Cloud native Microservices using Spring Boot

‹#›© 2014 Pivotal Software, Inc. All rights reserved. ‹#›© 2014 Pivotal Software, Inc. All rights reserved.
Cloud Native Applications
Hands on Workshop
David Ferreira Pinto <dpinto@pivotal.io>
Field Engineering - Pivotal

‹#›© 2014 Pivotal Software, Inc. All rights reserved.
> whoami
David Ferreira Pinto

Agenda
• What’s changing?
• Implications on applications
• help is here!
• Workshop
• Competition
• Deploying and running Cloud Native applications
• Wrap up
• Pizza & Beers

How it used to be
User RegistrationCatalog
Orders ProﬁleSearch
Recommendation ListsAccount
Middleware
OS
Middleware
OS

• Downtime for upgrades
• One small change/fix - whole new release
• Scalability - vertical/horizontal?
• Size for peaks - Crystal ball analysisTM
• Snowflakes

The world has changed

• Distributed computing
• Decoupling of components
• Latency of communication
• Dependencies
• Scale as you grow - Size for “now”
• Cost efficiencies

The rules for the cloud
1. S#!t happens!
• assume the worst!
2. See (1)
A central truth in cloud computing is that failure is inevitable. As
systems scale, we expect nodes to fail ungracefully in random and
unexpected ways, networks to experience sudden partitions, and messages to
be dropped at any time.
Rather than fight this truth, we embrace it.
HEATHER NAKAMA
Software Engineer, Azure Search

What does this mean for applications?
• Scalability
• Inter-dependency
• Platform dependency
• Location dependency
• Resiliency/Availability
• Traceability

© Copyright 2014 Pivotal. All rights reserved. 10
Scalability
Traditional applications have been
modeled as monolithic due an easy
deployment model
Several services are combined into one
massive single application
This model leads to a poor use of
resources when it comes to scaling out
your application
Greedy components steal resources from
others that are collocated with them
Recommendati
on ListsAccount

.

How you should scale
0 1000000 2000000 3000000 4000000 5000000 6000000
Catalog
Search
Recommendation
Lists
Orders
Profile
User
Account
Registration
Requests
Requests

Inter dependency
• Monolithic applications represent
component dependency as strong
bindings at the code level
• It makes harder to get an independent
scalable model promoted by service
oriented architectures

Platform specificity
import javax.servlet.http.HttpServlet;
import com.ibm.servlet.engine.webapp.*;
• Hard dependencies on the
runtime environment make
applications not portable
• At code level with runtime
dependencies
• At OS level (relying on cron as a
scheduler)

Location Specificity : Writing to disk
• Applications often need to write to disk, this includes form uploads with binary
data, or content in most CMS systems
• Containers are short lived and not guaranteed to be executed on the same
hardware every time they are needed to restart. Depending on a local file
system is a big lock dependency some applications impose on the runtime
• Some CMS vendors support usage of a service such as S3 to be the
persistent mechanism of choice

Location Specificity : service locations
Cache.hosts=10.68.27.41,10.68.27.4
2
#naming does not help either
Cache.hosts=cacheserver1,
cacheserver2
Properties file deployed
With application

12 factor Apps - 12factor.net
1. Codebase 
One codebase tracked in revision control, many
deploys
2. Dependencies 
Explicitly declare and isolate dependencies
3. Config 
Store config in the environment
4. Backing Services 
Treat backing services as attached resources
5. Build, release, run 
Strictly separate build and run stages
6. Processes 
Execute the app as one or more stateless
processes
7. Port binding 
Export services via port binding
8. Concurrency 
Scale out via the process model
9. Disposability 
Maximize robustness with fast startup and
graceful shutdown
10.Dev/prod parity 
Keep development, staging, and production as
similar as possible
11.Logs 
Treat logs as event streams
12.Admin processes 
Run admin/management tasks as one-off
processes

© Copyright 2014 Pivotal. All rights reserved.
Monolithic Architecture
18
Relational Database
Data Access
Service
HTML JavaScript MVC
Service
Monolithic ApplicationBrowser

Microservice Architecture
19
…
HTTP
HTTP
HTTP
HTTP
HTTP
HTTP
AMQP
AMQP
Relational
DB
Key/Value
Store
Graph
DB

Microservice Architectures
• Modularity Based on Component Services
• Change Cycles Decoupled / Enable Frequent Deploys
• Efficient Scaling
• Individual Components Less Intimidating to New
Developers
• Enables Scaling of Development
• Eliminates Long-Term Commitment to Technical Stack

Organize Around Business Capabilities
21
Data Access
Service
HTML JavaScript MVC
Service
UISpecialists
Middleware
Specialists
DBAs
BusinessCapability
BusinessCapability
BusinessCapability
Siloed
Functional
Teams
http://martinfowler.com/articles/microservices.html#OrganizedAroundBusinessCapabilities
Siloed
Application
Architectures
Cross-
functional
Teams
Microservice
Architectures

• Manage Configuration Data provided to µServices via the environment
reliably?(Factor-3)
• Update Configuration Data to all µServices instances simultaneously?
• Ensure µServices can discover each other, despite being Disposable?
(Factor-9)
• Handle load-balancing between µServices of the same type? (Factor-8)
• Guard against cascading failures when all instances of the same µService
fail?
• Ensure resiliency by restoring failed µServices?
12 factors is not enough,
Distributed Systems are hard!!!
HOW DO WE:

Help is Here!
Spring Cloud
Quickly add Distributed
Systems Patterns to your
µServices
with
Pivotal Cloud Foundry
Keeps your µServices
running!

Service registry & discovery - Eureka

Service registry & discovery - Eureka
@SpringBootApplication
@EnableEurekaServer //@EnableDiscoveryClient
public class RegistryApplication {
public static void main(String[] args) {
SpringApplication.run(RegistryApplication.class, args);
}
}

Client-side load balancing - Ribbon

Client-side load balancing - Ribbon
@Autowired
@LoadBalanced
private RestTemplate restTemplate;
Quote quote = restTemplate.getForObject("http://quotes/quote/
{symbol}", Quote.class, symbol);

Circuit Breaker - Hystrix

Circuit Breaker - Hystrix
@HystrixCommand(fallbackMethod = "getCompaniesFallback")
public List<CompanyInfo> getCompanies(String name) {
CompanyInfo[] infos = restTemplate.getForObject("http://quotes/company/
{name}", CompanyInfo[].class, name);
return Arrays.asList(infos);
}
private List<CompanyInfo> getCompaniesFallback(String name) {
List<CompanyInfo> infos = new ArrayList<>();
return infos;
}

Configuration Server

Spring Cloud Services for PCF
https://network.pivotal.io/products/p-spring-cloud-services

Workshop

Spring Trader Application

• Services:
• Quotes service
• Accounts service
• Portfolio service
• Web service
Spring Trader Application
http://bit.ly/1RLHzXQ
or
https://github.com/dpinto-pivotal/cf-SpringBootTrader
Quotes
AccountsPortfolio
Web
DB DB
Discovery Service
Circuit Breaker Dashboard

Competition
• Make the most money!!!
• Create account with 100,000
• Buy and sell as much as you want.
• When I say so, sell everything!
• Highest account balance wins!

Setup
• You will need some tools (lab 1).
• If you don’t want to install tools on your laptop, we
have a VM you can use.

Exercises
• Deploy applications to Cloud Foundry
1. Set up the environment
2. Deploy the registry service
3. Create a user-provided service
4. Deploy quote microservice
5. Deploy all services at one
6. Scaling applications
7. Zero downtime deployments
8. Application Performance Monitoring

Lab 1: Set up the environment
• Goal is to ensure you have all the tools and accounts to run
through the rest of the exercises.
• Java 8
• git CLI
— install git cli (or other git tool that allows you to clone repo)
— clone the repository locally.
• CF CLI
— install CF CLI to allow you to interact with PCF
• Build code - ./gradlew build - could take sometime…

Lab 1: Login to Pivotal Cloud Foundry
• Login to Pivotal’s test environment
• cf login -a api.emea.fe.pivotal.io
• user: PodX
• Password: PodX
• browser: http://console.emea.fe.pivotal.io

Lab 2 & 3: Create the Services
• Create a service called:
discovery-service
• bind it to your space
circuit-breaker-dashboard

And you may as well….
traderdb
• Can be any RDBMS ie.
Postgres, ClearDB, etc.

Lab 4 & 5: Pushing micro services
• cf push -f manifest-quotes.yml
• And same for accounts, portfolio and web.

Competition
Buy shares!!!!

Lab 6: Scaling the services
• cf scale quotes -i 2
• yes it is that simple!!!
• What happens when you scale the other services?
• is there one that behaves differently?
• How can we solve this?

Lab 7: Zero downtime deployments
• edit manifest-quote.yml
• cf push -f manifest-
quote.yml
• What happens in Discovery
Service dashboard?
---
timeout: 180
instances: 1
memory: 330M
env:
SPRING_PROFILES_DEFAULT: cloud
JAVA_OPTS: -Djava.security.egd=file:///dev/urandom
applications:
- name: quotes-2
random-route: true
path: springboottrades-quotes/build/libs/quotes-0.2.jar
services: [ discovery-service ]

Lab 8: Application Performance Monitoring
• Go To Marketplace
• Create new relic service and bind it to app
• Re-Push the apps
• In your App Manager UI, accept the license.
• wait a few minutes
• check newrelic dashboard.

And the winner is….

Wrap up

Summary
• It’s a different world!
• decoupling of application and infrastructure
processes
• decoupling of application components
• Platform to automate the running of applications
• Framework to simplify development of cloud
applications

What’s next?
• Topics for future meetups?
• What would you like to hear/discuss?
• Examples:
• Drill down into each cloud service e.g. Circuit Breaker
• API management
• Continuous Delivery
• Logging/Monitoring
• APM
• Any presenters?
• Format?
• Time?

Cloud native Microservices using Spring Boot

Cloud native Microservices using Spring Boot

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Ähnlich wie Cloud native Microservices using Spring Boot

Ähnlich wie Cloud native Microservices using Spring Boot (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Cloud native Microservices using Spring Boot