1.
NIST: national institute of science and technology
2. Cloud computing: model for enabling ubiquitous, convenient, on demand network access to
shared pool of resources
3. Chracterisitics of cloud computing
a. On demand self service: consumer can unilaterally provision computing capabilities
(such as time and storage) as needed automatically without requiring human
b. Broad network access: access share pole with different devices
c. Resource pooling: resources are pooled to serve multiple consumer according to
demand. Consumer has no control over the exact location of resource
d. Rapid elasticity: user ki demand ky matibiq services km zyada kr skty
e. Measured service: cloud systems automatically control and optimize resource. This is
done on pay per use basis
4. John Mccarthy present idea of cloud computing in 1961. In 160-1970 mainframe was leased
out by manufacturer. Idea of grid computing emerges in 1990 to use processing power of
networked pc for calculation. In 1990 [Link] start bringing remote software services
to enterprises. Amazon web service (AWS) were launch in 2002. In 2006, cloud computing
emerges to enable organization to lease computing capacity from cloud provider
5. Cluster computing: collection of interconnected standalone computers to act as single
resource. It become popular in 1990
6. HPC: high performance computing
7. In 2010 out of 500 supercomputer, 85% were computer cluster build with homogenous
nodes
8. Benefit of cluster computer
a. Scalability
b. High availability and fault tolerance
c. Use of commodity computers
9. Grid computing: integrated computing infrastructure for bringing together computers to
form large storage and network resource
10. Grid is used for large scale computation
11. Building a grid is done on voluntary basis. SETI@Home is applied over 3 million computers
to achieve 23.37TFlPos
12. Layers
a. Application: Top layer of grid consist of user application to be run on grid
b. Collective services: interaction among resources and provided link to connecting
services
c. Resource service: resource mlny ky bd usy manage krna
d. Connectivity layer: connecting server and network
e. Grid fabric: consist of all computational resource
13. Purpose of grid computing is to solve large scale problem. Difference between cluster and
grid computing is
a. Grid make use of computational resources spread over globe
b. These resources are owned by different organizations
c. Grids heavily depend on WAN/LAN
d. Virtual supercomputer is term derive from grid computing where multiple computers
collaborate together to make an illusion big computer
e. Several organization group together to form grid in shape of virtual organization (VO)
14. Software as a service (SaaS): capability provide consumer to use the provider application
running on cloud infrastructure
15. , platform as a service (PaaS): capability provides to the consumer to deploy onto cloud
infrastructure created or acquired application using program language
16. Infrastructure as a Service (IaaS): capability is to provide consumer to provision processing
where consumer able to deploy arbitrary software
17. Cloud infrastructure is collection of hardware and software that enable 5 essential
characteristics of cloud computing
18. Cloud deployment model:
a. Private cloud: provisioned for exclusive use by single organization. May exist on or off
premises
b. Community cloud: provision for exclusive use by specific community that have
shared concerns. Owned by one or more organizations
c. Public cloud: provision for open use by public. Exist on premises of cloud provider
d. Hybrid cloud: combination of two or more distinct cloud infrastructure
19. Mainframe: large server that can handle hundreds of server simultaneously. For example
IBM’s Z series can handle 500 server. In 1960-1970 mainframe were leased out by
manufacturer. Customer were charged on monthly basis. This model still exist in cloud
computing
20. Server: computer provide services to other computer. Server type
a. Web server: store websites and web apps and provide them on your desktop
b. Domain name server (DNS): store domain name and IP address
c. Database server: hosts database and provide access to data
21. Desktop: computer design to remain in stationary position
22. Data communication: exchange of data over medium. Factors essential for data
communication
a. Data must be delivered to correct destination
b. Data must not be altered
c. Timely delivery
23. There must not be uneven delay among packet arrival in audio or video
24. Component of data communication
a. Message
b. Sender
c. Receiver
25. Transmission medium: physical path. Protocol: set of agreed upon communication rules
between sender and receiver. 2 devices cannot communicate without protocol
26. Data can be represent by
a. Text: represented by bit pattern called code eg Unicode and American standard code
for information interchange (ASCII)
b. Numbers: directly convert binary of the numbers. ASCII is not used to represent
numbers
c. Images: sent as binary patterns. Image is represent by binary pixels
d. Audio: continuous stream of data
e. Video: can be continuous or sequence of image combinations
27. Data flow
a. Simplex: unidirectional. Eg keyboard, monitor
b. Half duplex: both device can communicate but one at a time. Eg walkie talkie
c. Full duplex: both devices can send and receive at same time. Eg telephone
28. Computer network are conceived in 1960 after invention of computer
29. Network is collection of computers and devices connected through transmission media
30. Connecting device:
i. Router: connect network with other networks
ii. Switch: connect device with network
iii. Modem: change form of data
31. Network criteria is measure by 2 metric, throughput and delay
32. Throughput: bulk of data transmitted in unit of time and delay. Increase throughput may
increase congestion and increase delay
33. Transit time: transit time and response time indicate network performance
34. Reliability: time to recover from a failure
35. Security: protecting data from unauthorized access
36. Link can be dedicated link (point to point) or shared (multipoint)
37. Physical topology:
a. Mesh topology: every device connected to every other. Robust but congested cables
b. Star topology: all devices connected to central device. Require only one I/O but if
central device fail whole system falls
c. Bus topology: one long cable act as a backbone. Easy of installation but signal drop
over length
d. Ring topology: connected in form of ring. Easy to expand but failure of one device
disable whole network
38. Network type:
a. LAN: privately owned network. Each packet contain source and destination/ 1Mbps
to 1Gbps
i. Ethernet (CSMA/CD): carrier sense with multiple access with collision
detection
ii. Local talk: CSMA with collision avoidance
iii. Wireless LAN: range <100m and speed 2Mbps
iv. Token ring: token travel along the network in ring
v. FDDI: token ring with fiber optic.
vi. ATM: star based,
b. WAN: wide area network, interconnected multiple LANs. WAN unlike LAN is owned
by network company
i. P2P WAN: connecting 2 devices through wired or wireless
ii. Switched WAN: network with more than 2 ends
c. MAN: Metropolitan area network is combination of large LAN and WAN. 5-50 km
39. Switch: connects two link to forward data from one another. . switch does not store data just
forward it. Types of switch
a. Circuit switch: dedicated physical connection between 2 devices
b. Packet switch: send in form of packets such as router. Multiplexing is used to forward
data
40. The internet: network of thousand of interconnected networks
41. Advance research project agency (ARPA) in mid of 1960 to share military info with eachother
42. TCP/IP was proposed in 1973. In 1981 it was included in OS. 5 layers
a. Application:
b. Transport: transport msg from application. Main protocol are TCP (transmission
control protocol and user datagram protocol (UDP). UDP is light and not connection
oriented
c. Network: host to host communication. Packets are called datagram
d. Data link: create frame of data, add MAC address of sending and recieving
e. Physical: bit to signal
43. TCP message: segment
UDP message: datagram
44. HTTP: hyper text transfer protocol, FTP: file transfer protocol, SMPT: simple mail transfer
protocol
45. IPV4 is 32bit, total address 232 and IPV6 is 128 bit, total address 2128.
46. IP address change with network and can have 2 ip address if it has 2 connection
47. ICANN: internet corporation for assigned names and number. Small organization can get IP
addresses from ISP (internet service provider)
48. IPV4 Ip address consist of prefix and postfix.
a. Class A: 8 bits for network. Total network: 27 first byte 0 to 127
b. Class B: 16 bits for network. Total network: 214 first byte 128 to 191
c. Class C: 24 bit for network. Total network: 221 first byte 192 to 223
d. Class D: used for multicasting. No prefix or network ID. First byte 224 to 239
e. Class E: reserved for future use. First byte 240 to 255
49. Classful address leads to depletion of IP address. Solution to this is address masking variable
sizes prefixes
50. DHCP: dynamic host configuration protocol. It is used to auto assign IP address from
acquired block of IP address
51. NAT (network address translation): organization can use private IP for LAN and use NAT
mechanism for having few IP to connect to internet. It replaces local IP with registered IP
52. Address resolution protocol (ARP): if any device want to find and IP address it sends with its
IP address signal and particular device replies with sender device address
53. IPV4 to IPV6
a. No more NAT
b. No more private address collision
c. Builtin authentication support
d. Easier administration (no more DHCP)
e. Simplified routing
54. 3 categories of IPV6
a. Unicast: for single connection
b. Multicast: one message transmitted to all
c. Anycast: one msg transmit to single interface
55.