CLOUD COMPUTING

WHAT IS
CLOUD
COMPUTING?
It is Internet-based computing, whereby shared
resources, software, and information are
provided to computers and other devices on
demand.
 HOW DOES CLOUD
COMPUTING WORK?
1. Various providers let you create virtual servers
1. Set up an account, perhaps just with a credit card

2. You create virtual servers ("virtualization")

1. Choose the OS and software each "instance" will have
2. It will run on a large server farm located somewhere
3. You can instantiate more on a few minutes' notice
4. You can shut down instances in a minute or so

3. They send you a bill for what you use

  System
serve Adminis
rs as trator
hardw a who s ofte
are, t le un n talk
appli he O S it tha a b ou

o na l
catio , the t i nclud t

cept
Serve
rs are
ns. storag
e, and
es the
funct o f t he
t
raditi
r C on ion i. en re
SQ L e. the ferred
serve Exch t o by
 If the
File s
r, t h e File a n
serve
g e serve
their
r, the
The T

Exch e r, etc.
ange rver fills
Serve

then serve up, o

the S r b ec r the
add i ystem omes
n a ne Adm overt
w ser inistr axed,
ver. ators
must
 The Traditional Server Concept

Web Server App Server DB Server EMail

Windows Linux Linux Windows
IIS Glassfish MySQL Exchange
  Unless there are multiple
o na l
servers, if a service experiences

cept
a hardware failure, then the
service is down.
raditi
r C on
 System Admins can implement
clusters of servers to make them
The T

more fault tolerant. However,

Serve

even clusters have limits on

their scalability, and not all
applications work in a clustered
environment.
 Cons
Pros
 Expensive
Easy to conceptualize
to acquire and maintain

The T
hardware
 Fairly easy to deploy

Serve

 Not
Easyvery scalable
to backup

 Difficult
Virtually to
anyreplicate
application/service can be


r ad i t i
run from thisistype
Redundancy of setup
difficult to implement

r C on
 Vulnerable to hardware outages
 In many cases, processor is under-

o n al
utilized

cept
  Virtual servers can be scaled out

r
easily.

ept
Serve  If the administrators find that the
resources supporting a virtual server are
Conc being taxed too much, they can adjust
the amount of resources allocated to that
rtual

virtual server
 A virtual server can be serviced by one
or more hosts, and one host may house
i
The V

more than one virtual server.

 Virtual servers themselves can be
migrated from host to host almost at
will.
 Cons
Pros

The V
 Slightly
Resourceharder to conceptualize
pooling
 Slightly more costly (must buy
Highly redundant
hardware, OS, Apps, and now the
 Highly available

rtuali
abstraction layer)
 Rapidly deploy new servers
 Easy to deploy

Co n c


Serve
Reconfigurable while services are
running
 Optimizes physical resources by doing

e pt
more with less

r
  You don’t have to own the
hardware
es
ext
 You “rent” it as needed from a
step
lizati uting tak
the n cloud
 There are public clouds
on to

 e.g. Amazon EC2, and now many

p
d com

others (Microsoft, IBM, Sun, and

others ...)

 A company can create a private

Clou
virtua

one
 With more control over security,
etc.
 LETS LOOK AT
THE VARIOUS GOALS
ACHIEVED BY
CLOUD COMPUTING
 Cost
os t
 Many systems have variable

rol
demands
1–C
Cont  Batch processing (e.g. New York
Times)
 Web sites with peaks (e.g. Forbes)
Go a l

 Startups with unknown demand (e.g.

the Cash for Clunkers program)
 Reduce risk
 Don't need to buy hardware until you
need it
  More than scalability -
ss
elasticity!

ty
u sine  Used to take 3 - 4 months to give a
Ag ili department a server cluster, then
they would hoard it!

 Using MICROSOFT CLOUD,

2-B

about 5 minutes!
 And they give it back when they are
Go a l

done!

 Scaling back is as important as

scaling up
  Various p
virtua rovide
l serve rs let y
rs ou cre
S et up a ate

ud
credit n account, p

rks
card erhaps
w Clo  You create
just w
ith a
virtual
g Wo ("virtu
alizati
on")
server
s
 Choose th
"instan e OS and
o

ce " wi softwa
putin

 l l have re eac
H

It will h
run on
somew a large
here server

You ca farm l
ocated
Com

n insta
minut n
es' not tiate more
 You ca
n shut
ice on a fe
w
or so down
instan
 They s
end y
ces in
a minu
te
ou a b
ill for
w
  Large s
rental cale infra
basis struct
 Operat ure av
ailabl
provi in g S e on a
des C ystem vir
 “Roll-y
ds
P U is tualiz
o ation

)
o u lation (e.g.
r-own

(Now
provi ” netw Xen)
C l ou des n
 Locall etwork iso ork provi

Fully
custo
y spe
cific
storag
lation sionin
g
 Service mer self- e abstractions
adver Level Ag
servic
e
ic

tized reem
 Reques ents (
S LAs
Publ

via w t s are acc ) a re

 Customeb services epted and resou
Intern ers acces rces g
rante
 Acco
untab
et s reso
urces
remo
d
tely v
 Web-b ility is e-c ia the
 “Pay-a ased transacti ommerce b
 Customs-you-go” andon ased
er ser flat-r
vice, ate su
r e f un bscrip
ds, et tion
c.
  Public
cloud
 What s are o
a
cloud pplicatio paqu
e
 Many
? ns w
ill wo
rk w e

rk?
Publi o f th e a ll in a
c Clo dv
antag
prem uds a es of
e y wo ise” I
 Self-se T
ppear
usefu
fered
by
l for
 Lega ce provi
r v i “on
 Flexibcy support sioni
n g
 What e
d o th

le res
ource
x t e alloc
are re nsion ation
quire s or m
varie d to s odifi
ty o f uppo catio
Ho w

 Data a services rt a wide ns

 Multip ilation nd applica r
s s im a
 Mobi layer gamin tions
?
le de g
v ices
Infrastructure (Mega Datacenters)
 Open Source Cloud Infrastructure

 Simple
 Transparent => need to “see” into the cloud
 Scalable => complexity often limits scalability
 Extensible
 New application classes and service classes may require new features
 Clouds are new => need to extend while retaining useful features
 Commodity-based
 Must leverage extensive catalog of open source software offerings
 New, unstable, and unsupported infrastructure design is a barrier to uptake,
experimentation, and adoption
 Easy
 To install => system administration time is expensive
 To maintain => system administration time is really expensive
SERVICES OFFERED BY
CLOUD
Cloud computing

Infrastructure as
Software as service Platform as service
service
 Cloud Computing – Simple Definition

Cloud Computing = Software as a Service

+ Platform as a Service
+ Infrastructure as a Service

 Software as a Service (SaaS)

 From end user’s point of view
 Apps are located in the cloud
 Software experiences are delivered through the Internet
 Cloud Computing – Simple Definition

Cloud Computing = Software as a Service

+ Platform as a Service
+ Infrastructure as a Service

 Platform as a Service (PaaS)

 From developer’s point of view (i.e. cloud users)
 Cloud providers offer an Internet-based platform to developers
who want to create services but don't want to build their own
cloud
 Cloud Computing – Simple Definition

Cloud Computing = Software as a Service

+ Platform as a Service
+ Infrastructure as a Service

 Infrastructure as a Service (IaaS)

 Cloud providers build datacenters
 Power, scale, hardware, networking, storage, distributed systems, etc
 Datacenter as a service
 Cloud users rent storage, computation, and maintenance from
cloud providers (pay-as-you-go; like utility)