Machine Learning
Pipeline

The basic process to learn from data

Big Data
Phenomenon

How data handling can be a challenge

Data Engineering
Evolution

From batch processing to fast data architectures

Batch Processing

Lambda (λ) Architecture

Kappa (κ) Architecture

Systems Engineering
Evolution

From big monoliths to small communicating services

Monoliths

Microservices

Event Sourcing + CQRS

Event Sourcing + CQRS

Reactive Manifesto

The system responds in a
timely manner if at all possible
(establish reliable upper bounds to deliver
a consistent quality of service)

Responsive

The system stays responsive
in the face of failure
(resilience is achieved by replication, containment, isolation and delegation)

Resilient

 The system stays responsive under varying workload.
(react to changes in the input rate by
increasing or decreasing resources)

Elastic

The system relies on async message-passing
(that ensures loose coupling, isolation and location transparency)

Observability

Systems Operations
Evolution

From on-premise to service meshes in the cloud

The Cloud

Virtualization

Orchestration

Service Meshes

Embrace failures instead
of trying to prevent them
(take advantage of the dynamic nature of running on a cloud platform)

Resiliency

Allow for fast deployments
and quick iterations
(the same idea behind the agile software development movement)

Agility

Add control of application life cycles from inside of it
(instead of relying on external processes
and monitors)

Operability

Provide information to know about the application state
(add ways of querying the current state
of a given application)

Observability

Intelligent Systems Architecture

Joining the best architectural practices

Cloud-Native Infrastructure

+

Kappa Architecture
+

Reactive Microservices

+

Reactive Machine Learning