Understanding Distributed Architectures - The Patterns Approach (Category Architecture)

Recently at the YOW conference! 2024 spoke Unmesh Joshi is a renowned distributed systems expert who is a Principal Consultant at ThoughtWorks with more than 20- years of experience in the industry. This talk is a summary of his book Patterns of Distributed Systems. 2023 Martin Fowler (on Martin's website There is a short version of the pattern catalog from this book.). Clearly, in 35 A minute's worth of talk. 30 He couldn't get the patterns out of the book, but here are the themes he was able to uncover.

• Introduction to Distributed Systems Modern systems include cloud services (cloud services)Kafka, Kubernetes and many other distributed technologies.
• Self-similarity of templates - patterns (patterns) Applicable at different levels of abstraction and often exhibit properties of self-similarity (self-similarity). It reminded me of fractals:)
• Open source learning Unmesh uses the “three-implementation rule” to identify patterns by analyzing open source distributed system designs. (Kafka, K8s, Cassandra, ...)
• Catalog of 30 pattern The template catalog created by the author helps to understand the internal structure of distributed systems, with examples from real products.
• Platform Sympathy. A concept similar to mechanical sympathy (mechanical sympathy) in motor racing, which helps to use the system effectively.

Mechanical sympathy is when you use a tool or system with an understanding of how it operates best. You don't have to be an engineer to be be a racing driver, but you do have to have Mechanical Sympathy. Jackie Stewart, racing driver.

• The importance of practical experience Real code eliminates ambiguity and turns abstract concepts into concrete solutions. The author is Github There are repositories with examples of distributed container orchestration systems (ala K8s), messaging system (kafka), nosql database (ala)
• Structure through templates Patterns help organize information and make code more understandable, eliminating ambiguities in documentation. I would add that they allow engineers to use a common language when designing complex systems.
• Architecture as a sequence of patterns Knowledge of templates makes it easier to understand complex architectures, as well as design your own systems.
• Harmonized core and metadata management An example pattern where a small cluster with a high degree of consistency (high consistency) metadataman (metadata)This allows the main cluster to scale to thousands of nodes. (nodes). In fact, this is a standard concept with the separation of control plane and data plane.
• Application of templates in real systems - the author has a demonstration of the implementation of patterns on the example of mniature versions of Kubernetes and Kafka for training (You can search the Github account of the author I cited above.)

And the final point of this talk is that these patterns help reduce random complexity. (accidental complexity) and clarify the essential complexity (essential complexity) distributed-system (distributed systems)This helps to deal with such a complex topic:)

#Software #Architecture #DistributedSystems #SystemDesign #Patterns #Sofware