing Infrastructure-as-code toolkits and contributing to open source projects that allows engineers to build infrastructure for applications in mere minutes and with no errors each time and even integrate it into the full software delivery pipeline, because that engineer was able to spend some initial development time in coding the infrastructure and then save a tremendous amount of time each time that code is used.
Of course, there’s no reason to state that it’s not possible to reach a similar level of functionality for key IaaS or PaaS services on home-grown private Cloud systems as there are plenty of examples from organizations that have done exactly that. The argument here is more about how easy it is to start and scale with a public Cloud provider versus starting to develop similar levels of automation from scratch.
Common fears of public Cloud adoption
While the benefits of public Cloud computing are more or less commonly understood, there are still some worries regarding primarily data security and Cloud cost.
Public Cloud security is a complex topic as it is comprised of multiple vectors. The first one, physical security is probably something that leaves very little worries. The facilities of the major public Cloud providers are among the most secure data centers in the world and are constantly being audited and certified by different organizations.
However, data locality and accessibility from the Cloud provider poses more questions. As the customer data is seemingly spread across a multitude of datacenters, there’s the impression that the data can be moved all over the world. While this is often true if the customer requires it, the Cloud Providers are very specific in which countries the data is stored and where the data can be moved in case of a disaster. They also state which of their services are not tied to a particular geographic location and make it easier to plan which services are suitable in scenarios where data location is one of the key requirements.
And while data technically is located on servers that are under the management of the provider, the data in the end still belongs to the customer. In order to provide an additional layer of encryption, the vendors have introduced hardware-based confidential computing solutions providing an additional layer of isolation and encryption on top of the usual encryption-at-rest capabilities. Besides that, the customer can always choose to introduce additional encryption at the application level with keys managed outside of the Cloud provider.
In addition to security, the cost of Cloud services is another topic that often gets a lot of attention. Simply by looking at the resource costs of virtual servers, even with applied discounts, there’s no arguing that the prices will look quite a bit higher than what one might be used to in their datacenters. Yes, the reliability of datacenters, built-in automation capabilities and flexibility costs, but for Cloud adoption to actually be something that saves money, a simple lift-and-shift migration will not suffice.
Many companies find themselves paying significantly more than they initially expected after moving to the Cloud because they treated the public Cloud as their own datacenter and might not have included every single line for the Cloud service. It is often the case that such companies are in a rush to leave their old infrastructure and after the initial lift-and-shift does not take on modernization and miss out on much more cost-effective and usually more user friendly PaaS services that Public Cloud providers have to offer.
Another frequent unexpected increase in the Cloud bill comes from the actual simplicity of consuming the cloud. With the speed and convenience of the public Cloud, developers are quick to build new environments, which might be slightly oversized for the purpose or are left running even when they are no longer required. This is a reminder that governance is an essential part of public Cloud cost management and should not be ignored.
Is public Cloud always the answer?
As mentioned in the beginning, there definitely are reasons why wide-scale Cloud adoption might not be the best solution. Besides the already mentioned regulatory aspect, a lot depends on the in-house development and planning capacity and investments that were already made earlier.
Scaling your infrastructure across 50+ datacenters across the world would not be an easy nor cheap task. But let’s be honest – almost no company needs that much scalability. And in case the in-house development teams are capable of ensuring sufficient levels of services, the total cost of ownership for infrastructure, depending on the size of the organization, can be lower than with the Cloud providers.
In terms of Cloud adoption at best there’s a “one size fits most” solution, but definitely not a “one size fits all”. However, in either scenario the experts at Emenit can help in providing the solution that suits your business needs, be it in the public Cloud or otherwise.