MICROSERVICES IN A MESSAGE-BASED WORLD Software Architect at ServiziCGN @manuelscapolan Manuel Scapolan 

Enterprise Applications Usually • involve persistent data There • ’s usually a lot of data Usually • many people access data concurrently There • ’s usually a lot of user interface screen Usually • they need to integrate with other enterprise applications What • about “business logic”? from Patterns of Enterprise Application Architecture (PoEAA - Martin Fowler) 

Monolithic Architecture • Simple to develop • Simple to test • Simple to deploy • Simple to scale Business Layer Presentation Layer Data Layer A good starting point… user 

Until it grows… • Large code intimidates developers • Small change – big impact • Long term commitment to technology stack • Little resilience to failure • Scaling can be difficult 

Vertical scaling • Adding more hardware… – CPU/core – RAM – … Business Layer Presentation Layer Data Layer Business Layer Presentation Layer Data Layer 

Horizontal scaling • Running multiple identical copies of the application behind a load balancer Business Layer Presentation Layer Data Layer Business Layer Presentation Layer Data Layer Business Layer Presentation Layer Data Layer 

Data partitioning Each • server runs an identical copy of the code Each • server is responsible for only a subset of the data (i.e. sharding) Business Layer Presentation Layer Data Layer Business Layer Presentation Layer Data Layer Business Layer Presentation Layer Data Layer data segment data segment data segment 

Functional decomposition • Splits a monolithic application into a set of services (or microservices) A B A B Microservices Traditional development model 

What are microservices? 

Microservices Service-oriented architecture composed of loosely coupled elements that have bounded contexts. Is an approach to developing a single application as a suite of small services, each running in its own process and communicating with lightweight mechanisms, often an HTTP resource API. These services are built around business capabilities and independently deployable by fully automated deployment machinery. James Lewis and Martin Fowler - ThoughtWorks Adrian Cockcroft – Cloud Architect at Netflix 

Characteristics of Microservices • Small (how much?) • Built around business capabilities (bounded context) • Independently deployable • Little centralized management • Smart end-points and dumb pipes • Lack of centralized, shared database 

A modern SOA? 

Classic SOA • Integrates different applications as a set of services WS* WS* WS* WS* WS* WS* WS* WS* Enterprise Service Bus 

Microservices architecture Architect a single • application as a set of services Backends Catalog service Ordering service Inventory service API API API Customer service Accounting service API 

Microservices at Netflix 

Communication mechanisms in a microservice architecture smart endpoints and dumb pipes 

Synchronous HTTP • As REST or SOAP Request/ + reply Firewall + friendly Doesn – ’t support publish-subscribe Both – the client and the server must be simultaneously available HTTP clients – needs to know host and port of the server 

Asynchronous messaging • As an AMQP-based message broker or a message bus + Support one-way requests and publish-subscribe + Producers are completely unaware of the consumers – Request-reply style is not a natural fit – Another moving part that add complexity to the system 

Communication issues • The granularity of APIs provided by microservices is often different than what a client needs • Different clients need different data • The number of service instances and their locations (host+port) changes dynamically • Partitioning into services can change over time and should be hidden from clients 

API gateway • Single entry point for all clients • Some requests are proxied/routed to the appropriate service • It handles other requests by fanning out to multiple services API gateway Service A Service B Service C Service D REST REST REST AMQP Protocol transaltion 

Decentralized data management each service has its own database (schema) 

Handling reads Order Service Customer Service PlaceOrder() VerifyCreditLimit() SaveOrder() The Customer Service must be running in order to place the order Calling thread is waiting for the result 

Handling reads (another approach) Order Service Customer Service PlaceOrder() CreditLimitUpdated SaveOrder() The Order Service stores a local copy of the credit limit StoreCreditLimit() 

Handling update requests • Distributed transactions  + Ensure that the data is always consistent – Reduce availability since all participants must be available in order for the transactions to commit – Not supported by modern stacks, e.g. REST, NoSQL databases, etc. 

Handling update requests • Event-driven asynchronous updates Message broker/bus Customer Service Order Service UpdateCreditLimit() CreditLimitUpdatedEvent CreditLimitUpdatedEvent 

Handling update requests • Event-driven asynchronous updates ☺ + If a consumer isn’t available to process an event then the message broker/bus will queue the event until it can – The application has to be written in a way that can tolerate eventually consistent data – Developers might also need to implement compensating transactions to perform logical rollbacks 

Organization & agile practices 

Conway’s Law If the architecture of the system and the architecture of the organization are at odds, the architecture of the organization wins. Organizations which design systems ... are constrained to produce designs which are copies of the communication structures of these organizations. Mel Conway, 1968 Ruth Malan, 2008 and 

Traditional team structure Separate silos • Source: Adam Cockcroft 

Microservices model • A team handles all aspects of software development for the microservice, from conception through deployment Source: Adam Cockcroft 

Microservices & DevOps 

The role of Containers • Create small independent deployments 

Handling failures Services can • fail at any moment Design • services to handle these • kind of failures Pattern: • circuit breaker, tolerant reader, bulkheads, timeouts, … Generating • failures with a “Chaos Monkey” Circuit Breaker - martinfowler.com 

How to get started 

Refactoring a monolith • Stop implementing significant new functionality by adding code to the monolith • Find a way to implement new functionality as a standalone service • You will have to write messy, complex glue code Monolith New Service Glue code 

In summary… 

Benefits of Microservices • Increase the autonomy of development teams • Aids the adoption of new technologies • Speed of change, and speed of responding to new opportunities with new services • Allow a fine-grained approach to performance tuning or scaling 

Microservices drawbacks • Catastrophic failure • Significant Operations Overhead • Substantial Devops Skills Required • Implicit Interfaces • Asynchronicity Is Difficult! 

To learn more about Microservices, Messaging system, DevOps,… 

Resources Microservices • [http://martinfowler.com/articles/microservices.html] InfoQ • eMag: Microservices [http://www.infoq.com/minibooks/emag-microservices] Udi • Dahan on Defining Service Boundaries [http://www.infoq.com/news/2015/02/service-boundaries-healthcare] SOA, EDA, and CEP a • winning combo [http://www.udidahan.com/2008/11/01/soa-eda-and-cep-a-winning-combo/] Event • -Driven Architecture (EDA), SOA, and DDD with Udi Dahan [https://vimeo.com/57644944] Evolutionary • Architecture and Microservices [http://www.infoq.com/presentations/evolutionary-architecture-microservices-cd] MicroServices • , yet another architectural style? [http://www.slideshare.net/aca_it/micro-services-40695502] Exploring • Microservices [https://xebialabs.com/assets/files/whitepapers/Microservices%20Whitepaper.pdf] Practical • Implications of Microservices in 14 Tips [http://www.infoq.com/articles/microservices-practical-tips] The • Role of Containers in Modern Applications [http://www.infoq.com/articles/roundtable-containers-microservices] Why • I love Docker [http://blog.thomasqvarnstrom.com/2015/01/why-i-love-docker.html] Centralised • vs. Decentralised Data [http://bill-poole.blogspot.it/2008/02/centralised-vs-decentralised-data.html] Cover image • [https://www.flickr.com/photos/matthewgriff/3434441969/] 

Thank you @manuelscapolan