15 Jun The Myths of Cloud Computing
The Cloud Computing paradigm has redefined and changed, not only the technology stack that vendors utilize to provide solutions but most importantly how the user experiences those solutions.
Let’s not lose our focus here. As software vendors or at least practitioners, we love the idea of a simple infrastructure, easy environment setup, continuous integration, scalability, and the list goes on. But we must keep in mind that ultimately, we are doing this so that the user has the best experience ever and chooses us day after day over our competitors.
Having said that there are lots of myths and misconceptions about the cloud that I’d love to address. Not everything is as it seems. Sometimes we are presented with alleged benefits, that might turn out to have hidden costs or factors that ultimately make that benefit turn into a disadvantage.
Alleged benefit 1 – Cost-effective:
The cloud “does not need any investments on physical investments or any kind of expert for the management as it is handled by the cloud service provider.” (Anuja Lath, 2019)
I wouldn’t say that is entirely true.
One key factor to be running a cost-effective cloud solution is to fully understand your usage consumption model. If you know your servers will on high demand from say 10 am to 5 pm, then you can prepare your infrastructure to scale in such period, and quickly downgrade in downtime usage periods, saving your budget.
If you have multiple servers running 24×7, on-premises solutions might give you a better performance for the same money.
Even more hardware vendors are continuously dropping prices to compete with cloud providers. Particularly hyper-converged infrastructures have proven to be more cost-effective compared to cloud infrastructures.
Take a look at the following comparison between cloud vs on-premises infrastructures in Europe and the UK:
Source: Matthew Beale (2018), “Cost comparison – cloud vs on-premise 2018/19”
You must also take into consideration the economic impact that represents having an experienced DevOps team required to deploy, support, and scale cloud solutions in any of the major cloud vendors available today. Even more, specialized DevOps engineers that are certified in public cloud infrastructures tend to have higher salary wages.
With all of this into account, I hope the cost-effective myth is more clarified, and we all raise awareness on how to better manage our cloud services.
After all, the cloud is about money.
Alleged benefit 2 – Easy implementation:
“Businesses don’t have to deal with the backend technicalities.” (Stratosphere Networks).
Even though the cloud does provide pre-built solutions that ease the initial deployment of applications, chances are your service might start being simple at first and easy to handle but it will most definitely grow to become more and more complex turning that so-called “easy implementation” into a complicated engineering problem.
Alleged benefit 3 – Easily scalable:
Let’s picture the following scenario, you have an EC2 server running and handling several requests. Your service starts growing, more and more requests keep coming in until you reach the point where your service starts responding with HTTP error code 408 Request Timeout. You might think, hey I’m in the cloud, no problem, I’ll just upgrade my server to twice the capacity and problem solved … well, that might help but it may as well not solve your problem at all. You most likely will have to deal with your application server configuration, workers, threads, calculations, and lots of other engineering things to finally solve the scaling problem properly. It might sound easy, but it’s actually far from that.
Alleged disadvantage 4 – Downtime:
“Downtime is considered as one of the biggest potential downsides” (Hemant Sharma, 2020). You might see this throughout several pieces of literature. Though it is definitely an issue you must consider, especially for mission-critical applications, through be said that, according to Gartner, “Amazon Web Services and Google had nearly identical uptime statistics for the virtual machines at the heart of cloud services — 99.9987 percent and 99.9982 percent, respectively” (Tom Krazit, 2019).
If we dive deeper, according to a study made by Zeus Kerravala, from the beginning of 2018 through May 3, 2019, AWS had only 338 hours of downtime, followed by Google with 361 while Microsoft Azure failed far behind totaling 1,934 hours of downtime.
Therefore evidence shows that downtime is not really as important factor as we may have through initially.
It all comes down to selecting the right application and the right vendor to move to the cloud.
I’m not saying the cloud is misleading and you shouldn’t use it, in fact, we @Cualit believe just the opposite and I’m the first one to promote it because of the beautiful innovative solutions and experiences you can accomplish, but I do say be prepared, especially when you are scaling.