Cloud computing: state- of-the-art and research challenges

meatcologneInternet και Εφαρμογές Web

3 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

68 εμφανίσεις

Cloud computing: state
-
of
-
the
-
art and research
challenges

Qi Zhang, Lu Cheng,
Raouf

Boutaba

Presented by Abhishek Sharma CS 598

Agenda


Overview


Related technologies


Architecture and design


Key features and characteristics


State
-
of
-
the
-
art implementation


Research challenges

Overview


NIST definition


Cloud Computing is a model for enabling convenient,
on
-
demand network access to a shared pool of
configurable computing resources that can be rapidly
provisioned and released with minimal management
effort or service provider interaction.


Popularized by Eric Schmidt in 2006

Cloud Service Providers


Two kinds


Infrastructure providers: Manage cloud platforms and
lease resources.



Service Providers: Rent resources from one or many
infrastructure providers to serve end users.

Advantages


No up
-
front investment



Pay
-
as
-
you
-
go model.


Lowering operating cost



Rapid allocation and de
-
allocation


Highly scalable



Easy to handle rapid increase in service demands


Easy access



Web
-
based services accessed through different devices


Reduced business risks



Service provider shifts business risks to infrastructure
providers



Related Technologies


Grid Computing


Distributed computing paradigm that coordinates
networked resources to achieve a common objective
.


Utility Computing


Represents model of providing resources on demand
.


Virtualization


Technology that abstracts away details of hardware and
provides resources for high
-
level abstractions
.


Autonomic Computing


Computing systems capable of self
-
management
.

Cloud Computing Architecture


Layered model


Hardware Layer: Responsible for managing the
physical resources of the cloud.


Infrastructure layer: Creates pool of resources using
virtualization technologies.


Platform Layer: Consists of operating system and
application frameworks.


Application Layer: Actual cloud applications.



Business model


IaaS: On
-
demand provisioning of infrastructural resources,
usually in terms of VMs.


PaaS: Refers to providing platform layer resources, including
OS support and software development framework.


SaaS: Refers to providing on
-
demand applications over the
internet.

Types of Cloud



Public Clouds


Service providers offer their resources as services to
general public.


Private Clouds


Designed for exclusive use by a single organization.


Hybrid Clouds


Facilitate on
-
demand expansion and contraction(public
clouds), but with tighter security and control(private
clouds)


Cloud Computing Characteristics



Multi
-
tenancy


Shared resource pooling


Ubiquitous network access


Service oriented


Dynamic resource provisioning


Self
-
organizing


Utility
-
based pricing


State
-
of
-
the
-
art Implementation


Architecture


Layered design


Uniform high capacity


Free VM migration


Resiliency or Fault
-

tolerance


Scalability


Backward capability


Distributed file systems


GFS: Provides efficient, reliable access to data. Designed
and optimized to run on data centers. Provides extremely


high data throughput, low latency.


HDFS: Stores large files across multiple machines. Data
stored across multiple geo
-
servers. Reliability by replicating
the data across multiple servers.



Distributed application framework


MapReduce

Research Challenges


Automated service provisioning


Satisfying service level objectives while minimizing
operational costs, resource provisioning decisions
must made online.


Virtual Machine Migration


Avoid hotspots. Currently, detecting workload
hotspots lacks agility to respond to sudden workload
changes



Server consolidation


Research challenges: Resource congestion



Energy management


Research challenges: Designing energy
-
efficient
data centers



Traffic management


Research challenges: Higher density of links in data
centers. Exiting methods compute traffic matrices
between few hundreds of data hosts, data center may
have thousands. Change in traffic patterns


Data Security


Research challenges: Goals


confidentiality and
auditability. Remote attestation(for auditability) not
sufficient.



Software Framework


Research

challenges
:

VM

allocated

to

each

Hadoop/MapReduce

node

may

have

heterogeneous

characteristics,

performance

modeling

of

Hadoop

jobs
.



Storage

technologies


Research

challenges
:

Compatibility

issues

with

legacy

file

systems

and

applications
.



Novel

cloud

architectures


Research

challenges
:

High

energy

expense,

high

initial

investment

for

constructing

data

centers
.

Summary


Cloud computing rapidly changing the landscape
of information technology.



Despite benefits, current technologies are not
developed enough to realize full potential.



Many challenges only starting to receive
attention.