Jakub Nabrdalik had a talk about mistakes made when moving to microservices at GeeCon Prague 2019. You can watch it also in polish from Confitura 2019.
He did a thesis that majority of companies that migrated to microservices from a legacy monolith made similar mistakes. From his experience as a consultant, he described what was done poorly and how to do it better.
Big part of talk was about software architecture and how it was enforced in companies. My favorite quote from this part is:
Architecture is the shared understanding of people on the project.
My conclusions that came from this quote:
- all developers create the software architecture consciously or not
- sharing knowledge is the best way to enforce architecture changes
- all developers must have a single baseline to compare their individual understanding of architecture
Jakub Nabrdalik introduced a solution that covers part of these conclusions – architecture guilds. Guilds are a way for spreading and cultivate the knowledge. To be effective, they need to be open and share there work with whole company. Members of the guild have to guide and teach.
To cover the rest of these conclusions Jakub Nabrdalik was convincing the audience to implement C4 model of theirs software architecture in open for all developers repository. This way every developer could introduce changes as pull requests that would be reviewed. This repo would be the baseline for everyone to compare.
He covered other topics of migration from monolith to microservices. I will not cover them, so if you want to know more go watch it.
Event-Driven Systems & CQRS
Jakub Nabrdalik in his talk also recommended another book – Designing Event-Driven Systems by Ben Stopford. You can download it right now for free from Ben Stopford’s site. I did it and read it in less than a week.
Ben Stopform is CTO of Confluent – a company that describe itself as The Complete Event Streaming Platform For Apache Kafka.
No one will be surprised that his book about the event streaming is based on Apache Kafka. Fortunately, this book is not just an advertisement for technology. Examples are made in Apache Kafka, but the knowledge is universal.
I like the fact that this book introduces more and more difficult concepts one by one. Each concept is break down with an example. All examples are about the same use-case, so it easy to compare them.
I like this book, because it gave me confidence on how and when to use the event streaming. Here on of my favorite quotes from this book. I hope they will encourage you to read it.
Centralize an immutable stream of facts. Decentralize the freedom to act, adapt and change.
Event Sourcing ensures every state change in system is recorded, much like a version control system.
If an event stream is source of truth, you can have as many different views in as many different shapes (…) as you may need.
If you squint a bit, you can see the whole of your organization’s systems and data flows as a single distributed database. Jay Kreps, 2013
The more mutable copies, the more data will diverge over time.
Talk at Poznan University of Technology
On Monday I had a presentation at put.net. It is the science club at Poznan University of Technology. Their main focus is C#, but they talk about many other aspects of developing software. Jakub Riegel, who takes care of Web Development section, invited me to talk about REST API. I was very happy to help and share my experience with students.
I introduced them to common REST API implementation with HTTP and JSON. Oliver Drotbohm’s presentation inspired me to show students usage of HATEOAS. Here you can read the slides from my presentation.
- Hermitage - project that verifies how isolation levels are actually implemented in different database systems. Dr Martin Kleppmann wrote it as a background for his book – Designing Data-Intensive Applications. If you don’t believe Dr Martin Kleppmann, you can check them yourself. He left instructions to manually verify his summary. Currently project contains tests for: PostgreSQL, MySQL, Oracle DB, MS SQL Server and FoundationDB
- Awesome Conference Practices by Kitze sums up what you need to do if you want to organize a conference
- Awesome EventStorming by Mariusz Gil. It is a collection of resources – books, articles, videos, presentation slides, people to follow on Twitter, projects and communities – about Event Storming. If you are interested in the topic, you should star this repo and contribute.
- Microsoft released article – Reasons to move to Java 11 – that covers changes between Java 8 and Java 11.
- Claes Redestad from Oracle compared startup times for OpenJDK from Java 8 to 14. I am impressed with changes that came with new Java versions. Read it and see for yourself.
- Angie Jones and Dr Nicole Forsgren did a webinar – Test Automation as a Key Enabler for High-performing Teams – this week. It is based on book Accelerate and revelations that came out of it.
- Article The Service Mesh: What Every Software Engineer Needs to Know about the World’s Most Over-Hyped Technology by William Morgan debunks myths about service mesh. It describes what is services mesh and what isn’t. If you want to know better how service mesh could be helpful and when to use them, you should read it.
- If you missed, ToughtWorks released Tech Radar vol. 21 this week. I wish to cover it in detail in the next week article.
- Article How to hire sane engineers 2.0 by vas3k. It is suited for HR, but for me it is always interesting to see things from the other side. I also like to compare and benchmark myself by measures in those kinds of articles.