Building a Globally Distributed, Scalable Azure Serverless Applications

linkedin.com/in/ryanninodizon
sessionize.com/ryanninodizon
youtube.com/@LearnWithJon
medium.com/@ryanninodizon
github.com/ryanninodizon

Ryan Niño Dizon

Cloud Architect @

Event-Driven Architecture

Message Broker

Event

Event

Producer

Consumer

Consumer

Consumer

Request-Response vs Event-Driven

  • Synchronous vs Asynchronous

  • Inversion of Control

Request-Response Model                                            Event-Driven Model

Sender

Receiver

Request

Publisher

Message Broker

Consumer

Consumer

Consumer

Consumer

Consumer

Request-Response Model                                            Event-Driven Model

Sender

Receiver

Request

Publisher

Message Broker

Consumer

Consumer

Consumer

Consumer

Consumer

Service 1
Api 1

Service 2
Api 2

Service 3
Api 3

Need to know

User

API
Gateway

Subscription Service

Payments Service

Recommendation Service

New Subscriber

Success!

New Subscriber

Payment
Processed

Notification Service

Successful Subscription

What is Serverless?

No infrastructure management
 

Using fully managed services enables developers to avoid administrative tasks and focus on core business logic. With a serverless platform, you simply deploy your code, and it runs with high availability.

Dynamic scalability



With serverless computing, the infrastructure dynamically scales up and down within seconds to match the demands of any workload.

Faster time to market


Serverless applications reduce the operations dependencies on each development cycle, increasing development teams’ agility to deliver more functionality in less time.

More efficient use of resources

 

Shifting to serverless technologies helps organizations reduce Total Cost of Ownership (TCO) and reallocate resources to accelerate the pace of innovation.

Serverless means...

Serverless vs Traditional Computing

Serverless Computing

  • Gets rid of the requirement for developers to handle servers or infrastructure.
  • Automatically assigns resources according to demand.
  • Charges solely for the real usage of resources..

Traditional Computing

  • Developers need to handle and adjust their own servers.
  • Manual action is needed to increase or decrease resources.
  • Costs are incurred for a set infrastructure, regardless of how much it's actually used.

Why Build Global Applications?

  • Minimize latency

  • Meet regional compliance

  • Ensure uptime (multi-region failover)

  • Better customer experience

"If your users are spread across regions, running your app close to them reduces delays, and helps with local data laws"

Key Challenges

  • Data synchronization across regions

  • Maintaining consistency and state

  • Security at scale

  • Monitoring distributed systems

Going global isn’t easy — managing state, syncing data, securing endpoints, and monitoring are all more complex.

Core Azure Serverless

Azure Serverless computing helps developers to build powerful applications by eliminating the need to provision and manage infrastructure.

Azure Container Apps

Azure Functions

Azure Logic Apps

Azure Event Grid

Azure Application Insights

Azure Cosmos DB

Supporting Azure Services

Azure Front Door

Azure Key Vault

Blob Storage

Common Architecture Patterns

  • Event-driven microservices

  • Pub/Sub messaging

  • Durable functions for workflows

  • Stateless API endpoints

"Architecting serverless solutions requires a mindset shift—design for events, decouple components, and make systems resilient to change."

Multi-Region Strategy with Azure

  • Deploy Functions in multiple regions

  • Geo-replication of Cosmos DB

  • Use Azure Front Door for routing

  • Traffic routing by latency or priority

"Azure Front Door makes it easy to route user traffic to the closest region. Cosmos DB handles data across those regions."

Scaling Functions

  • Auto-scale based on events

  • Use queues to buffer burst loads

  • Consumption vs Premium plans

"Serverless scaling isn’t just about compute — buffering with queues prevents overload and supports graceful scaling"

Serverless Design Example

  • Scheduling

Serverless Design Example

  • CQRS

Serverless Design Example

  • Event-based processing

Serverless Design Example

  • File triggers and transformations

Serverless Design Example

  • Web apps and APIs

Serverless Design Example

  • Data pipeline

Serverless Design Example

  • API

Sample Architecture

Demo

Made with Slides.com