Cloud computing101

Fabien Hermenier a.k.a Guillaume Urvoy-Keller

…..starting from 2017

1

Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction.

2011

“”2

1 self-provisioning,no human intervention

On-demand self-services

3

4

2availability over the

networkstandard mechanisms

broad network access

5

3multi-tenant

virtual or physical resourceson-demand allocation

location independance

resource pooling

6

reserved instances (yearly based)on-demand instances (hourly based)

hotspot instances (market based)7

Amazon EC2HotSpot instances

bid over the market price to get the instance8

4fast (de-)allocation of resourcesscale to infinity

rapid elasticity

9

verti

cal

elas

ticity

Tiers 1

Tiers 2

Tiers 3

10

horizontal elasticity

Tiers 1

Tiers 2

Tiers 3

11

5 metering capabilitiestransparent reporting

measured service

12

13

and I will call it cloud computing

14

Cloud Computingorigins

15

If computers of the kind I have advocated become the computers of the future, then computing may someday be organized as a public utility just as the telephone system is a public utility... The computer utility could become the basis of a new and important industry.

John McCarthy, 1961

“”

16

cluster computing

loosely coupled co-located serverssingle tenant

non-interactive workload rigid jobs

80s17

Clo

ud o

r no

t ? on demand self-services

broad network access

resource pooling

rapid elasticity

measured service18

Clo

ud o

r no

t ? on demand self-services

broad network access

resource pooling

rapid elasticity

measured service19

ApplicationService

Provider

service oriented

pay as you go

95+1

Client2

Client3

Client

remote access todedicated applications

20

Clo

ud o

r no

t ? on demand self-services

broad network access

resource pooling

rapid elasticity

measured service21

Clo

ud o

r no

t ?

on demand self-services

broad network access

resource pooling(not real hw resources)

rapid elasticity

measured service22

computers on demand.2002Deploy full custom stacks (OS to applications)

23

to (re)deploy reproduciblenetwork experiments

multi-tenant,(limited on purpose) resource pooling

24

to (re)deploy reproduciblenetwork experiments

100Mb/s10ms

10ms50ms,

5% loss

25

Clo

ud o

r no

t ? on demand self-services

broad network access

resource pooling

rapid elasticity

measured service26

Clo

ud o

r no

t ? on demand self-services

broad network access

resource pooling

rapid elasticity

measured service27

S O Aervice

orientedrchitecture

2001+

composable unassociated, loosely coupled units

28

exponential grows since 2001 private and public services to support its

growth

29

800 x

tons of API, mini-services devoted to automation, flexibility, on-demand services for public and private use

30

2006

scalable web services for other websites or client-

side applications31

SOAP & REST over HTTPpay as you goelastic *-oriented services

*data, network or computation32

Clo

ud !

on demand self-services

broad network access

resource pooling

rapid elasticity

measured service33

?aaS34

web access to commercial sw.“one to many” model

customers don’t handle upgradesAPI for integration

Softwareas a ServiceS

aaS

35

36

37

jailed runtime available to host applicationsgeneric or provider-specific APIsno control over the environment

Platformas a ServiceP

aaS

38

39

$ heroku login…

$ git clone https://github.com/heroku/java-getting-started.git$ cd java-getting-started

$ heroku createCreating warm-eyrie-9006... done, stack is cedar-14http://warm-eyrie-9006.herokuapp.com/ | git@heroku.com:warm-eyrie-9006.gitGit remote heroku added

$ git push heroku master… http://warm-eyrie-9006.herokuapp.com/ deployed to Heroku

$ heroku ps:scale web=1

Infrastructureas a Service

low-level resources to deploy arbitrary software stackscomplete control over its network, storage and OS

IaaS

40

41

42

Things will crash. Deal with it!

Assume you could start with super reliable servers (MTBF of 30 years)Build computing system with 10 thousand of thoseWatch one fail per day

Dean Keynote, LADIS 2009

“

”

43

~0.5 overheating (power down most machines in <5 mins, ~1-2 days to recover)~1 PDU failure (~500-1000 machines suddenly disappear, ~6 hours to come back)~1 rack-move (plenty of warning, ~500-1000 machines powered down, ~6 hours)~1 network rewiring (rolling ~5% of machines down over 2-day span)~20 rack failures (40-80 machines instantly disappear, 1-6 hours to get back)~5 racks go wonky (40-80 machines see 50% packetloss)~8 network maintenances (4 might cause ~30-minute random connectivity losses)~12 router reloads (takes out DNS and external vips for a couple minutes)~3 router failures (have to immediately pull traffic for an hour)~dozens of minor 30-second blips for dns~1000 individual machine failures~thousands of hard drive failuresslow disks, bad memory, misconfigured machines, flaky machines, etc.Long distance links: wild dogs, sharks, dead horses, drunken hunters, etc.

Typical first year for a new google cluster

44

Building fault tolerant

services

deal with failuresdeal with inconsistency

be pessimistic

at every level

45

October, 21th 2016: dynDNS targeted by a DDoS

Affected services:

1.2 Tb/s of DNSlookups

46

applicationsruntimes

integration/security

databaseservers

virtualisationserver HW

storagenetwork

you

man

age

ol’school IT

47

SaaSapplications

runtimesintegration/security

databaseservers

virtualisationserver HW

storagenetwork

man

aged

by

vend

or

48

PaaSapplications

runtimesintegration/security

databaseservers

virtualisationserver HW

storagenetwork

you manage

man

aged

by

vend

or

49

Open-source PaaS stacks

IaaSapplications

runtimesintegration/security

databaseservers

virtualisationserver HW

storagenetwork

you

man

age

man

aged

by

vend

or

51

Open-source IaaS stacks

(2008+)

cloudstack

(2008+)

(2010+) (2012+)

52

53

vendor lock-in

IaaS PaaS SaaS

Deployment models

54

public cloudgeneral availability to

everyone

the “real” cloudreduced coststrust issues ?

55

56

Trust in megive me your code & data

private cloudself hosted cloudworldcompany SA

might reduce TCOstronger trust

better manageability

57

worldcompany SA hybrid cloud

58

multi-cloudsyou spread your application

avoid Single Point of Failures*take the benefits of each cloud

LB

59