Cloud architecture that avoids risk and complexity

One of the truisms about cloud architecture is that there are typically 100×100 architecture solution patterns. They all work, but only one is the right solution. If your head is swimming, let me explain.

Indeed, many technology stacks and approaches will solve the same problems using cloud-based technologies. They all work, but the majority will bring more risk, cost, and complexity than you should endure.  

In my years of reviewing architectures and plans for the Global 2000, I’m often engaged in a struggle between “it works, doesn’t it?” and “what’s optimal?” with the lowest risk, cost, and complexity. Why should we care about this?

Cost is easy. You can spend ten times what you need to, to solve the same problem.  Typically, the architecture team layers on more technology than necessary or doesn’t take advantage of cloud-native features. This means that the applications burn ten times more public cloud resources.   

Often I come upon disturbing realities, such as a technology being used because of an existing enterprise license agreement with that technology provider, which really means “funny money” that needs to be spent.

Risk is another core factor and is not as easy to spot as cost. Overengineering of the cloud solution can cause additional unnecessary complexity, which can lead to more attack surfaces for hackers and the additional likelihood that data on premises or in the cloud will be breached.