Architecture

​

​

Good software architecture reduces complexity. It does not wear a fancy garment that disguises its purpose. At first glance, you can see what it represents, its technicality shines through, and it literally "screams" at you. Not the RESTful ASP.NET core app, but a digital appointment calendar for spa clinics, not the hip GraphQL API, but an ERP system for bakeries, not the snappy Angular app with PyTorch-driven AI in the backend, but a photo retouching app for professional photographers.

The customer's domain is the focus,

not the database,

not the cloud,

not Docker.

​

Challenge


The digital metamorphosis, spurred by events like the - thankfully overcome - Corona pandemic and the growing adoption of cloud solutions, has hit the business world like a hurricane.

Personally, I consider the metaverse as outlined by Mark Zuckerberg to be rather utopian (see also my podcast on this subject: https://www.wackyworld-podcast.de/e/metaverse-bis-zur-unendlichkeit-und-noch-viel-weiter/), but an industrialized 3D digital universe populated by Digital Twins is, in my view, a thoroughly plausible vision of the near future (now: 2023/07/20 - 20:20). It recently became known that Siemens is planning investments of considerable magnitude here (https://press.siemens.com/global/de/pressemitteilung/siemens-investiert-eine-milliarde-deutschland-und-legt-grundstein-fuer).

Add to this the monumental buzz around AI. Hardly a day goes by without a new StartUp sprouting from the ground like a genetically modified mushroom to advertise its services for the benefit of mankind. Many of them will undoubtedly share the fate of a mushroom, either being plucked or, better, ripped apart by the big players (63 billion euros is said to have cost the training of GPT 4 [ https://the-decoder.de/leaks-zeigen-gpt-4-architektur-datensaetze-kosten-und-mehr/] , a sum I don't have liquid at hand right now). Or they are made obsolete by the open source community, which Google considers to be the greatest danger, just think of Stable Diffusion and Llama 2.

Nevertheless, the examination of AI and its background remains my passion (see https://www.wackyworld.de/artificial-intelligence-studies).

The demands on data and its handling are daunting: high volumes, a variety of data formats, fast processing, and variable data quality. Data engineering skills are needed here, and often a fair amount of cloud knowledge as well.

The migration away from on-premise solutions to flexible, scalable and cost-efficient cloud solutions has long since begun. Large cloud providers such as Microsoft, Google and Amazon are leveraging their size to achieve economies of scale and pass these benefits on to their customers.

Despite the enormous competitive and innovation pressure, especially in the automotive industry, many companies lack the necessary resources and know-how to become more "cloudy" or at least "cloud-ready". Additional pressure arises because AI support is offered in many areas and is now also expected by many customers. Anyone who comes around the corner with long release cycles, expensive on-premise solutions that are cumbersome for the end customer to install, and which also solve tasks programmatically that can be better solved by AI training, can quickly be forced out of the market. This is happening faster today than it did a few years ago.



Solution



As an experienced software architect coming from a software development background and an (IT) lawyer with a passion for Artificial Intelligence and a Masters in AI in Progress, I offer a unique combination of skills. I understand both the technical and legal aspects of software development and can assess what is feasible and what is not. And I love automation, preferably with AI. I am well aware of the risks of this technology, but nonetheless, for me AI is the next stage of digital evolution.

What can I offer you? My specialty is software architectures of distributed systems, where modular building blocks communicate asynchronously with each other. I favor microservices when they make sense, and am an advocate of clean architecture and clean code principles when they are appropriate.

 

You can see what I mean by Clean Architecture here:

https://www.wackyworld.de/en/post/clean-architecture-for-those-in-a-hurry

You can "taste" a snippet of my microservices knowledge here:

https://www.wackyworld.de/en/post/microfrontends-in-practice-the-ultimate-guide-for-software-developers-and-architects

I am not a dogmatist, but a pragmatist. I understand that not all principles make sense in every context and that sometimes compromises are necessary. For example, a Clean Architecture approach is not a panacea and not suitable for every project or situation. Because during my years as a software architect I have learned (often painfully) that it is the business value that matters and not the bulging ego of the architect and/or developer. If an application that is slightly removed from youth, has lost its sex appeal, and is based on the classic 3-tier model, still fills the customer's coffers and each new release is not disproportionately more expensive than the previous one, then as a consultant I would advise against switching to Clean Architecture just for the sake of coolness.

As a lawyer, I also have the advantage of being able to legally assess whether something is feasible (DSGVO, copyright) or not.

And as an AI enthusiast, I can usually realize what is feasible much faster than someone who does not "request" AI support.



In this sense: Beam me up, Scotty!