This document discusses the history and evolution of cloud computing, including its origins in concepts like virtualization, supercomputing, cluster computing, grid computing, and distributed computing. It traces the development of these technologies over time, from early concepts in the 1960s proposed by John McCarthy regarding computation as a public utility, to modern implementations like Google's large network of clustered commodity servers. Key people and technologies that advanced each concept are described, such as Seymour Cray's work in supercomputing and Ian Foster and Carl Kesselman's contributions to grid computing through projects like Globus Toolkit.

1. Virtualization, Cloud
Computing,
and Business Continuity
Michael S. Sunggiardi
michael@sunggiardi.com

2. Why Cloud?
• When drawing Computer Network Diagrams, we used to
represent internet as a Cloud
• Connect a computer to Cloud by a line, meaning that the
computer is connected to internet, and accesses internet
services or content
 Web mail
 Instant Messaging
 Web Pages
 Online Virus Scanning
CLOUD

3. Definition
• The underlying concept of Cloud Computing dates back
to 1960 and is attributed to computer scientist John
McCarthy
• John McCarthy proposed the concept of "Computation
being organized and delivered as a public utility"
• The characteristics of the concenpt are similar to the
"Service bureaus" in 1960s (a service bureaus is a
company which provides business services for a fee)

4. John McCarthy
(September 4, 1927 – October 24,
2011) was an American computer
scientist and cognitive scientist. He
invented the term "artificial
intelligence" (AI), developed the Lisp
programming language family,
significantly influenced the design of
the ALGOL programming language,
popularized timesharing, and was very
influential in the early development of
AI.

5. Cloud Progress
• Super Computing
• Cluster Computing
• Grid Computing

6. Super Computing
• Supercomputers were introduced in the 1960s and
primarily designed by Seymor Cray, called "the father of
supercomputing"
• Supercomputers is a massive computer with hundreds or
thousands of CPUs sharing common memory and I/O
• Cray XT5 Jaguar is currently the fastest supercomputers
in the world achieving 1,75 petaflopfs (quadrillion
calculations per second) on November 16, 2009

7. Seymour Roger Cray
(September 28, 1925 – October 5,
1996)
By 1960, at the age of 34, Seymour had
established his reputation for genius in
designing high performance computers.
He had completed the design of the
Control Data 1604, the first computer
to be fully transistorized and had begun
the design of the first system that
earned the title of supercomputer, the
CDC 6600 which was also the first major
system to employ three-dimensional
packaging and an instruction set that
was later to be referred to as RISC.

8. FLOPS
Name FLOPS
yottaFLOPS 1024
zettaFLOPS 1021
exaFLOPS 1018
petaFLOPS 1015
teraFLOPS 1012
gigaFLOPS 109
megaFLOPS 106
kiloFLOPS 103
In computing, FLOPS (or flops or flop/s, for
floating-point operations per second) is a
measure of a computer's performance,
especially in fields of scientific calculations
that make heavy use of floating-point
calculations, similar to the older, simpler,
instructions per second.

9. Quad to Six Core
The XT5 Jaguar, located
at Oak Ridge National Lab
in Tennessee, recently
received a refresh that
included upgrading its
quad-core CPUs to hex-
core Opteron processors.
That means a 2.3
petaflop per second
theoretical performance
peak (”nearly a quarter of
a million cores”), and
1.75 petaflops measured
by the Linpack
benchmark.

10. Cluster Computing
• Another approach to build a supercomputer, as adopted
by Google
• A cluster (group) of hundreds of thousands of COTS* low
cost computers interconnected via fast LAN, configured
so that they appear as a single machine.
• Achieve high availability and scalability.
• Typically cluster computing is much more cost effective
than a single supercomputer.
*COTS: Commercial off-the-shelf

11. Cluster Computing - Programming Tools
• Message Passing Interface (MPI)
- API for sending/receiving messages
- Designed for high performance computing on both
massively parallel machines and on workstation
clusters
- Group communication, scalable file I/O, dynamic
process management, synchronization

12. Message Passing Interface
Message Passing Interface
(MPI) is a standardized and
portable message-passing
system designed by a group of
researchers from academia and
industry to function on a wide
variety of parallel computers.
The standard defines the syntax
and semantics of a core of
library routines useful to a wide
range of users writing portable
message-passing programs in
Fortran 77 or the C
programming language.

13. Cluster Computing - Programming Tools
• Parallel Virtual Machine (PVM)
- Emulate a general purpose heterogeneous computing
framework on interconnected computers
- Present a view of virtual processing elements
- Create and manage processing tasks, basic message
passing

14. Parallel Virtual Machine
The Parallel Virtual Machine
(PVM) is a software tool for
parallel networking of
computers. It is designed to allow
a network of heterogeneous Unix
and/or Windows machines to be
used as a single distributed
parallel processor.

15. Cluster Computing - Google Example
• Estimated 450.000 low cost commodity server in 2006
(900.000 servers in 2011)
• A Google's rack content twenty 2U or forty 1U servers on
each side
• The servers on each side of a rack are interconnected via
a 100Mbps ethernet switch
• Each rack ethernet switch has one or two Gigabit uplinks
to a core Gigabit switch that connect all racks together

16. Google Data Center
The Data Center

17. Grid Computing
• Early 1990s Ian Foster and Carl Kesselman came up with a
new concept of "The Grid: Blueprint for a new
computing infrastructure“
• Expand the techniques of cluster computing where
multiple independent computer clusters act like a grid
due to their nature of not being located in a single
administrative domain, as distributed and large scale
cluster computing, as well as a form of network
distributed parallel processing

18. Ian Foster and Carl Kesselman
Dr. Carl Kesselman is Professor of Industrial
and Systems Engineering and a Fellow in the
Information Sciences Institute in the Viterbi
School of Engineering at the University of
Southern California. Dr. Kesselman also serves
as Chief Scientist of Univa Corporation, a
company he founded with Globus co-founders
Ian Foster and Steve Tuecke.
Dr. Foster is a fellow of the American
Association for the Advancement of Science,
the Association for Computing Machinery, and
the British Computer Society. He was a co-
founder of Univa UD, Inc., a company
established to deliver grid and cloud
computing solutions.

19. Grid Computing
• What distinguished grid computing from conventional
cluster computing systems is that grids tend to be more
loosely coupled, heterogeneous and geographically
dispatched
• Applied to computationally intensive scientific,
mathematical and academic problems a volunteer
computing
• Used in commercial entrprises for diverse applications and
commercial transactions in support of e-commerce and web
service as utility computing

20. Grid Computing
• SETI@home
- A grid computing project using internet connected
computers for scientific experiments in the Search for
Extraterrestrial Intellegence (SETI), hosted by Space
Sciences Laboratory at the University of California
Berkeley

21. Grid Computing
• BOINC
- Non commercial middleware system for volunteer and
grid computing
- For researchers to tap into the enormous processing
power of personal computers around the world
- About 550.000 active hosts worldwode processing on
average 1,5 petaflops as of March 2009
- Supported platforms: Linux (32-bit), Linux (64-bit),
Mac OS X (10.4.0+), Windows 2000/XP/Vista/7 (32-bit)
- Total credit: 137,881,433,424
Total user count: 1,280,634

22. Grid Computing
• Globus Toolkit
- A fundamental enabling technology for the "Grid"
letting people share computing power, databases
and other tools securely online across corporate,
institutional and geographic boundaries without
sacrificing local autonomy
* Open Grid Services Architecture
* GRIP (Grid Resource Information Protocol)
* GRAM (Grid Resource Access and Management)
* GridFTP

23. Globus Toolkit
Scientists in the National Fusion
Collaboratory are learning to
use the Access Grid and Globus
Web services to participate
remotely in pulsed plasma
fusion experiments. The remote
interface provides sensor
readings, data analysis, audio,
and video available in the
control room and allows the
team to discuss what is
happening. The Access Grid is
integrated with Grid services
and applications using the
Globus Toolkit's security and
communication libraries.

24. Distributed Computing
• Multiple CPUs across multiple computers over a network,
working together to solve distributed computing
problems
• A process is split up into parts that run simultaneously on
multiple computers communicating over a network
• A form of parallel programming. Parallel programming
usually means vector processing of data or multi tasked
programming.

25. Distributed Computing

26. Distributed Computing
• Distributed Problem Domains:
- Large amounts of data
- Lots of Computation
- Parallelizable applications
- Many computers
• Indexing the World Wide Web (Google)
• Analyzing the Internet threats (Trend Micro)

27. Distributed Computing
• Distributed System
- An aplication that executes a collection of protocols to
coordinate the actions of multiple processes on a network,
such that all components cooperate together to perform a
single or small set of related tasks
* Fault Tolerance
* High Availability
* Recoverability
* Consistency
* Scalability
* Performance Predictability
* Security

28. Distributed Computing
• Distributed System
- An aplication that executes a collection of protocols to
coordinate the actions of multiple processes on a network,
such that all components cooperate together to perform a
single or small set of related tasks
* Fault Tolerance
* High Availability
* Recoverability
* Consistency
* Scalability
* Performance Predictability
* Security

29. Distributed Computing
• Expectation and handling of failure on both hardware and
software
• 8 Facilities:
- The network is reliable
- Latency is Zero
- Bandwidth is infinite
- The network is secure
- Topology does not change
- There is one administrator
- Transport cost is Zero
- The network is homogeneous

30. Distributed Computing
• One of the basic architectures in
distributed computing – Client-Server
• Clients request services from
servers over a computer network
• Services are provided by a group of servers of a particular
type
• DNS services. Web Services

31. Distributed Computing
• Actual Service server processing client requests may base
on locality or load balance
• Locality
- Local machine Local server Remote server
• Load balance
- Spread request works among serice servers
- Optional resource utilixation, maximize performance,
minimize response time

32. Distributed Computing
• Data replication
- Multiple copies of service data at multiple servers
- Permit access of multiple locations
- Increase availability
- Avoid SPOF (Single Point of Failure)
• Caching
- Local data copy for quick access
- Validate cached data before use commonly TTL
- Browser and web proxy cache frequently accessed
web pages

33. Distributed Computing
• Reliable communication via TCP/IP protocols
• Internet Protocol (IP)
• Protocol for communication data accross a packet
switched internetwork
• Addressing methods IPv4 or IPv6
• Transmission Control Protocol (TCP)
• A reliable stream delivery service that gurantees in order
delivery of data stream between sender dan receiver
• Socket programming socket(), bind(), connect(), send(),
recv() ...

34. Internet Stack
Application Layer
- HTTP, FTP, RPC, SSH
Transport Layer
- TCP, UDP
Network Layer
- IP, ICMP
Link Layer
- Physical Layer
Transmission of Data
- Ethernet, MAC
Address

35. Google Bigtable

36. Distributed Computing
• Hadoop Distributed File System (HDFS)
- Distributed file system designed to run on COTS
hardware
- Detection of failures and automatic recovery
- High throughput of streaming data access
- Batch processing rather interactive
- Large data set, favor large files
- Write once read many coherency model
- Moving computation is cheaper than moving data
- Portability (Java)

37. MapReduce
• A simple programming model that applies to many large
scale computing problems and processes large data sets
• Automatic parallelization and distribution of large scale
computations on a large cluster of commodity machines
• It allows programmers without any experience with
parallel and distributed systems to easily utilize the
resources of a large distributed system

38. MapReduce
• Automatic
parallelization and
distribution
• Fault tolerance via re-
execution
• I/O scheduling
• Status and Monitoring

39. Web 2.0
• The web as a platform
- Communication
- Information sharing
- Collaboration
- Interaction
- Content
• Delivery of web sites to
users as services
• Allow users to run software
applications via browsers

40. Software as services
• A model of software deployment whereby a provider
licenses an application to customers for uses as a service
on demand
• Globally and remotely access and manage commercial
software via web
• No client side software/hardware installation,
maintenance update. Reduce total cost of ownership
• Integration of interconnected software services, Mashups.

41. Software as services Providers
• Trend Micro Hosted Security
• Google Apps
• Salesforce.com Customer Relationship Management
(CRM)
Marc Benioff (Salesforce)

42. Platform as services
• By salesforce.com. Delivery of a platform as a service for
building and hosting web applications
• Cloud based web application development testing,
deploymenr, hosting lifecycle
• http://www.youtube.com/watch?v=EbnCUqAL6UM

43. Platform as services Providers
• Force.com
(from salesforce.com)
• Google App Engine
• Microsoft Asure
Services Platform
• Amazon Web Services

44. Infrastructure as services
• (originally Hardware as a Service, HaaS) delivery of
computing infrastructure as a service
• Platform/hardware virtualization environment
• Server, Network equipment, RAM, disk, CPU etc
• Dynamic resource allocation based on the needs of your
applications
• Only pay for what you use

45. Infrastructure as services Providers
• Amazon EC2
(Elastic Compute Cloud)
- a web service that
provides resizeable
compute capacity in
the cloud

46. Cloud Computing
• Integrate the concept of IaaS - PaaS - SaaS - Web 2.0 and
related technologies (ex MapReduce, Ajax, Virtualization)
• Based on the internet cloud to satisfy the computing
needs of the users on demand
• A distributed computing model
• In the internet cloud use dynamic, scalable and virtualized
computing resources to provide we services

47. Integrate the concept of IaaS - PaaS - SaaS

48. Cloud Computing
• For complex computation on large scale data, use dynamic
and virtualized computer clusters to execute distributed
and parallel computation in the Cloud
• The architecture of the software systems involved in the
delivery of clouding computing. It typically involves
multiple cloud components communicating with each
other over application programming interfaces, usually we
services

49. Cloud Architecture
• Design of software applications that use internet
accessible on demand services
• Applications built on Cloud Architectures are such that the
underlying computing infrastructure is used only when it is
needed (for example to process a user request), draw the
necessary resource on demand (like compute servers or
storgae), perform a specific job, then relinquish the
unneeded resources.

50. Cloud Architecture - Platform
• Cloud platform for developing
and hosting the application
services
• Google App Engine
- Let you run your web
applications on Google
architecture
- Java and Phyton runtime
environments
- Full support for common
web technologies and
services
- Distributed Datastore (Big
Table)

51. Cloud Architecture - Platform
• Hadoop MapReduce
- O
p
e
n
s
o
u
r
c
e
d
i
s
t
r
i

52. Cloud Architecture - Storage
• Distributes persistent data storage in the cloud
• Unstructured data storage
- Traditional file system
* HDFS
- Key Object Pair
* Amazon S3 (Simple Storage Service)
• Structured Data storage
- Amazon SimpleDB
- Google App Engine DataStore

53. Cloud Architecture - Infrastructure
• Infrastructure as a Service
- Pay for what you use
- On demand
• Virtualized computing infrastructure
- Computer clusters
- Hardware abstraction
- Virtualization
* Xen, VMWare

54. Cloud Architecture - Infrastructure
• User configure computing
environment
- Operating System,
networking
- Memory, Disk, CPU

55. About Cloud

56. Private Cloud

57. Virtualization
Virtualization Properties

Virtualization Layer

58. Virtualization
Service Level Assurance (Automate)
High Availability Dynamic Resource On Demand Automate
Scheduling Capacity
• High Availability
Virtual Center (Manage)
• Dynamic Resource Balancing
Virtual Machines
• On Demand Capacity
Manage
Aggregated Virtualization • Resource pooling & allocation
(Consolidate & Simplify)
• Instant Provisioning
Server Farm
• Easy, Non-disruptive Scaling
Consolidate
Network
• Multiple Workloads / Server
Storage • Simplified Management
• Opex and Capex Savings

59. Virtualization
Automated Resource Increased Availability
+
On Demand Capacity
+
Assurance
Dynamic Balancing Automated Non-disruptive Scaling
Continuous Optimization Across Applications Flexible, Reconfigurable
X

60. Today’s Challenges
SPEED EFFICIENCY LEGACY COMPLEXITY
Not fast enough Need more with Must use what Too much everywhere
less you have
PROCESSES INNOVATION ACCESS COLLABORATION
Too hard to Not able to use Limited Too difficult
deliver

61. How We Defines Cloud
Cloud computing increases business agility by offering
access to a pool of computing resources (e.g. networks,
servers, storage applications and service) either on or
off-premise.
• There is a lot of hype of around cloud…it is simply a
long-term movement away from local processing to
networks and web-based applications.
• Cutting through the hype can be a daunting
experience, with costly missteps along the way….

62. Cloud Enables the Efficient Enterprise

63. Cloud Enables the Efficient Enterprise

64. Cloud Enables the Efficient Enterprise

65. Cloud Services Point-of-View
• While there is a lot of needless hype around cloud – it will
fundamentally transform the way many companies do business
• How a company uses cloud depends on the type and size of the
company and how it currently handles business processes
• Cloud offers the possibility of cheaper and faster but it some cases it
simply makes it “better”
• Cloud is a journey with many paths and incremental improvements
along the way to minimize risk and maximize profit to find cloud that
best meet their business needs.

66. Where are you on the Journey
• In early stages of virtualization
• Need to manage their virtual infrastructure to increase
business agility and infrastructure optimization
Flexible
Sourcing
• Manage 1000’s of applications with complex infrastructure
Consumption Based
and business process requirements Billing
• Have significant investments in legacy technologies, do not Self-Service Consumption
Dynamic Resource Pooling
want to forklift upgrade their infrastructure
• Have their own specific business needs and realize that
Fully Stateless Infrastructure
“one size fits all” does not apply
Rapid Deployment
Improving Availability
Enabling Mobility
Improving Utilization

67. Where Customers are on the Journey
• In early stages of virtualization
• Need to manage their virtual infrastructure to increase
business agility and infrastructure optimization
Flexible
Sourcing
• Manage 1000’s of applications with complex infrastructure
Consumption Based
and business process requirements Billing
• Have significant investments in legacy technologies, do not Self-Service Consumption
Dynamic Resource Pooling
want to forklift upgrade their infrastructure
• Have their own specific business needs and realize that
Fully Stateless Infrastructure
“one size fits all” does not apply
Many are here,
early in the journey Rapid Deployment
Improving Availability
Enabling Mobility
Improving Utilization

68. Example

69. Thank You
michael@sunggiardi.com