Farms, Fabrics and Clouds Steve Loughran Julio Guijarro HP - - PowerPoint PPT Presentation

farms fabrics and clouds
SMART_READER_LITE
LIVE PREVIEW

Farms, Fabrics and Clouds Steve Loughran Julio Guijarro HP - - PowerPoint PPT Presentation

Farms, Fabrics and Clouds Steve Loughran Julio Guijarro HP Laboratories, Bristol, UK December 2007 steve.loughran@hpl.hp.com julio.guijarro@hpl.hp.com Julio Guijarro Researcher at HP Laboratories Area of interest: Deployment In charge of


slide-1
SLIDE 1

Steve Loughran Julio Guijarro HP Laboratories, Bristol, UK December 2007

Farms, Fabrics and Clouds

steve.loughran@hpl.hp.com julio.guijarro@hpl.hp.com

slide-2
SLIDE 2

Page 2 Farms, Fabrics and Clouds www.smartfrog.org

Julio Guijarro

Researcher at HP Laboratories Area of interest: Deployment In charge of OSS release http://smartfrog.org/

slide-3
SLIDE 3

Page 3 Farms, Fabrics and Clouds www.smartfrog.org

Steve Loughran

Researcher at HP Laboratories Area of interest: Deployment Author of Ant in Action

slide-4
SLIDE 4

Page 4 Farms, Fabrics and Clouds www.smartfrog.org

Our research

  • How to host big applications

across distributed resources

– Automatically – Repeatably – Dynamically – Correctly – Securely

  • How to manage them from

installation to removal

  • How to make dynamically

allocated servers useful

slide-5
SLIDE 5

Page 5 Farms, Fabrics and Clouds www.smartfrog.org

Who had breakfast this morning? Question

slide-6
SLIDE 6

Page 6 Farms, Fabrics and Clouds www.smartfrog.org

Who harvested wheat or corn,

  • r killed an animal

for that breakfast? Question

slide-7
SLIDE 7

Page 7 Farms, Fabrics and Clouds www.smartfrog.org

Farms provide food. It is somebody else's problem

slide-8
SLIDE 8

Page 8 Farms, Fabrics and Clouds www.smartfrog.org

Who is wearing clothes they wove or knitted themselves? Question

slide-9
SLIDE 9

Page 9 Farms, Fabrics and Clouds www.smartfrog.org

Provisioning of clothing -fabrics- is outsourced It is somebody else's problem

slide-10
SLIDE 10

Page 10 Farms, Fabrics and Clouds www.smartfrog.org

All new applications are on the Web

  • Web Browser, AJAX clients
  • Richer: Flash, XUL, Silverlight
  • "… as a Service "

⇒Lots of code running in the server ⇒Data mining/analysis problems ⇒Unpredictable demand

slide-11
SLIDE 11

Page 11 Farms, Fabrics and Clouds www.smartfrog.org

Old world installation: single server

Single web server, Single DB RAID filestore

  • SPOF
  • limitations of scale
slide-12
SLIDE 12

Page 12 Farms, Fabrics and Clouds www.smartfrog.org

yesterday: clustering

Multiple web servers, Replicated DB RAID Network filestore Load-balancing router

  • Cost
  • Complexity
  • Limitations of scale
  • Maintains the illusion of a single server
slide-13
SLIDE 13

Page 13 Farms, Fabrics and Clouds www.smartfrog.org

Now: server farms

500 web servers, Distributed filestore Rented storage & CPU Scales up No capital outlay Agile infrastructure

slide-14
SLIDE 14

Page 14 Farms, Fabrics and Clouds www.smartfrog.org

tomorrow? grid fabric. 50000 servers

slide-15
SLIDE 15

Page 15 Farms, Fabrics and Clouds www.smartfrog.org

Application architectures and deployment problems change radically in this world

slide-16
SLIDE 16

Page 16 Farms, Fabrics and Clouds www.smartfrog.org

Application architectures

  • ROA/REST
  • Virtualized
  • Map/Reduce
  • Shards
  • Tuple-spaces
  • Grid
slide-17
SLIDE 17

Page 17 Farms, Fabrics and Clouds www.smartfrog.org

Virtualization

slide-18
SLIDE 18

Page 18 Farms, Fabrics and Clouds www.smartfrog.org

Why?

  • Save on hardware (and power, space)
  • Dynamically move running servers
  • Demand creation of new images
  • Testing complex system configurations
  • Redistributing entire machine image
  • 'virtual appliance'
slide-19
SLIDE 19

Page 19 Farms, Fabrics and Clouds www.smartfrog.org

Assumptions that are now invalid

  • Systems have a long lifespan
  • It is slow/expensive to create a new system
  • It is expensive to duplicate one
  • Systems can/should be managed by hand
  • Clocks proceed at the same rate
  • Physical RAM doesn’t get swapped out
  • Running machines can't be moved/cloned
slide-20
SLIDE 20

Page 20 Farms, Fabrics and Clouds www.smartfrog.org

Server Farms

slide-21
SLIDE 21

Page 21 Farms, Fabrics and Clouds www.smartfrog.org

Assumptions that are now invalid

  • System failure is an unusual event
  • 100% availability can be achieved
  • Data is always near the server
  • You need physical access to the severs
  • Databases are the best storage form
  • You need millions of $/£/€ to play
slide-22
SLIDE 22

Page 22 Farms, Fabrics and Clouds www.smartfrog.org

Who has the servers?

  • Yahoo, Google, MSN, eBay: services
  • MMORPG Game Vendors:

Word of Warcraft, Second Life

  • EU Grid: Scientists
  • HP, IBM, Sun: rent to companies
  • focus on CPU performance
  • Amazon: rent to anyone with an Amazon account
  • focus on startups
slide-23
SLIDE 23

Page 23 Farms, Fabrics and Clouds www.smartfrog.org

Host

Amazon EC2

S3 Storage AMI (Xen VM) AMI (Xen VM) /mnt Host AMI (Xen VM) AMI (Xen VM) Public Internet /mnt /mnt /mnt Fast (free) network

free access; slow initial read time pay per GET; per megabyte

$ $ $ $ $

slide-24
SLIDE 24

Page 24 Farms, Fabrics and Clouds www.smartfrog.org

Amazon EC2

  • Pay as you go Virtual Machine Hosting
  • No persistent storage other than S3 filestore -

uses HTTP GET/PUT/DELETE operations

  • $0.10 per CPU/hour
  • S3 Storage has own billing

(by MB & by access -cheaper in bulk)

slide-25
SLIDE 25

Page 25 Farms, Fabrics and Clouds www.smartfrog.org

Demo

slide-26
SLIDE 26

Page 26 Farms, Fabrics and Clouds www.smartfrog.org

Map/Reduce

Run code near the data, then merge the results

slide-27
SLIDE 27

Page 27 Farms, Fabrics and Clouds www.smartfrog.org

Assumptions that are now invalid

  • Terabyte datasets are hard to work with
  • Code runs on a single machine
  • Sequential code is better than parallel code
  • RAID hardware is the best way to store data
  • Databases are better than filesystems
slide-28
SLIDE 28

Page 28 Farms, Fabrics and Clouds www.smartfrog.org

Shards

slide-29
SLIDE 29

Page 29 Farms, Fabrics and Clouds www.smartfrog.org

Assumptions that are now invalid

  • A single farm needs to scale to infinity
  • You need to provide 100% availability to 100%
  • f users
  • You have to roll out simultaneous updates to

the application, changes to the DB schema, globally

slide-30
SLIDE 30

Page 30 Farms, Fabrics and Clouds www.smartfrog.org

Changes for developers

  • Many classic assumptions are invalid
  • Design for scale
  • Rent servers from the outset

—every developer can have their own set

  • Cover your server costs from the outset and

you are in the black from day 1

slide-31
SLIDE 31

Page 31 Farms, Fabrics and Clouds www.smartfrog.org

Problems for us farmers

  • Power management
  • Predictive disk failure management
  • Load balancing for availability, power
  • Data cache management
  • Billing
  • Security/Isolation
  • How this will change server hardware
  • Managing/Configuring Machine Images
  • Diagnostics when things go wrong
slide-32
SLIDE 32

Page 32 Farms, Fabrics and Clouds www.smartfrog.org

Where is all this heading? Topic for discussion