Article 1: From monolith to microservice — A series on the future of enterprise software development.
I’ve lately made the slightly aggressive 5–5–5 projection: in 5 years 5 computer science students will be able to code a complete Oracle Peoplesoft implementation in just 5 weeks from scratch. I believe this will be possible because of three core factors: reusability of code in microservice architectures, joint authentication layers and automation in configuration. I have the pleasure to spend my workdays alongside an amazing group of people here at Humanitec making software development fast, fair and fun plus I’m quite an optimist so I might be biased. Nevertheless I deeply believe it’s utmost time to push this topic because the way we write code today is redundant and in many ways ancient. This is the first in a series of articles I’ve written to explain why I think we will witness one of the biggest shifts in the history of this industry and see software step into the sharing economy. This will have a wast impact on provider of legacy systems, on monetization models, team setup and the way we work and collaborate in software teams. Super exciting times ahead!
We start off with the evolution of enterprise software and the underpinning technical breakthroughs. Based on that, we point out respective cutting-edge developments in 2019. Finally, we depict where we believe enterprise software is heading towards and what this has to do with the sharing economy.
Let’s get to it
Over the last decades, enterprise software has evolved in three waves: From 1982 and up to the 1990s, the industry was dominated by firms like SAP, HP, and Oracle with strictly closed-source monolithic enterprise software solutions running on-premise; In the 2000s, firms like Salesforce started providing cloud-native enterprise software running on-demand; in the 2010’s, it became increasingly apparent that reality hasn’t lived up to the widely-held prediction of one dominant player arising in every software category — and that one-size-fits-all solutions, as exemplified by the infamous Lidl and Haribo cases in Germany, are doomed to fail. As Handelsblatt puts it precisely: “another example of Germany’s digital failure”.
So what’s next? To grasp where we are heading in enterprise software development, we will start off the first article of our series on enterprise service software in the 21st century with some history, Harari-style.
Enterprise software, ancient Rome and the Victorian Era
Chris Stephenson has a background in building internal developer tools as a product manager at Google. He compares the evolution of enterprise software development to how the construction industry unfolded over the last millennia. The Romans layered massive stones on top of each other. Everything used to be custom made, which is extremely redundant and virtually unplannable. As a result, the deadlines for the completion of theaters, temples and bathhouses had to be adjusted again and again. Nowadays — and with rare exceptions like major airports in Berlin — construction projects in developed countries are usually finished exactly to the day. Only in the rarest cases one would literally build a window made to measure but rather use a standard of the shelf.
On the contrary, the development of customized enterprise software applications is still a highly individual process. Yet, this isn’t surprising. We’ve been making software for a couple of decades now — so even if the evolution in bytes is twice as fast as the one in bricks, we’re still in the midst of the Victorian Ages. As a result, ever-growing monolith systems require a tremendous effort today, both from internal and external developers in firms: For internal IT organizational structures, on the one hand, this is hardly manageable, as exemplified by the Toyota/Microsoft Visual Studio case made by James Somers in his recently published article “The coming software apocalypse” that drew much attention. Custom software providers, on the other hand, can seldom reuse anything from their last project because most architectures are subject to little standardization. As the coders among you will be well-aware, this ranges from programming language to frameworks and coding conventions and social authentication, user administration, and search functions need to be built again and again. In a nutshell, monolithic enterprise software solutions are both costly and resource-intensive for everyone involved except for their providers.
Digital aqueducts and railways
As a result, businesses are increasingly relying on many small, specialized systems that are employed in concert, just like a mosaic. As pioneered by U.S. trailblazers (e.g. Netflix and Twitter) and API-based providers (e.g. Twilio and Mulesoft), we’re now witnessing a shift from monolithic to microservice-based solutions that are kept in sync through API connections. To come back to Stephenson’s analogy, APIs are like digital aqueducts that provide standardized waste and water pipes. They resemble the railways built in the Victorian Era that allowed goods, raw materials, and people to be moved around the country — which facilitated trade and industry, not least via the formation of economic centers and industry clusters. This importance of APIs has been taken to heart even by former closed-system behemoths, most notably Microsoft, which is often arguably decried as the archetype of the monolithic system firm. Yet, under the helm of its new CEO Satya Nadella, Microsoft has recently made massive inroads into open API systems. This culminated in the acquisition of Github in October 2018.
Technically, these developments are enabled by a combination of steady progress in cloud-native environments, APIs and all sorts of evolutions around microservice architectures. In our next article, we will thus elaborate on the technical breakthroughs underpinning the “monolith-to-microservice” shift in enterprise software development described in this article. These evolutions will bring us one step closer in automating configurations and reducing redundancy in enterprise software development. All to help developers focus on true innovation and prevent to manually build the same functions over and over again.
Article 1: From monolith to microservice — from ancient Rome to the Victorian age.
Article 2: Goodbye stone age: Technical breakthroughs in enterprise software development (coming up soon)
Article 3: Quo vadis: Where enterprise software is heading — the 5–5–5 projection (coming up soon)
Article 4: Speak flats, cars and offices: How software moves into the shared economy, too (coming up soon)
Article 5: Enterprise applications at lightning speeds: How Humanitec disrupts the industry (coming up soon)
