All through my profession as a practitioner and trainer of cloud structure and normal enterprise structure, I’ve centered on the excellence between logical and bodily architectures.
It’s no secret what they comprise: A logical structure is a construction of expertise ideas that doesn’t title particular applied sciences or manufacturers, whereas a bodily structure does include particular applied sciences and types.
After I design a cloud computing structure, I first ask what issues have to be solved. I then decide what an appropriate answer would require. Subsequent, I checklist the objects that have to be a part of the structure and their relationships to one another: databases, middleware, encryption, storage, compute, software growth, and so on. Once more, no particular applied sciences, solely technological ideas that may contribute to fixing the general issues.
The first profit to this method is that it permits the architect to assume extra objectively about an total technical answer. Whenever you title a selected expertise or model of expertise, you’ll construct the answer round what works with that particular product and model.
We’ve been making this error with public cloud suppliers for years. We choose a selected supplier earlier than understanding the core enterprise and technical necessities (aka, the issues the expertise should clear up), after which we proceed with an answer that’s native to that particular cloud supplier. By selecting the model too early within the course of, we should then select expertise that matches inside the constraints of a single cloud supplier. Greatest case, the answer will run on underoptimized expertise. To totally optimize an answer, we have to select the perfect expertise for the job after which select the perfect suppliers.
That is the largest mistake I see architects make as of late: selecting public clouds earlier than they know and perceive the issue. It occurs most frequently with cloud-based databases that restrict the manufacturers of middleware, enterprise intelligence methods, database admin instruments, and so on. Limiting your selections too early within the planning course of nearly all the time turns into a mistake.
Sadly, firms typically find yourself with a cloud structure that works. “It really works” doesn’t describe an optimum answer. The functional-yet-less-optimal structure will ship far much less worth than one which leverages the best expertise to unravel the particular enterprise issues at hand.
To recap, a logical structure accommodates an inventory, configuration, and interrelationships with expertise as ideas. After getting a logical structure with conceptual options to particular issues, then and solely then do you progress to the bodily structure.
In different phrases, map conceptual applied sciences within the logical structure to precise applied sciences within the bodily. The bodily structure turns into the final word answer candidate, however the expertise should nonetheless be verified with testing earlier than you progress to any kind of deployment.
There’ll nonetheless be bumps within the highway with this course of. As an example, you could outline a expertise idea within the logical structure that doesn’t exist. Particular applied sciences available on the market could lack sure choices and options, resembling a database that doesn’t have built-in danger analytics for the kind and mannequin of database you logically outlined. You’ll want to regulate, however these points are the exception fairly than the rule.
Additionally, cloud architects must have a working information of normal expertise ideas in addition to precise applied sciences that may map to those ideas.
If you happen to do logical structure planning earlier than committing to a selected expertise, you’ll make higher selections. This additionally removes the “architects are individuals” issue since most architects have expertise preferences and favorites. They may have a bias for a selected cloud supplier, database, safety method, and so on., which finally ends up being an answer that works however is way costlier than a extra optimized one.
Take into account that your aim as a cloud architect—or as anybody who defines enterprise options from expertise—is to create an answer that returns probably the most worth to the enterprise. Use the logical structure method first to present your self a higher likelihood to succeed in that aim.
Copyright © 2022 IDG Communications, Inc.
