PRESENTED TO : DR. NADIR SHAH

1

OutlineWhat is it?Why now?Cloud killer appsEconomics for usersEconomics for providersChallenges and opportunitiesImplications

2

What is Cloud Computing?Old idea: Software as a Service (SaaS)

Def: delivering applications over the InternetRecently: “[Hardware, Infrastrucuture,

Platform] as a service”Poorly defined so we avoid all “X as a service”

Utility Computing: pay-as-you-go computingIllusion of infinite resourcesNo up-front costFine-grained billing (e.g. hourly)

3

Why Now?Experience with very large datacenters

Unprecedented economies of scaleOther factors

Pervasive broadband InternetFast x86 virtualizationPay-as-you-go billing modelStandard software stack

4

Spectrum of CloudsInstruction Set VM (Amazon EC2, 3Tera)Bytecode VM (Microsoft Azure)Framework VM

Google AppEngine, Force.com

5

EC2 Azure AppEngine Force.com

Lower-level,Less management

Higher-level,More management

Cloud Killer AppsMobile and web applicationsExtensions of desktop software

Matlab, MathematicaBatch processing / MapReduce

Oracle at Harvard, Hadoop at NY Times

6

Unused resources

Economics of Cloud Users

7

• Pay by use instead of provisioning for peak

Static data center Data center in the cloud

Demand

Capacity

Time

Demand

Capacity

Time

Unused resources

Economics of Cloud Users

8

• Risk of over-provisioning: underutilization

Static data center

Demand

Capacity

Time

Economics of Cloud Users

9

• Heavy penalty for under-provisioning

Lost revenue

Lost users

Demand

Capacity

Time (days)1 2 3

Demand

Capacity

Time (days)1 2 3

Demand

Capacity

Time (days)1 2 3

Economics of Cloud Providers5-7x economies of scale [Hamilton 2008]

Extra benefitsAmazon: utilize off-peak capacityMicrosoft: sell .NET toolsGoogle: reuse existing infrastructure

10

ResourceCost in

Medium DCCost in

Very Large DCRatio

Network $95 / Mbps / month $13 / Mbps / month 7.1x

Storage $2.20 / GB / month $0.40 / GB / month 5.7x

Administration ≈140 servers/admin >1000 servers/admin 7.1x

Adoption Challenges

Challenge Opportunity

Availability Multiple providers & DCs

Data lock-in Standardization

Data Confidentiality and Auditability

Encryption, VLANs, Firewalls; Geographical Data Storage

11

Growth ChallengesChallenge Opportunity

Data transfer bottlenecks

FedEx-ing disks, Data Backup/Archival

Performance unpredictability

Improved VM support, flash memory, scheduling VMs

Scalable storage Invent scalable store

Bugs in large distributed systems

Invent Debugger that relies on Distributed VMs

Scaling quickly Invent Auto-Scaler that relies on ML; Snapshots

12

Policy and Business Challenges

Challenge Opportunity

Reputation Fate Sharing Offer reputation-guarding services like those for email

Software Licensing Pay-for-use licenses; Bulk use sales

13

Short Term ImplicationsStartups and prototypingOne-off tasks

Washington post, NY TimesCost associativity for scientific applicationsResearch at scale

14

Long Term ImplicationsApplication software:

Cloud & client parts, disconnection toleranceInfrastructure software:

Resource accounting, VM awarenessHardware systems:

Containers, energy proportionality

15