24/11/14. During this course. Internet is everywhere. Frequency barrier hit. Management costs increase. Advanced Distributed Systems Cloud Computing

Advanced Distributed Systems

Cloud Computing

2014-11-26

Cristian Klein Department of Computing Science

Umeå University

2014-11-26 Cloud Computing 1/43 2014-11-26 Cloud Computing 2/43

During this course

•  Treads in IT

•  Towards a new data center

•  What is Cloud computing?

•  Types of Clouds

•  Making applications Cloud-ready

Cloud Computing

2014-11-26 3/43

Trends in IT

Cloud Computing

2014-11-26 4/43

Internet is everywhere

Source: w3.org

Source: ITU

Total Internet bandwidth is increasing exponentially Home bandwidth follows Nielson's law:

It doubles every 21 months (some say 12) Number of mobile, always-connected users is increasing

More and more users are connected

Cloud Computing

2014-11-26 5/43

Frequency barrier hit

Since 2003, (single-core) processor speed stopped increasing

Cloud Computing

2014-11-26 6/43

Management costs increase

Cloud Computing

2014-11-26 7/43

Power & Cooling expensive

Cloud Computing

2014-11-26 8/43

Increasing prosumarism

•  Users both produce and consume content

•  Need efficient means to share it

Cloud Computing

2014-11-26 9/43

Towards a new data center

Cloud Computing

2014-11-26 10/43

Business face rapid growth

Cloud Computing

2014-11-26 11/43

… regular peaks

•  Daily, weekly and yearly patterns

Wikipedia Traces

Cloud Computing

2014-11-26 12/43

… expected peaks

•  Marketing campaigns –  Christmas

•  Tax filling

Load on the FIFA 98 Website

Cloud Computing

2014-11-26 13/43

… unexpected peaks

•  Triggered by news/events

Source: Google blog

Cloud Computing

2014-11-26 14/43

time

Over-provisioning

time

Under-provisioning lost clients

time

Ideally: auto-scaling wasted

resources

Legend:

actual load provisioned load

Cloud Computing

2014-11-26 15/43

What is Cloud computing?

Cloud Computing

2014-11-26 16/43

NIST definition

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

You do not own computing infrastructure, you rent it

Cloud Computing

2014-11-26 17/43

Essential Characteristics

•  On-demand self-service

–  User can provision resources without human intervention

•  Broad network access

–  Service is accessible through the network

•  Resource pooling

–  Resources are pooled to several costumers

•  Rapid elasticity

–  Can rapidly increase and decrease capacity

•  Measured service

–  Resource usage can be monitored, reported

Cloud Computing

2014-11-26 18/43

Service models

•  SaaS: Software as a Service –  Ready to use applications

–  E.g.: Gmail, Google Docs, YouTube, Facebook

•  PaaS: Platform as a Service –  Ready to use tools

–  E.g.: Google Apps Engine

•  IaaS: Infrastructure as a Service –  Fundamental computing resources –  E.g., Amazon EC2, Amazon S3, Windows

Azure

Cloud Computing

2014-11-26 19/43

Deployment models

•  Private cloud (“my stuff”)

–  Institution resources managed in a

“Cloud-like” fashion

•  Community cloud (“our stuff”)

–  Share resources among institutions

•  Public cloud (“anyone's stuff”)

–  Resources to rent by anybody with a

credit card

•  Hybrid cloud

–  A combination of the above

Cloud Computing

2014-11-26 20/43

Types of Clouds

Cloud Computing

2014-11-26 21/43

Infrastructure as a Service (IaaS)

•  Fundamental computing resources

–  CPU, RAM, storage, network, etc.

•  Mostly in an virtualized way

–  Details of the underlying physical

resources are hidden from the user

Cloud Computing

2014-11-26 22/43

Physical Machine Physical Machine 2

Virtualization

Physical Machine 1

OS 1 OS 2

Application 1 Application 2

Virtual Machine 2 Virtual Machine 1

OS 1 OS 2

Application 1 Application 2

Hypervisor

•  Hypervisors: Xen, VMware, KVM, Microsoft Hyper-V

•  Isolates different users

•  Flexible infrastructure management:

consolidation –  At VM start: mapping –  During VM execution: migration

Cloud Computing

2014-11-26 23/43

Public IaaS Clouds

•  Amazon: EC2, S3

•  Windows Azure

•  RackSpace

•  E.g.: Amazon EC2 –  1 year micro instance for free

• Try it yourself!

–  Small: 0.60 $/hr –  8XL: 4.60 $/hr

•  For comparison:

–  8XL in your own datacenter: 1.50 $/hr (estimated)

–  Abisko @ HPC2N: 0.07 $/core/hr (source: P-O)

Cloud Computing

2014-11-26 24/43

Open-source solutions

•  Useful to build private/community/

hybrid clouds

•  Production-oriented

–  Eucalyptus, CloudStack, OpenStack

•  Research-oriented

–  OpenNebula, Nimbus

Cloud Computing

2014-11-26 25/43

IaaS Cloud standards

•  De-facto: Amazon –  WSDL-based –  createInstance()

–  startInstances(), stopInstances()

•  Open Cloud Compute Interface (OCCI) –  REST-based

POST /compute/ HTTP/1.1 Category: compute [...]

X-OCCI-Attribute: occi.compute.cores=2 X-OCCI-Attribute: occi.compute.hostname="foobar"

HTTP/1.1 201 OK

Location: http://example.com/vms/foo/vm1

Cloud Computing

2014-11-26 26/43

“Cloud-ready” applications

•  Almost any application can run in the Cloud, but they are not “Cloud-ready”

•  Use API to auto-scale –  Application needs to be elastic

time load

2014-11-26 Cloud Computing 27/43 2014-11-26 Cloud Computing 28/43

1 VM per tier

Client

Application

Database

Cloud Computing

2014-11-26 29/43

Scalable architecture

Client

Load balancer

Application 1

Master DB Slave DB 1 Application 2

read write

replicate

Cloud Computing

2014-11-26 30/43

Add auto-scaling

Client

Load balancer

Application 1

Master DB Slave DB 1 Slave DB

n-1 Application

2 Application

n read write

replicate replicate

Cloud Computing

2014-11-26 31/43

IaaS pros and cons

• 

Pros

– General purpose – Very flexible

• 

Cons

– Low-level, fine-grained – Need to handle application

• Deployment

• Elasticity

• Failure

Cloud Computing

2014-11-26 32/43

Platform as a Service (PaaS)

•  Provides programming languages,

libraries, services and tools

• 

Higher level than IaaS

• 

Special purpose

• 

Less flexible

•  Examples

–  For web applications: Google Apps Engine

–  For BigData processing: MapReduce

Cloud Computing

2014-11-26 33/43

Google Apps Engine

•  Develop and host web applications

•  Take advantage of Google

technologies for scalability, resilience

•  Python, Java, Go

•  Several APIs

–  Images, Mail, Log

–  Blobstore, Datastore, BigTable

Cloud Computing

2014-11-26 34/43

helloworld.py import webapp2

class MainPage(webapp2.RequestHandler):

def get(self):

self.response.headers['Content-Type'] = 'text/plain' self.response.write('Hello, webapp2 World!') app = webapp2.WSGIApplication([('/', MainPage)], debug=True)

app.yaml application: helloworld version: 1 runtime: python27 api_version: 1 threadsafe: true handlers:

- url: /.*

script: helloworld.app

Hello Google Apps Engine

Try it, it's free!

Cloud Computing

2014-11-26 35/43

MapReduce

•  A programming model for processing

large data sets (BigData)

–  Difficult due to scalability, failure, etc.

•  Useful for: Indexing, Financial

Analysis, etc.

•  Used by Google, Yahoo!, Facebook

•  Open source implementation:

Hadoop

Cloud Computing

2014-11-26 36/43

MapReduce: Logical View

•  Operates on several key-value spaces:

(1) input, (2) intermediate, (3) output

•  Map : (key

1, value

1) → (key

2, value

2)

•  Reduce : (key₂, list(value₂)) → list(value₃)

Source: blog.maxgarfinkel.com

Cloud Computing

2014-11-26 37/43

MapReduce: Word count

function map(String name, String document):

for each word w in document:

emit (w, 1)

function reduce(String word, list partialCounts):

sum = 0

for each partialCount in partialCounts:

sum += partialCount emit (word, sum)

Spark

•  General-purpose BigData engine

•  Most operations in-memory

•  Allows interactive use

Cloud Computing

2014-11-26 38/43

Spark: More Examples

Cloud Computing

2014-11-26 39/43 2014-11-26 Cloud Computing 40/43

Final thoughts

Cloud Computing

2014-11-26 41/43

Clouds vs. Grids

Clouds Grids

Origin Industry Academic

Economic model Public: pay-per-use Other: domain-specific (e.g., grant models)

Grant models

Size of allocation Many “small” users Few “big” users Waiting time On-demand Queuing Duration Long (months) Short (< 4h)

Usual workload Interactive

Long-lived services Non-interactive Jobs Organization One provider Multiple providers

Cloud Computing

2014-11-26 42/43

Cloud issues

•  Privacy

–  What happens to your data?

•  Security

–  Sometimes VM provide insufficient isolation

•  Standardization is lagging behind

–  Important to prevent vendor lock-in –  IaaS: less of a problem

–  PaaS: major problem

Cloud Computing

2014-11-26 43/43

Conclusion

•  IT trends

•  Frequency barrier

•  Infrastructure complexity

•  Omnipresent network connection

•  What are Clouds: rent computing

resources, rapid elasticity

•  Types of Clouds: IaaS, PaaS, SaaS

•  Making an application Cloud ready

•  Challenges remaining