Google Cloud Computing compares GCE, GAE and GKE

Post on 13-Apr-2017

266 views 2 download

Transcript of Google Cloud Computing compares GCE, GAE and GKE

Google Cloud Computing comparesGCE, GKE and GAE

Simon Su @ GCPUG.TW

I’m Simon Su...

var simon = {/** I am at GCPUG.TW **/};

simon.GDE = 'https://developers.google.com/experts/people/peihsin-su';

simon.nodejs = ‘http://opennodes.arecord.us';

simon.googleshare = 'http://gappsnews.blogspot.tw'

simon.nodejsblog = ‘http://nodejs-in-example.blogspot.tw';

simon.blog = ‘http://peihsinsu.blogspot.com';

simon.slideshare = ‘http://slideshare.net/peihsinsu/';

simon.email = ‘simonsu.mail@gmail.com’;

simon.say(‘Let’s go Cloud!');

Welcome to join us!

Google Cloud Platform

BigQuery

Big Data

Pub/Sub Dataflow Dataproc Datalab

Compute

Compute EngineApp Engine Container Engine

Storage

Cloud Storage

Cloud SQL Cloud Datastore

Bigtable

Machine Learning

Speech API

Machine Learning

Translate APIVision APINatural Language API

GCECompute Engine

- GCE VM, Disk, Network- Load Balancer- Instance Group

GAEApp Engine

- Standard Runtime- Flexible Runtime

- Standard Env- Custom Env

GKEContainer Engine

- Docker & rkt - Pod, deployment- Service vs. NLB- Ingress vs. HLB- Scaling

Getting Start… You want to deploy...

Like this...

Or this...

What we care about...● Loading high

● Run out of resource

● Management

● Monitoring & alert

● Cost

● ...

Over Bare-metal Deployment

Using GAE Deploy

Over Kubernetes

ContainerCode and libraries

ContainerCode and libraries

ContainerCode and libraries

ContainerCode and libraries

GCE / GKE / GAE

PaaSPreset run-times

Java, Go, PHP, Python...Focus is application logic

Pay for what you useLess management overhead

IaaSRaw compute, storage and

networkMore granular control

Pay for what you allocateMore management overhead

Towardsmanaged

infrastructure

Towardsmanaged services

Compute Engine App EngineContainer Engine

- restrictions and conditions

- dockerized service

- micro services- fast scale- no limitation

GCE / GKE / GAE need to do before use

Towardsmanaged

infrastructure

Towardsmanaged services

Compute Engine App EngineContainer Engine

- operation & monitor

- architecture planning

- more integrated pricing

- daily free 28 instance/hour quota

- master node managed by google

- at last 1 worker node

- extend base on template vm size

GCE / GKE / GAE compares in Cost

Towardsmanaged

infrastructure

Towardsmanaged services

Compute Engine App EngineContainer Engine

- custom node size / scalable pricing

- load balancer pricing

- web services with large scaling

- need quick scale

- using GAE services

- micro services- containerized

services- plan to cross

cloud

GCE / GKE / GAE good for...

Towardsmanaged

infrastructure

Towardsmanaged services

Compute Engine App EngineContainer Engine

- small scale for small service deploy

- large scale for high performance service

There are more features...

GCE~ more ~

Functional

● Integrated cloud permissions● Routing and firewalls● Snapshot● Images● Custom machine type

Pricing

● Sustain pricing● Sub-hour billing

GAE~ more ~

Project

App EngineApp Servers

Application instances

Application instances

Application instances

Task queues

Scheduled tasks

Search

Memcache

Logs

● Federation clustering● Auto scaling● GCE resource integrated

○ HLB as K8S Ingress ○ NLB as K8S Service

● Secret● GKE UI● Host auto-scaling● Version upgrade & management● Container registry (gcr.io)● Image builder

GKE~ more ~

StackDriverAll you used in GCP can be Monitored….

MonitoringPlatform, system, and application metricsUptime/health checksDashboards and alerts

LoggingPlatform, system, and application logsLog search/view/filterLog-based metrics

TraceLatency reporting and samplingPer-URL latency and statistics

Error ReportingError notificationsError dashboard

DebuggerDebug applications

ThanksQ&A

More GCP events...

Join us….