Building effective Java applications for the Cloud: The DHARMA principles - Daniel Bryant

Building Java applica.ons for the Cloud: The DHARMA principles

Daniel Bryant Principal Consultant, Open Credo

[email protected]

@danielbryantuk

What to expect •  Problems when developing for the Cloud –  “LiG and shiG” –  Smashing the monolith –  Greenfield (“Cloud na.ve”)

•  Some sugges.ons on where to focus efforts –  Not specific plaPorm/framework reviews

•  Tools and techniques

•  Lots of informa.on… (slides will be available)

14/10/2014 @danielbryantuk

Who Am I?

•  LJC Associate

•  Adopt OpenJDK and JSR


•  Principal Consultant at OpenCredo ü  Agile transforma.ons

ü  DevOps ü  Microservices and Cloud

The Current Industry Wish List…

•  Service-‐Oriented Architecture

•  Cloud-‐based deployments

•  DevOps Culture

@danielbryantuk 14/10/2014

The Current Industry Wish List…

•  Service-‐Oriented Architecture – Twi`er’s Story (bit.ly/1j1WbmI)

•  Cloud-‐based deployments – Today!

•  DevOps Culture – Devoxx UK talk (bit.ly/1BylnZb) – Previous LJC Event (bit.ly/1elVPJz)

@danielbryantuk 14/10/2014

Common Cloud Problems

TL;DR…



Not respec<ng the underlying environment


Lack of applica<on/plaDorm monitoring…


Bizarre failure modes…


Difficulty in understanding the new architecture


Confusion over environment provisioning and config


Not tes<ng in the Cloud… (hint: here be dragons!)

We’ve created the “Cloud DHARMA Principles” to act as a checklist when building Cloud apps


dharma /ˈdɑːmə,ˈdəːmə/ noun 1.  Signifies behaviors that are considered to be in

accord with order that makes life and universe possible (Hinduism)

2.  "cosmic law and order”, but is also applied to the teachings of the Buddha (Buddhism)


Documented (just enough)

Highly cohesive/loosely coupled (all the way down)

Automated from commit to Cloud

Resource aware Monitored thoroughly

An.fragile


Documenta.on (just enough)

•  Provide a map for developers

•  Component purpose (and contract)

•  Ini.alisa.on instruc.ons (mocks/stubs)

•  Highlights areas of opera.onal risk


Simon Brown’s C4 Model


www.codingthearchitecture.com

14/10/2014 @danielbryantuk 14/10/2014 @danielbryantuk

API Docs with Swagger


helloreverb.com/developers/swagger

Create a PACT


github.com/DiUS/pact-‐jvm





An.fragile


High Cohesion / Loose Coupling (all the way down…)

•  Code

•  Architecture –  Components –  Services

•  Public API –  PayPal (bit.ly/1hnZNly)


Smashing the Monolith…

•  Business func.onality -‐“Cart Service”

•  Technology chunk -‐ “Email Service”

•  Ver.cal Slice –  “Registra.on” (Groupon: vimeo.com/105880150)

•  Horizontal Slice –  “User Repo” (Microservices: oreil.ly/1pp6qmx)


Microservices

•  Probably won’t catch on… ;-‐)


Twelve Factor Apps


h`p://12factor.net/





An.fragile


Automated from Commit to Cloud


•  Con.nuous Integra.on

•  Con.nuous Deployment

•  Con.nuous Delivery

Our Build Pipeline

Jenkins, with plugins…

•  Build Pipeline –  wiki.jenkins-‐ci.org/display/JENKINS/Build+Pipeline+Plugin

•  Parameterized build –  wiki.jenkins-‐ci.org/display/JENKINS/Parameterized+Build

•  Promoted Builds Plugin –  wiki.jenkins-‐ci.org/display/JENKINS/Promoted+Builds+Plugin


Our Build Pipeline

•  Component Build – Compile – Unit Tests (surefire) –  Integra.on Tests (failsafe)

•  Deployment onto QA Cloud – Python Scripts + Chef to provision – Verify success using Python – Serverspec serverspec.org


Our Build Pipeline

•  Acceptance Tests – Cucumber (and Selenium)

•  Performance Tests –  Jmeter + Jenkins performance plugin – Make sure environment is realis.c!!

•  Live Deployment?


Automa.ng QA

•  Intra-‐component integra.on tes.ng – U.lise embedded datastore/middleware – “Scassandra” (github.com/scassandra) – Service virtualisa.on (www.mbtest.org)

•  Inter-‐component integra.on tes.ng – The hardest part of SOA… – Consider ‘synthe.c txns’ (ac.ve monitoring)


Infrastructure: Say No To Snowflakes!

•  Automate all provisioning (store in SCM)

•  Fry... –  Chef, Puppet, SaltStack, Ansible –  Bash, Python (Fabric) –  Vendor APIs

•  …or bake? –  Packer.io – NePlix Aminator


Infrastructure: Say No To Snowflakes!

•  Doing “Proper Development” – Gareth Rushgrove at CraG Conf (bit.ly/1njuc49) – Chef Conf (www.youtube.com/user/getchef)

•  Local tooling/tes.ng – Vagrant (www.vagrantup.com) – Docker (www.docker.io)


Configuring the Cloud

•  Bundle config with app – Upload new config file – Re-‐deploy en.re app (easier with Docker?)

•  Store externally – Zookeeper & Curator curator.apache.org – NePlix Archaius github.com/NePlix/archaius – etcd github.com/coreos/etcd – Consul.io www.consul.io






An.fragile


Deployment PlaPorm: What you’ve got…


What you think you want…


What you actually get…


Fact: 9 out of 10 cheetahs prefer the taste of an Ops team over <nned food

Thou Shalt Know thy Cloud…

•  AWS “Magne.c” EBS 100 IOPS – New SSD EBS 3K IOPS (burst, PIOPS available) – My Mac SSD does 49K IOPS

•  1000Mbps network max transfer ~125MB/s –  My Mac does 400+ MB/s Sequen.al Write to SSD

Reference for Mac sta.s.cs: bit.ly/1GJZH8


Thou Shalt Know they Cloud…

“Everything fails all the <me [in the cloud]” Werner Vogels, CTO, Amazon.com

•  Everything is ephemeral

•  Vola.lity •  Noisy (virtual) neighbours –  bit.ly/1w1HQy7


Thinking/Ac.ng Opera.onally

•  Cul.vate “Mechanical Sympathy”

•  Virtualisa.on – Tech Target (bit.ly/1kDVqyG)

•  Networking –  ‘Unix and Linux System Administra.on Handbook’ – aws.amazon.com/documenta.on


Thinking/Ac.ng Opera.onally

•  Learn Linux fundamentals

•  Diagnos.c skills –  top, netstat, vmstat, tcpdump –  Java u.ls: jps, jstat, jmap, jhat – “DevOps Troubleshoo.ng” by K. Rankin






An.fragile


Monitor All The Things!

•  Infrastructure monitoring – Nagios / Zabbix – AppDynamics

•  Distributed Tracing –  twi`er.github.io/zipkin

•  Centralised Logging –  logstash.net


Component Metrics


•  Dropwizard’s Metrics – metrics.codahale.com – Spring Boot (bit.ly/1rGo76V)

•  NePlix’s Servo – github.com/NePlix/servo

•  Etsy’s StatsD – github.com/etsy/statsd/wiki

Health Checks


Gauges, Counters, Meters, Timers…


Graph It!


dashing.io


Phrase borrowed from Etsy!





An.fragile


An.fragile

•  The opposite of fragile? – Robust… – An.fragile…

•  NePlix are best-‐in-‐class – bit.ly/1gs5n3q

•  System must be robust first!


Design for Failure

•  Distributed Compu.ng Principles –  ‘For young bloods’ (bit.ly/1pKVepz) – NePlix (bit.ly/1h5GMid)

•  Design pa`erns – Timeouts / retries – Bulkheads / circuit-‐breakers


Retries


github.com/rholder/guava-‐retrying

Circuit-‐breaker


github.com/NePlix/Hystrix github.com/NePlix/Hystrix/tree/master/hystrix-‐contrib/hystrix-‐javanica

java.dzone.com/ar.cles/hystrix-‐and-‐spring-‐boot projects.spring.io/spring-‐cloud/

An.fragile Pa`erns: Async FTW Asynchronous Communica.on -‐ queues, pub/sub Command Query Responsibility Segrega.on (CQRS) Event Sourcing (ES)

mar.nfowler.com/bliki/CQRS.html


An.fragile Pa`erns: Respect the CAP

Eventual consistency (ACID vs BASE)

Clever caching (soG-‐state)


en.wikipedia.org/wiki/CAP_theorem cloudshankar.blogspot.co.uk/2013/05/eventual-‐consistency.html www.dataversity.net/acid-‐vs-‐base-‐the-‐shiGing-‐ph-‐of-‐database-‐transac.on-‐processing/

An.fragile Pa`erns: Elas.c Scaling

Stateless components

Distributed data stores / caches 14/10/2014 @danielbryantuk





An.fragile 14/10/2014 @danielbryantuk

So, Cloud Apps are ‘done’ when…

Thanks For Listening

•  Massive thanks to all the OpenCredo team!

•  Ques.ons / comments? – [email protected] – @danielbryantuk


Building effective Java applications for the Cloud: The DHARMA principles - Daniel Bryant

Presentations & Public Speaking

Transcript of Building effective Java applications for the Cloud: The DHARMA principles - Daniel Bryant