microservice architecture

Make Developing software is easier, cheaper, faster and smarter.

Agenda

Scope - Why and How , Biz-Tech Drivers

Domain Concerns - Technology Options

Best Practices - Design/Code/Test/Deployment

API Governance - Elements and implementation. Spring Boot Application, Demo

Dev-Ops - Culture, Practices

Microservices

Is an architectural style that structures an application as a collection of loosely coupled services, which implement business capabilities that can be deployed and scaled independently

Something strange is happening ...

The rise of the application cloud
End of the road for monolithic software architectures
De-productisation through microservices
Infrastructure shifts to ‘serverless’ event-driven programming models
Rules of software distribution being rewritten

TRENDS

Scope

Defining process of publishing, promoting, and overseeing APIs in a secure, scalable environment - Policy driven approach
API Key life-cycle management, tenant quota management and metering of APIs, Managing, securing, and mediating API traffic
Ensuring that developers are productive.
Define domain concerns and technology stack selection for managing APIs
Define best practices for API design/development and ensure they are being followed. - Quality Gatekeeper of API
Ensure to limit the use of technologies that are tied to particular vendor as much as possible with preference open source.

Goals

Stateless, independently deployable and scalable
Rapid development, provisioning, and deployment of artifacts with independent life cycle
RESTful to be easily consumed by modern client JS frameworks.
Services APIs based on OpenAPI standards
API governance where in various service concerns are managed.
APIs need to be based on modern tech stack and should be quick to adapt to changes and services easily amenable.
Cloud friendly APIs (should be able to leverage the cloud standards), Easy to implement cloud design patterns. Services are SaaS friendly with the ability to turn into on-premise easily.
Efficient & scalable storage/data transfer with high volume/velocity and the variety of data.

Economic

API-first development strategy is the state of art, provides unprecedented up-sell opportunity (disruptive, as ~30 million developers can write new way applications with multiplier effect)
Developer first, then build products has proved to be good strategy for various consumption methods (web, mobile, 3rd party apps and in house products)
Automated faster releases will improve agility enormously,
Bounded contexts to limit cognitive load. Increases the adoption rate

Technology

Independently deploy-able/testable services with independent release cycles
Makes it easy to scale with inherent stateless architecture
Works well with large teams, Easy to manage the releases
Trend for SaaS solutions is to develop API as for all internal and external usage as well with API as scaling with that remains the only credible option
Quicker technology adoption and increases happiness quotient for developers

Drivers

Economic

Enables diverse technologies to co-exist,new busines oppurtunity, speed to market, data provided by API can be mixed and remixed as per business needs to provide various visualizations, adapt to new trends faster.
Eat your own API - Same APIs for internal as well as external. reduces the cost with early feedback cycles

Technology

It builds on loosely coupled but highly cohesive set of services.
Eliminates long-term commitment to a technology stack.

Drivers

"API becomes digital glue" - Forrester.

Technologies

Selecting Best persistent model - Options

Data really powers everything that we do. —Jeff Weiner, CEO of LinkedIn

Polyglot persistence
Autonomous Tech Teams
Voluntary adoption,
Lot’s of small apps

ElasticSearch
Low latency for indexed read (return set of documents)
Low latency for random search (return set of documents)
Low latency for full text and aggregates queries (return facets)
MySQL/Mongo
Low latency indexed read
High latency random search

Dynamic schema - Mongo
Controlled schema and transactions - MySQL

Redshift/Vertica
High write throughput
Low latency aggregation queries

S3
High read throughput (MB/sec)
Moderate latency single document Read

Everything in few words

Designed with the same next generation technologies powering Netflix, Amazon, LinkedIn viz...

MODERN, CLOUD-SCALE TECHNOLOGY

Lightening fast performance regardless of number of users or data volume.

FAST INGESTION, SEARCH, EXPORT

Dramatically improve throughput and response time.

NO SINGLE POINT OF FAILURE

High availability with truly distributed processing.

MT AWARE MANAGED SERVICES

It is estimated TCO of these services reduces by 40-60%

Its all about "scaling" at all levels