Is a service run at a single physical location for a specific need, but lacking self-service capabilities and a pay-as-you-use model, truly a cloud service?
What if you do not need to make an up-front investment but pay a monthly, weekly, daily or hourly fee for a service? Is the computing capacity delivered by a managed service provider a cloud service? Or are the services delivered by service providers such as Amazon Web Services, Microsoft or Google, Salesforce and so on the only true cloud services?
So, what is a cloud service? There are as many definitions as there are people writing about the issue. One very good definition of cloud services is provided by the National Institute of Standards and Technology (NIST). It is a few years old, but still a very valid definition and a summary of it is given below:
Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model is composed of five essential characteristics, three service models, and four deployment models.
More detailed definition information can be found here.
Let's compare this definition with the examples given earlier.
Can a specific service, such as a contact centre solution running in a specific physical location, be called a cloud service? Well, I would argue with that it cannot, since you cannot assume that, in the case of an on-demand service, there won't be a need to sign a contract before agreeing with the service provider on the kind of contact centre solution you would like. We cannot say that we have a highly standardised solution if it requires customisation to fit in with specific needs, or an on-demand self-service if it requires a service provider to tweak the solution.
A managed service provider provides automated computing capacity, such as virtual servers, in one or more physical locations based on one or more virtualisation technologies. In such a case, we might have some limited self-service capabilities in place for each IaaS platform and the service provider will probably have added some standardisation to ensure competitive pricing. There may even be a pay-as-you-use model and fast delivery, but definitely not infinite scalability or public availability, since this would require an agreement between the customer and service provider and the customers cannot determine the operating systems. This would be closer to a cloud service, but a great deal would still be required in order to fulfil the definition.
Finally, if we take cloud platforms such as AWS, Azure or Google, I would argue that they provide exactly the same features as the above examples, but are not infinitely scalable. They have limits which can be modified, but they are still limits. Some people may regard such limits as the standardisation required in order to provide the service. There is no doubt that these players have much higher capability to enhance the user experience and improve the packaging of different services for consumers, but this is ultimately about hardware-based networks, servers, storage, applications and services.
Does it really matter to end users what kind of platform, virtualisation technology or hardware the service is run on, as long as the service is relevant to and provides value for the consumer? Does a consumer really care whether service providers such as Netflix or Salesforce use AWS, Azure, Google, a managed service provider or a hand-crafted self-built solution, as long as they can use the service as expected? Does a customer really care where a mobile pay solution is running, as long as it is secure to use, you can trust the service and it works every time you need to make a payment at a grocery store or petrol station? The solution must meet the need in question.
I feel that many people are starting to use the cloud because it is a must for staying relevant in their business, without really thinking about why they are doing so. The same applies to service providers, for whom the question should be: Why I am building a cloud service, what problem am I trying to solve, what do I aim to provide for consumers and is it what they expect?
For consumers and customers, the question should be: What will I solve by moving my computing capacity or application from one place to another? How will that help my business to become more relevant for my customers?
I would argue that no business can benefit from just starting to use any "cloud service". Instead, this may simply complicate systems that are important to the business and could even slow down service development. Cloud services, in whichever form they come, should be used to gain agility, enable enterprises to establish new business models through digitalisation, and make business application hosting easier and more transparent.
Everyone considering the cloud should forget about holding onto a specific hardware and hypervisor technology and calling it a cloud service. Instead, they should ensure that they gain an agile, trustworthy and reliable platform on which they can build new business opportunities and services. Cloud services should be built to enable the consumer to consume, allowing the consumer to transfer, expand or split their activities between one or more cloud platforms. The capacity layer should be abstracted from the service, no matter how great it looks, and consumers should be allowed to "just use the capacity" in order to achieve the desired outcome.
As I see it, a cloud service is much more about changes in people, in their mindset and thinking, than changes in technology. Virtualisation, automation and containers have been around for a long time, so we should now focus on other service aspects. Once people and their cultures are ready to utilise or provide cloud services on the right grounds, the journey into the cloud will have much more chance of succeeding.
To keep up with such demand, it is equally if not more important that service providers begin by changing their culture. Only then will it be possible to start building services based on service design principles, where the core element involves collecting data in order to build relevant solutions. I will cover service design in greater detail in my next blog. Once all of the above elements are in place, service providers can provide solutions which enable businesses and service consumers to be agile and innovative in relation to services. This will ultimately enhance the service consumer's experience. The time for selling or buying automated capacity is already long gone.
I would therefore encourage everyone to begin by thinking carefully about the problem they are trying to solve and whether their culture and mindset support new practices, before embarking on a migration to the cloud.
If you have a different opinion, have any comments, or share my views, I would love to receive your feedback through my social media channels.
Toni is a senior solution architect with the Tieto Hybrid Infrastructure and service design team. He has been working for several years as an architect and trusted advisor at the customer interface, involved in the design of various solutions in the ever-changing infrastructure and cloud landscape. He has recently developed a greater interest in service design and design thinking methods, with the aim of ensuring that Tieto’s services provide value for its customers.