Scribd
Upload
22 views

PDF # L4 Virtualization and Cloud Computing

Uploaded by

Rahul Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
22 views

PDF # L4 Virtualization and Cloud Computing

Uploaded by

Rahul Singh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 24

Virtualization and Cloud

Computing
Sorav Bansal
Administrivia
• Instructors: Sorav Bansal, Huzur Saran, Gautam
Shroff (Tata Consultancy Services)
• Webpage:
http://www.cse.iitd.ernet.in/~sbansal/csl862
• Syllabus: Lectures + Papers on webpage +
Gautam Shroff’s book (see webpage)
• Lecture Hours:
• Assignments: Will be posted on Webpage
• Teaching Assistant:
What to Expect
• Deep Understanding of Virtualization Internals
• Heavy Programming
– Pre-requisite: Must have done significant
programming in OS course
– Many new concepts and a new environment
• Understanding of Cloud Computing and Related
Technologies
• Systems Research Papers
– Expect you to read the paper before attending lecture
– Often, the lecture will be organized as a discussion
based on your understanding of the paper
History of Computing

Babbage Difference Engine


1879
Harvard Mark-I, 1944 MIT Whirlwind, 1951

IBM 360 Mainframe DEC PDP-8 minicomputer ATARI microcomputers


1964, $2.5-3 million 1965, $18,000 Gaming + home computing
1979.
History of Computing - 2

Apple Macintosh, 1984 IBM PC/AT, 1984 Microsoft Windows, 1985


Intel 80286 Microprocessor

Linus Torvalds, 1991


Intel Pentium Processor Internet and Search, 1998
Diagram, 1993
History of Computing - 3

IPAD, 2010 SaaS Computing

The Million Server Datacenter


Providing IaaS
Source acknowledgement: Raghavan Subramaniam, Infosys
Source acknowledgement: Raghavan Subramaniam, Infosys
More Buy-vs-Rent examples
• Each IIT student must buy a house in Delhi
versus Rent a hostel room
• One workstation per student versus GCL lab
• Networks: Circuit-switched versus Packet-
switched

Common Theme: Virtualization of Resource +


Scheduling
Virtualization inside an OS
• CPU  Process
• Memory  Virtual Memory
• Disk  Files
• Network Card  TCP/UDP Sockets
• Screen  Windows

NOW: Physical Machine (CPU+Mem+Disk+Net+…)


 Virtual Machine (VCPU+Vmem+Vdisk+Vnet+…)
How to Virtualize
• Divide a resource in Time and/or Space
• Share
• Protect
• Schedule
• Make Pre-emptible
• Build a Cost Model

Consider examples: CPU, Memory, Disk, Cloud


Why Virtualize?
• Usually cost
– $$, power, space, maintenance personnel, ….

• Also, Hardware Trends

• Plus, much more


– Mobility, Security, Encapsulation, …
Hardware Trends: Moore’s Law

Number of Transistors double every 2 years


Not expected to stop until 2015 or later!
Hardware Trends: CPU Frequency
• Altair 8800 (1975), used an Intel 8080 CPU with a
clock rate of 2 MHz
• IBM PC (1981) : 4.77 MHz
• Pentium (1995): 100 MHz
• Pentium4 (2002): 3 GHz

CPU Speed scaled by 1500x over 30 years


Wirth’s Law (1995):
``Software is getting slower more rapidly than
hardware becomes faster’’
Hardware Trends: Memory/Disk
• Memory and Disk Size have also followed
exponential trends:

Disk Size
Hardware Trends: Power
2020 IT Carbon Footprint

820m tons CO2

360m tons CO2

2007 Worldwide IT
carbon footprint:
2% = 830 m tons CO2
Comparable to the
global aviation
industry

Expected to grow 260m tons CO2


to 4% by 2020

Source: David Patterson


17
Virtualization Economics 101

Capacity $
Machines

Demand Capacity

Demand
Time Time
“Statically provisioned” “Virtual” data center
data center in the cloud

Unused resources
Risk of Under-utilization
• Underutilizations occur if “peak” predictions
too optimistic
Capacity
Unused resources
Resources

Demand

Time

Static data center


Risk of Under-Provisioning

Resources
Capacity

Demand
Resources

Capacity 1 2 3
Time (days)
Demand Lost revenue
1 2 3

Resources
Time (days)
Capacity

Demand
1 2 3
Time (days)
Lost users
20
“Risk Transfer” to Cloud
• Not (just) Capital Expense vs. Operation Expense!
• “Cost associativity”: 1,000 CPUs for 1 hour same
price as 1 CPUs for 1,000 hours (@$0.08/hour)
• Major enabler for SaaS startups
– Animoto traffic doubled every 12 hours for 3 days when
released as Facebook plug-in
– Scaled from 50 to >3500 servers
– ...then scaled back down
• Gets IT gatekeepers out of the way
– not unlike the PC revolution
Classifying Clouds
• Instruction Set VM (Amazon EC2)
• Managed runtime VM (Microsoft Azure)
• Framework VM (Google AppEngine)
• Tradeoff: flexibility/portability vs. “built in”
functionality
Lower-level, Higher-level,
Less managed More managed

EC2 Azure AppEngine


Course Outline
• Virtualization
– Dynamic recompilation, JVM, LLVM, …
– Vmware, Xen, HVM, …
– CPU virtualization, memory virtualization, I/O device
virtualization
– Applications
• Cloud Computing
– Parallel Programming
– Data in the Cloud (MapReduce, BigTable, …)
– Present Day Clouds
– Private/Public Clouds
Next Lecture
• Compiling Java Just In Time

• Xen and the Art of Virtualization

You might also like