There is no getting away from the fact that cloud can be complex, as enterprises have to make sense of the varying propositions of myriad providers, and the veritable jungle of options and services they have to offer.
Finding a provider whose services are a good fit for your company and its requirements can be quite simple if done correctly, but there are plenty of variables that can stand in the way.
It is worth remembering that cheapest does not always equate to the best in the business world, as any seasoned professional will tell you, and loss of service can easily result in loss of customers and revenue in the worst-case scenario.
There seem to be some very specific cloud suppliers frequently mentioned by management, but they do not seem to understand the options, costs and flexibility that come at a not insignificant premium.
The reality is many cloud service providers appeal to specific markets and do one (or several) things and do them well, rather than dozens of things in an average way. These markets are businesses with different business drivers, capabilities and outcomes.
To put it into context: Developers have different requirements to web designers. Developers want to be able to stand up a host instance quickly, with minimal intervention.
Therefore, a development house may look at how good and complete the available API is, and how easy it is to automate and stand up a virtual machine in it that may only need to last a few hours for development and testing, rather than longer-term use in a production environment.
Read more about cloud deployments
- Nationwide Building Society is in the throes of a cloud and DevOps-focused effort to re-platform its digital banking and mortgage services.
- Barclays Bank has revealed it is two years into a digital transformation project that will see it shut datacentres and go all-in on the AWS public cloud.
Being able to call the API means the build of a virtual machine can be automated and may last minutes, hours or days, which recently gave rise to the concept of per-second billing.
Those environments with rich environmental APIs that automate often have increased flexibility, but can come at a monetary cost, and the cost verses flexibility debate is a critical one for businesses to get involved with.
Compare and contrast this to a web developer who just needs a cloud-based hosting environment for SMEs and micro-businesses. For several dollars a month, a web developer can have a robust test and development environment.
A lot of small developers want a simple service. Indeed, some of the top-tier suppliers offer an overly complex set of options, and in business, time is money. Excess options do not always make happy customers, who adhere to the mantra: “Do what is needed and get out”.
It may lack the bells and whistles, and perhaps even the high availability that comes with some providers, but when the bill is paid for on a monthly basis, the upfront costs may not be that much cheaper per virtual machine but the specifications tend to be better.
For example, provider A offers 1GB memory and one CPU core for £5. Vendor B, who is not as “API driven”, may have a cost of £5 but comes with 2GB memory and two CPU cores.
It all comes down to what is required to satisfy the business requirements. For instance, load balancing, backups, per-second billing and APIs all come with costs. Smaller businesses usually do not require the more advanced capabilities that the bigger, web-scale businesses, or startups often do.
Furthermore, the underlying functionality can be further whittled down to the realms of serverless computing, where the cost discussion starts getting extremely complicated, not least in developing the application itself.
In summary, though, there really is no one-size-fits-all type of cloud supplier for every situation, and before committing to using a specific one, consider the following.
- What do we want to achieve for the business? Do we just want hosting for our websites or do we need capabilities that will make our services more dynamic and reactive?
- Do we have any existing in-house expertise with specific suppliers that may fit the bill and provide us with the functionality and capabilities required?
- What functionality is key to a successful outcome (management capability, automatic backups, service-level agreements, disaster recovery, etc).
- Do we use (or intend to use) automation or scaling? Nine times out of 10, the answer is no, despite what some cloud administrators may say.
- Does the provider need to offer team-based infrastructure, role-based capabilities, management and auditing?
- The simpler the interface and fewer the options, the easier the management can be. This, however, depends on your requirements. Complexity is the enemy of simplicity and uptime.
For most, the answer usually comes down to billing period (months verses seconds or minutes) and resiliency (high availability and redundancy versus single instance servers with no disaster recovery infrastructure.
For small test and development servers without the complexity, small suppliers often win out in terms of just getting things done.
At the end of the day, it really comes down to doing quality research and ensuring the chosen company provides the functionality needed without adding in a load of billable extras that can cause confusion and make things more complicated than they need to be.