As so many enterprises confidently add an “s” to their “years” of cloud computing experience, some traditional IT architecture concepts are starting to resurface. These concepts can make or break a cloud implementation, but they are largely forgotten by those who design, build and deploy today’s cloud computing systems. How did this happen?
Two factors: First, many architecture certification courses from specific cloud providers omit a lot of general basic architecture training. Second, many modern tools eliminate the need to think about the details that surround a cloud architecture solution.
All cloud architects must fully understand key IT architecture concepts to create the most optimized cloud computing architecture, be it single or multicloud. Here are the top three overlooked concepts:
Abstraction. This concept requires very complex things, such as poorly designed databases, overly complex network designs, or overly complex applications, and provides a more simplified picture for the person or application using these resources.
Perhaps the best example of abstraction in cloud systems is data virtualization, where layers of abstraction or virtual data structures are placed over any type of physical data storage system. No matter how poorly designed the database and how many applications are closely tied to the physical databases, you can leverage the data using custom structures mapped to each back-end database structure.
The bottom line is that you can handle any complex or poorly designed database with your own access structure that provides a layer of abstraction above the physical database structure. Because you don’t change the physical database, you don’t force all applications associated with the database.
Orchestration. If you ask in a cloud architect meeting who understands orchestration, chances are all hands will go up in the air. Most cloud architects have experience with the concept of orchestration through container orchestration systems, such as Kubernetes.
Orchestration is the automated coordination of computer systems, applications and services. Like abstraction, orchestration helps cloud system designers more easily manage complex tasks that need to coordinate actions between systems, applications, and databases, at least as applied to IT architecture, including cloud architecture.
The reality is that orchestration is a much more powerful concept when considered in a typical cloud architecture, especially multicloud. We need to think more about building orchestrations and abstraction over public cloud provider groupings. Too often orchestrations and abstractions only exist in the walled garden of a single cloud provider, which only makes your multicloud more complex as you have to orchestrate the orchestrations.
Automation. Everyone knows what automation is, right? Yes, we know the definition of the word, but maybe not what it means in the context of a cloud computing architecture. Automation is the process of looking at all systems and making them more efficient. The goal is to automate manual or overly complicated processes so that they can function autonomously, which logically should make them more efficient.
As for cloud computing architecture, automation takes away the need for people. The fewer people we need to run core processes, applications and operations, the more complete our cloud architecture solutions will be. We maximize the business value of the solution by maximizing the business benefit at the least cost.
While automation can also apply to traditional IT architecture automation, cloud architects often overlook automation capabilities. Many think they already understand the definition, but they may not understand exactly how to use automation to improve single or multicloud solutions.
Abstraction, orchestration and automation can help you address the complexities of today’s cloud computing architectures. Do yourself a favor. Learn more about them.
Copyright © 2022 IDG Communications, Inc.