cloud computing: evolution or redefinition

11/10/2012

1

Cloud Computing: Evolution or Redefinition

Prof. Mario Dantas Federal University of Santa Catarina (UFSC) Informatics and Statistics Department (INE)

Florianópolis - Brazil E-mail: [email protected]

This talk addresses fundamental concepts of cloud computing which are related to parallel and distributed environments. This is followed by a discussion of challenges faced by this computational paradigm in order to meet the requirements of applications from different domains. Additionally, limitations of the cloud computing paradigm will be highlightedand finally commercial and academic study cases will be presented.

Motivation for this talk a) Internet x Web Number of computers connected to the internet

Date Computers Web Server

1979, Dec 188 0

1989,July 130,000 0

1999, July 56,218,000 5,560,866

Motivation for this talk b) Downsizing x Rightsizing (90's) From centralised environments to distributed computing

Motivation for this talk c) Unix x Linux Standards (BSD and AT&T) against standard

Motivation for this talk d) wired x wireless networks (90's and early 00's) Reliable, high speed links against unsecure and slow networks

11/10/2012

2

Motivation for this talk e) HPC x Cloud Environments Are these environments excludents or complementaries paradigms?

Motivation for this talk f) Academic x Commercial Approach Several challenges in the computer industry were studied in the past with an initial collaboration from the academic community. BUT, without this approach we can have figures like these:

“I’ve never seen something more powerful than this computation combined with this network that we now have... In the last seven years, do you know how many times I’ve lost any personal data? Zero. Do you know how many times I’ve backed up my computer? Zero.” – Steve Jobs, 1997.

Motivation for this talk

g) A special user view

Agenda > Concepts > Challenges > Requirements > Limitations > Study Cases > Conclusions and Recommendations


• CPU • Memory

• Output

• Device • Input

• Device

l Main bus

l I/O bus

l Von Neumann Architecture

Concepts

l Architecture evolution

11/10/2012

3

l Source: [www.nvidia.com]

l Fermi Architecture

Concepts


l Old fashion clusters

Concepts


•

• Collaboration

• Groupware

• Virtualization

• Web Portal

• Remote Viz.

• [Dominic Lam, IBM]

Concepts l Architecture evolution

l Cloud computing buzzwords

l PaaS l PaaS

l Private

l Cloud l Public

l Cloud

l SaaS

l EC2 l Google

l APPs

Concepts

But, how can we understand these cloud buzzwords?

Concepts

Cloud computing is a model > for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) > that can be rapidly provisioned and released with minimal management effort or service provider interaction.

Concepts

11/10/2012

4

This cloud model is composed of:

> five essential characteristics; > three service models; and > four deployment models.

Concepts

Essential Characteristics: 1) On-demand self-service; 2) Broad network access; 3) Resource pooling; 4) Rapid elasticity; 5) Measured service.

Concepts

Service Model: 1) SaaS (Software as a Service); 2) PaaS (Plataform as a Service); 3) IaaS (Infrastructure as a Service).

Concepts

l [Zhang, Cheng, Boutaba, 2010]

Concepts

Concepts

Service Model

Concepts

11/10/2012

5

Deployment Models: 1) Private cloud. The cloud infrastructure is provisioned for exclusive use by a single organization 2) Community cloud. The cloud infrastructure is provisioned for exclusive use by a specific community of consumers 3) Public cloud. The cloud infrastructure is provisioned for open use by the general public 4) Hybrid cloud. The cloud infrastructure is a composition of two or more distinct cloud infrastructures

Concepts

“We will probably see the spread of ‘computer utilities’, which, like present electric and telephone utilities, will service Individual homes and offices across the country.”

[1969, Len Kleinrock]

But, is this idea of cloud effort new?

Concepts

“A computational grid is a hardware and software infrastructure that provides dependable, consistent, pervasive, and inexpensive access to high-end computational capabilities.”

Concepts

[1998, Ian Foster and Carl Kesselman]


Challenges

But, does the cloud approach have any challenge?

Challenges

But, does the cloud approach have any challenge?

Essential characteristics Examples to remember 1) On-demand self-service; Self-service restaurant queues; 2) Broad network access; Different roads from different states/provinces/countries; 3) Resource pooling; Shared or distributed memory; 4) Rapid elasticity; EURO monetary elasticity; 5) Measured service. bills never delay.

11/10/2012

6

Challenges

> Interoperability: Lack of (or to many) standards;

> Human Resources: Small number of people with real

good knowledge of distributed systems;

> Security : . Byzantines attacks; . not yet known...... (e.g. Cyber attacks such as Stuxnet and flame)


Requirements

But, which are the requirements to have a cloud environment?

Deployment Models: 1) Private cloud. It may be owned, managed, and operated by the organization, a third party, or some combination of them, and it may exist on or off premises. 2) Community cloud. It may be owned, managed, and operated by one or more of the organizations in the community, a third party, or some combination of them, and it may exist on or off premises.

Requirements

Deployment Models: 3) Public cloud. It may be owned, managed, and operated by a business, academic, or government organization, or some combination of them. It exists on the premises of the cloud provider. 4) Hybrid cloud. cloud infrastructures (private, community, or public) that remain unique entities, but are bound together by standardized or proprietary technology that enables data and application portability (e.g., cloud bursting for load balancing between clouds).

Requirements


11/10/2012

7

Limitations

But, does the cloud computing approach have any limitation?

Limitations

> Heterogeneity: cloud providers facilities;

> Elasticity: this number is not known yet;

> (full) Interoperability: between low level applications (e.g. virtual machine tools);

> Human resources: developers for this

Hybrid environment (shared, distributed memory and GPU environments)


Study Cases Public Clouds

l Public Clouds


Study Cases

11/10/2012

8

l AZURE (PaaS)

Study Cases

l Source: [www.amazon.com]

l S3 (PaaS)

Study Cases

l S3 (PaaS)

Study Cases

l Source: [www.amazon.com]


Study Cases

Study Cases Private Clouds (UFSC/CTC/INE/LaPeSD)

11/10/2012

9

Study Cases Private Clouds (UFSC/CTC/INE/LaPeSD)

l [Dantas et al., 2009]

l Private Cloud

l Interface

l Services

l Catalog

l Provisioning

l Tools

l System

l Management

l Monitoring

Study Cases

l A) Context oriented approach

Study Cases

Resource Reservation

Study Cases

l User Centric Authentication

Study Cases

l Spatio-Temporal Model

Study Cases

11/10/2012

10

l Experimental Results

l Experimento 1 – Comparação das abordagens de

autenticação adotando o consumo energético como

métrica

Study Cases

l Experimento 2 – Comparação da eficiência das

abordagens de autenticação em termos percentuais

Study Cases

l Experimental Results

l VO-C l # processors l VO-B l # processors

l Cluster_01 l 7 l Cluster_1 l 4

l Cluster_02 l 5 l Cluster_2 l 4

l Cluster_03 l 3 l - l -

l Multi-cluster (cloud) configuration


Study Cases

l B) Ontology Approach


l Reference Brazil Canada

Study Cases

l Ontology Approach

• Interactive Interface

• Test without dynamic information


Study Cases Experimental Results


• Test with dynamic information

Study Cases


Experimental Results

11/10/2012

11


• No dynamic information considered

Study Cases




• Dynamic Information

Study Cases



C) Advance reservation of resources through augmented reality

Study Cases


a. The cloud approach can be considered now as redefinition on how to use several existing paradigms, such as:

Grid computing; l Utility computing; l Virtualization; l Autonomic computing.

a.b. This centralized environment adopts several approaches from:

a. > computer architecture; a. > computer networks: a.> distributed and parallel computing b. paradigms;

11/10/2012

12

a.c. The cloud approach has some news challenges to be considered in large scale:

a.> heterogemeity (hw & sw); a.> security; and a.> mobile computing interaction.

d. The cloud approach in the future could be a. considered as an Evolution. if the used paradigms form a new approach

•Recomendations

• Consider multi- disciplinary subjects, such as: DB; Distributed systems (e.g. Mosix, Condor); Mobile computing (e.g. sensors); Programming languages (e.g. CUDA, PGAS); ●Computer newtorks (e.g. Infiniband, Quadrics); ● Computer architectures (e.g.multi-computers, multi-procesores);

• When developing to cloud environments, do'nt forget: > The context aware, ontology and fault tolerance approaches; > How a better computing performance can help (e.g. speed up web paradigm); > Mobility facilities; > Security issues.

•Recomendations

Cloud Computing: Evolution or Redefinition

QUESTIONS?

Prof. Mario Dantas Federal University of Santa Catarina (UFSC) Informatics and Statistics Department (INE)

Florianópolis - Brazil E-mail: [email protected]

cloud computing: evolution or redefinition

Documents

cloud model

cloud approach

community cloud

public cloud

hybrid cloud

private cloud

cloud computing buzzwords

cloud computing paradigm