Cloud journey and its primary aspects
Have you ever been asked if moving to the cloud is worth than keeping your infrastructure on-prem?
Many people ask me every day about cloud technologies, commonly, they already know some of the benefits that cloud providers like Microsoft, Google, Amazon provide. But commonly they’re facing doubts when adopting cloud technologies.
If you’re worried about the cloud journey and you ask me about cloud adoption. I’ll ever say “It depends”. In my experience, I’d say that there’re two main points that you must consider in a cloud journey (among many others).
Strategy and sizing are two basic things that will impact your experience in cloud. The strategy will impact the adoption and sizing will impact every time you get a new invoice from your cloud provider 😅.
From the strategy side, I usually say that it’s so big that’s impossible to cover all aspects in few words.
Taking Microsoft Azure as an example of the cloud provider. Microsoft has what it’s called Shared responsibility model, which defines what is Microsoft responsibility x what’s your responsibility. Why is it so important? Because it defines how much work and money you’ll have to spend on your journey.
Some of the most basic aspects that you must take into account when planning your on-prem infrastructure are:
- Physical Hardware and its upgrade
- Physical security
- Networking appliances
- UPS and electricity
- Cooling
- OS-Patches
- High-Skilled people
These points are just the beginning, you aren’t even close to launching your application. That’s why I call strategy a must in the cloud journey.
Do you have enough Capex? When moving to the cloud, your costs are fully operational (Opex) once you haven’t spent money buying underlying things.
In a shared responsibility model, the cloud provider I’ll take care of such primary things like these for you, even if you have chosen IaaS (Infrastructure as a Service) over PaaS (Platform as a Service).
IaaS x PaaS is another mandatory aspect of strategy that you must decide on in the cloud journey. Are you willing to allocate high-skilled workers to manage database, OS, firewall configurations, and many others components to host your applications?
In PaaS services, Microsoft, as the cloud provider, will manage all components and keep them up-to-date, and more, in case of any failure, Microsoft will manage everything for you. In this sense, your focus is on using it not configuring.
Bringing it back to costs, PaaS services tend to be more expensive than IaaS services. For example, to host a database we can use Azure Pricing Calculator, to size it. Supposing that we are using MySQL as RDBMS in a small environment, we can notice that the MySQL PaaS service is significantly expensive than a regular VM.
In another hand, you’ll have to hire some skilled workers to configure your database before starting to use it.
Let’s take sizing into account to avoid surprises. In an on-prem environment, you certainly have extra hardware, predicting some access spikes in events like Black Friday. In this sense, it’s very common to see oversized resources.
For example, when creating a VM to host a web server you allocated 8GB RAM instead of 4GB because you don’t want to worry about it so soon (Yes, it happens a lot 🤷🏼♂️).
In cloud, this strategy will leave you in trouble. Considering that you’re billed for what you’re consuming, you’ll be paying for 8GB. That’s why the sizing is critical if you don’t want to be surprised when getting your invoice.
A relevant component in cloud platforms is autoscaling, it helps you to save money. In these scenarios, you’re scaling as your traffic grows. Some examples of autoscaling are VMSS, App Service, Load Balancers, etc.
Another opportunity to save money in cloud is using reserved instances. From the base resources needed you can sign a commitment to reserved instances. In some cases, it can help you to save up to 60% on some specific resource.
In practice terms, the cloud provider will dedicate that hardware for you in that certain period that you’ve committed.
From reserved instances, you can choose between SQL Databases, VMs, and storages. We can notice that these features are available in both models of infrastructure, IaaS, and PaaS.
Finally, Microsoft provides a Total Cost of Ownership (TCO) calculator where you can input your existing workloads and the underlying costs, and compare how much you will save when adopting Azure services.
When moving to the cloud there’re many others aspects that you might want to consider in your strategy related to reliability as latency, planned maintenance strategy, etc.
Also, another must is to take a look at the Service-Level Agreement (SLA). Microsoft has its SLA, in which you can check at Microsoft SLAs page for each service before starting using it.
