What the reader will learn:
• How
service-oriented architectures relate to cloud computing
• About
the ‘cloud stack’
• Details
of the three service models
• About
the kind of services that are being offered by the major vendors
• How
to create a basic cloud application on Google App Engine
5.1 Introduction
 |
Wikipedia de fi nes cloud computing as:
the delivery of computing as a
service rather than a product, whereby shared resources, software, and
information are provided to computers and other devices as a metered service
over a network (typically the Internet). ( http://en.wikipedia.org/wiki/Cloud_computing )
The GartnerGroup similarly lists
‘servicebased’ as one of the key de fi ning
attributes of cloud computing. Indeed, we could say that one of the defi
n ing c haracteristics that makes cloud computing different from traditional
software development and deployment is the focus on service delivery. The
National Institute for Standards and Technology (NIST) also sites services as
an essential component of cloud computing and defi n es three service models.
I n this chapter we are going to
explore the different ways in which you can build and use services in the
cloud. We will examine what is known as the SPI model where the letters stand
for ‘services’, ‘platforms’ and ‘infrastructures’ and see how the model applies
to the systems currently available in the cloud. As an emerging and rapidly
advancing area, a number of key terms used in cloud computing do not yet have
an agreed designation. However, NIST provides the most widely used defi n
itions, particularly relating to the SPI model, and so we include them in the
relevant sections below. We also include a number of example technologies for
each
R. Hill et
al., Guide to Cloud Computing: Principles and Practice, Computer 91
Communications and Networks, DOI 10.1007/978-1-4471-4603-2_5,
© Springer-Verlag London 2013
of the service types. However, the list is far from
exhaustive, and of course, new services and providers are continually emerging
and transforming the rapidly changing ‘cloudscape’. At the end of this chapter,
you will develop the virtual machine you created previously and then use it to
create and deploy your own applications in the cloud. We should note that
although we do touch on data storage issues in this chapter, the topic is
discussed in detail in Chap. 6 .
T here are a number of technologies
which have been identifi e d as precursors to cloud computing such as grids,
utility computing, autonomic computing and virtualisation. Perhaps the most
relevant to cloud services is the development of web services and
service-oriented architectures (SOAs). We begin this chapter by examining these
technologies which are still highly relevant to the ongoing evolution of cloud
services.
5.2 Web Services
A web service
refers to software which provides a standardised way of integrating facilities
offered by web applications and supports communication between organisations
without requiring detailed knowledge of how the services are implemented or
even which language or platform they are implemented on. Services should
perform easily describable, focused and isolated tasks such that they are
independent of the particular state of other services. Unlike traditional client–server
architectures, web services provide a programmatic and machine-friendly
interface rather than a graphical user-friendly interface presented in a
browser or desktop application.
XML is the typical format of communication for
web services which provides the signi fi cant advantage of platform
independence. The Web Service Description Language (WSDL) can be used to
describe the services available, to structure requests and responses and to
provide information regarding the kind of protocol needed in order to invoke
the service. Implementations for working with WSDL based on C#, Java and other
languages are available.
Many web services are completely free, whereas
others may charge for their use. If you create a web service, it is obviously
useful if you can publish this fact such that potential users are able to
locate, evaluate and use the web service you are offering. The most widely used
directory of web services is available via UDDI (Universal Description,
Discovery and Integration) specifi c ation which defi n es a way to publish and
discover information about web services. SOAP(Simple Object Access Protocol) is
a standardised XML format commonly used to tag the information and transport
the data. The collection of standards (developed by the World Wide Web
Consortium (W3C)) used for web services are collectively known as the WS*
stack.
Perhaps it is easier to understand web
services by looking at some examples. If you go to the home page of
WebserviceX.NET (h ttp://www.webservicex.net)
, h undreds of web services are listed. You can search by keyword or browser by
popularity or subject category. You should be able to see commonly used
examples such as barcode generators, currency converters and suppliers of
weather data. If you are interested, you can use the site to obtain the WSDL
service description together with the required SOAP structuring and other
useful information regarding the services.
5.5 Composability
5.3 Service-Oriented Architecture
T he goal of service-oriented architecture is to bring
together various pieces of f unctionality to form new applications which are
built largely from existing software services. SOA provides a set of
methodologies for designing software as a set of loosely coupled interoperable
services suitable for use in a distributed computing environment. In a somewhat
similar way to other software engineering approaches such as object
orientation, service-oriented architectures enable us to create components or
building blocks that make sense to human users and designers and can be reused
for different purposes. The complex internal working of the components is
hidden, and a simple interfacewith well-de fi ned behaviour is presented.
The services are self-contained modules
that provide standard business functionality and are independent of the state
or context of other services. Services are described in a standard de fi nition
language and have a published interface.
The W3C refers to
SOA simply as:
A set of components which can be
invoked, and whose interface descriptions can be published and discovered. (
http://www.w3.org/TR/ws-gloss/ )
SOA enables communication between providers
and consumers and gives a mechanism to manage available services. The
communication should be in the form of loosely coupled services or software
functionalities that can be reused for different purposes. In summary we can
say that an enterprise application that follows the SOA paradigm is a
collection of services that together perform complex business logic.
5.4 Interoperability
I nteroperability is a key SOA requirement. SOA
components should be independent of programming language and platform and
should be unassociated with each other. SOA aims to allow users to combine or
mesh collaborating services from a number of sources to build new, potentially
complex applications. SOA allows organisations to expose and access an
application’s services and to consume information bound to those services using
well-de fi ned interfaces.
5.5 Composability
Applications are
frequently built from a collection of components: a feature called
composability. A composable system uses components to assemble services that
can be tailored for a speci fi c purpose using standard parts. One of the
compelling advantages of the SOA approach is the way that services can be
composed when creating applications. Combining and interleaving of services is
the heart of serviceoriented computing. SOA implies that services can
participate equally well in simple or complex compositions. Composability has
been taken to a new level with cloud computing; for example, fundamental
computing resources such as storage and
processing power are provided as services which can be composed to
create a required application environment (see below).
5.6 Representational State Transfer (REST)
R epresentational state transfer (REST) is a style of
software architecture originally introduced by Roy Fielding who was one of the
principal authors of HTTP speci fi cations 1.0 and 1.1. As with HTTP, the REST
architecture envisions clients and servers in a conversation or
request/response cycle. Requests are centred on representations of the resource
and use existing HTTP methods. This is in contrast to other systems, such as
SOAP where users can de fi ne their own methods and resource identifi e rs. In
REST a given URI is used to access the representational state of a resource and
also to modify the resource. For example, a URL on the web can be used to
provide information to users about the resource or to modify the resource:
• GET
is used to transfer the current representational state of a resource from a
server to a client.
• PUT
is used to transfer the modi fi ed representational state of a resource from
the client to the server.
• POST
is used to transfer the new representational state of a resource from the
client to the server.
• DELETE
is used to transfer the information needed to change a resource to a deleted
representational state.
REST uses HTTP for all CRUD
(Create/Read/Update/Delete) operations. REST is an increasingly popular
lightweight alternative to mechanisms like RPC (remote procedure calls) and
SOAP. Like SOAP and RPC, a ‘RESTful’ service is platform and language
independent but is considered simpler and easier to use. Whilst REST services might
use XML in their responses (as one way of organising structured data), REST
requests rarely use XML; indeed we can think of the web page itself as the
service. A good example is given by a site offering an ISBN service which
returns the details of a book on receipt of an ISBN number. Isbndb.com offers
such a service which requires you to simply use an HTTP GET with parameters
including an access key to allow use of the service (to gain an access key and
use isbndb.com, you need to complete the free registration at h ttp://isbndb.com/ and follow the instructions to gain a veri fi ed
account).
If you enter a URL
as shown below:
but replace the
value for the access key with one you have generated at isnbdb.com, the result
returned in XML format as shown below:
<?xml version=”1.0”
encoding=”UTF-8”?>
<ISBNdb
server_time=”2012-02-22T13:19:46Z”>
<BookList total_results=”1” shown_results=”1”> <BookData book_id=”alan_turing_a04”
isbn=”0045100608” isbn13=”9780045100606”>
5.7 The Cloud Stack
<Title>Alan Turing</Title>
<TitleLong>Alan Turing: the enigma of intelli-gence</
TitleLong>
<AuthorsText>Andrew Hodges</AuthorsText>
<PublisherText publish-er_id=”unwin_paperbacks”>
London : Unwin Paperbacks,
1985.</PublisherText>
</BookData>
</BookList>
</ISBNdb>
A t the end of this chapter, we
will build a small web application using Google App Engine (see below) which
will use the RESTful service provided by isbndb. com to retrieve and process
book information in XML based on ISBN numbers entered by the user.
U nlike SOAP services which require
responses packaged in XML, REST does not require the response to be in any
particular format. A simple illustration is provided by Twitter which will
return the results in various formats (RSS, ATOM and JSON) using the URLs shown
below to send a REST request to Twitter’s search service.
http://search.twitter.com/search.atom?q=turing&count=5
http://search.twitter.com/search.rss?q=turing&count=5 http://search.twitter.com/search.json?q=turing&count=5
The ‘ q ’ following the question mark
is the query term (‘ turing ’), and the count represents the maximum number of
items to return.
REST is based on the HTTP protocol and so is a
‘stateless’ architecture. It is straightforward to make AJAX-based applications
RESTful, and returning the response in JSON format makes the applications
simple to program whilst minimising the bandwidth required for the response.
REST is ubiquitous in cloud services, for example, Amazon.com offers a RESTful
API with HTTP verbs for their S3 storage solution (see below).
Resources are the key element of a RESTful
approach. Resources are identi fi ed by the URLs, and both the state and
functionality of the application are represented using resources. SOAs can be
built using REST services—an approach sometimes referred to as ROA
(REST-oriented architecture). The main advantage of ROA is ease of
implementation, agility of the design and the lightweight approach. The latest
version of WSDL now contains HTTP verbs and is considered an acceptable method of
documenting REST services. There is also an alternative known as WADL (Web
Application Description Language).
5.7 The Cloud Stack
It is useful to view
cloud offerings as a stack, where each layer has distinct features and
capabilities. This is sometimes referred to as the ‘cloud ontology’, and
various models of differing complexity have been proposed. We present a
relatively simple model here as shown in Fig.
5.1 .
Fig. 5.1 The cloud stack 
A t the top of the stack is the
application with the user interface delivered to the user via a web browser. At
the base of the stack lies the hardware of the physical machines actually
running the application, normally located in a large data centre.
Virtualisation technology allows for the abstraction and presentation of the
hardware resources in a discrete and scalable manner to the higher layers.
Actually, a particular layer can be defi n ed according to the level of
abstraction, so we can say that a layer is classi fi ed higher in the cloud
stack if the services offered by that layer can be composed from the services
of the underlying layer. The layer can also be identifi e d according to the
type of user targeted by the layer. The service models build on one another and
defi n e the areas of responsibility amongst vendors, end users and developers.
In this chapter, we are primarily interested
in the three service layers in the middle of the stack, and the following
sections describe each of these with examples.
5.8 Software as a Service (SaaS)
The capability provided to the
consumer is to use the provider’s applications running on a cloud
infrastructure. The applications are accessible from various client devices
through either a thin client interface, such as a web browser (e.g., web-based
email), or a program interface. The consumer does not manage or control the
underlying cloud infrastructure including network, servers, operating systems,
storage, or even individual application capabilities, with the possible
exception of limited user-speci fi c application con fi guration setting. NIST (
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf )
Software as a service (SaaS) is a hosted
application that is available over the Internet via a web browser. SaaS,
sometimes referred to as ‘on-demand software’, is the most complete of the
cloud services. Computing hardware, software and the solution are offered by a
vendor. The cloud application layer is the only layer visible to end users
5.8 Software as a Service (SaaS)
who are the target users for this layer. The end user
does not manage or control the underlying infrastructure. Their only
responsibility is for entering and managing their data based on interactions
with the software. The user interacts directly with the hosted software via the
browser. We should note that SaaS existed well before the concept of cloud
computing emerged; nevertheless, it is now an integral part of the cloud model.
C reating and delivering software
via the SaaS layer is an attractive alternative to the more traditional desktop
applications which must be installed on the users’ machine. With SaaS the
application is deployed in the cloud so the work of testing, maintaining and
upgrading software is greatly simpli fi ed since it can all occur in one place
rather than being rolled out to the desktops of potentially thousands of users.
Con fi guration and testing is reduced in complexity due to centralisation and
the preset restrictions in the deployment environment. Developers can also use
a simpli fi ed strategy when applying upgrades and fi xes. Furthermore, composition, as
discussed above, becomes a straightforward option as soon as the cloud services
are developed. Last but not least, the providers also bene fi t from greater
protection to their intellectual property as the application is not deployed
locally and pirated versions of the
software will be much harder to obtain and distribute. A number of typical characteristics of SaaS
are listed below:
• Software
is available globally over the Internet either free or paid for by subscription
based on customer usage.
• Collaborative
working is easily provided and generally encouraged.
• Automatic
upgrades are handled by the vendor with no required customer input.
• All
users have the same version of the software.
• The
software will automatically scale on demand.
• Distribution
and maintenance costs are signi fi cantly reduced.
T here are a huge variety of SaaS
applications already available, and their number appears to be growing at an
exponential rate. A small selection of prominent examples of SaaS are discussed
below, which hopefully give a good illustration of the SaaS approach. We
recommend that you investigate for yourself by viewing the sites mentioned,
even if only brie fl y.
5.8.1 Salesforce.com
S alesforce.com (h
ttp://www.salesforce.com ) provides
the most widely used customer relationship management (CRM) service used for
managing fi nancial, logistics and staf
fi ng issues related to business system operation. A CRM consists of a set of
business processes together with software that helps manage customerrelated
activities, information and analysis. These activities include:
• Sales
• Marketing
and leads
• Content
analysis
• Accounts
• Forecasting
• Partnerships
and collaboration
• Customer
service
Salesforce.com offers a comprehensive list of
features which can be integrated with applications from Google. The
Salesforce.com SaaS offering is highly customisable and tailored for particular
industries.
5.8.2 Dropbox
Dropbox (
https://www.dropbox.com ) uses
cloud storage to enable users to store and share fi les and folders across the Internet.
Dropbox starts with a free tier with up to 2 GB of storage and supports
multiple user clients across desktop and mobile operating systems, including
versions for Windows, Linux, Mac OS X, Android and iPhone. Any fi les or folders the user places or updates
in the dropbox folder of their local machine are synchronised with the web
version. When a user logs onto another machine with dropbox installed, any
updates are automatically applied. Dropbox also provides a revision history so
that user can revert to older versions of their fi l es. Dropbox currently uses
Amazon’s S3 storage (see below) although it may switch to a different storage
provider in the future.
5.8.3 Google Services
Google (
http://www.google.com/services/ ) offers a wide range of cloud
applications including email, web hosting and, of course, searching. A number
of Google services target business users, such as Google Analytics which
support its targeted advertising business.
Google Docs is somewhat similar to Microsoft
Of fi ce and contains a word processor, spreadsheet, drawing and presentation
program and is free to use up to a certain data size. With Google Docs,
collaborators can share documents and work in real time from different
locations. REST APIs enable users to directly access the Google Data Protocol
(which underlies Google products) and to search Google Docs contents. Once
published to the web, the users are provided with the public URL and the HTML
code required to embed the document in a web page.
G oogle Apps Script gives users a
high level of control over a number of Google products. Google spreadsheets
amongst other products can be accessed and controlled using JavaScript programs
which run directly on Google servers. Actually this means that the Google
spreadsheet is offering more than ‘user-specifi c application con fi guration settings’ and so
strictly speaking is falling outside the NIST defi n ition of SaaS. This is not
an uncommon situation for products which otherwise naturally fall into the SaaS
category.
5.8.4 Prezi
Prezi ( http://prezi.com/
) is a collaborative technology, which is free to use when creating and showing
publicly available non-linear presentations (a ‘prezi’) in the
cloud up to 100 MB in size. Users may simply replace Microsoft
PowerPoint with Prezi for interactive presentations but much more is on offer.
The service also allows users to collaborate in real time for editing or
brainstorming, and Prezi allows the user to upload video fi les and image fi les.
A
fee is payable for greater storage and for additional features such as
the ability to create Prezis of fl ine and to make the Prezis private.
5.9 Platform as a Service (PaaS)
The capability provided to the
consumer is to deploy onto the cloud infrastructure consumer-created or
acquired applications created using programming languages, libraries, services,
and tools supported by the provider. The consumer does not manage or control
the underlying cloud infrastructure including network, servers, operating
systems, or storage, but has control over the deployed applications and
possibly confi g uration settings for the application-hosting environment. NIST
F or many businesses, specifi c rather than general requirements may mean
that a generic SaaS application will not suf fi ce. PaaS creates a managed environment
in the cloud where complex, tailor-made applications can be constructed, tested
and deployed. The provider supplies a hardware platform together with software
environments speci fi cally designed to support cloud application development.
The target users are developers who build, test, deploy and tune applications
on the cloud platform (for the end user the result is still a browser-based
application). Automatic scalability, monitoring and load balancing are provided
so that an increase in demand for a resource such as a web application will not
result in degradation in performance (of course, this may mean an increased
charge is incurred). PaaS vendors will also ensure the availability of
applications, so, for example, should there be any problems on the underlying
hardware, the application will be automatically redeployed to a working
environment, without a detrimental effect on the end users experience. PaaS
should accelerate development and deployment and result in a shorter time to
market when compared with traditional software development using an
organisations’ own data centre.
The developer does not control and has no
responsibility for the underlying cloud infrastructure but does have control
over the deployed applications. The vendor is responsible for all operational
and maintenance aspects of the service and will also provide billing and
metering information. The vendor commonly offers a range of tools and utilities
to support the design, integration, testing, monitoring and deployment of the
application. Typically the tool set will include services to support:
• Collaboration
and team working
• Data
management—sometimes referred to as data as a service (DaaS)
• Authentication
or identi fi cation—sometimes referred to as authentication as a service (AaaS)
• Performance
monitoring and management
• Testing
• Queue
services
• Email
and messaging
• User
interface components
Often the supporting tools are con fi gured as
RESTful services which can readily be composed when building applications. It
is important that developers are able to write applications without needing to
know the details of the underlying technology of the cloud to which they will
be deployed. Standard programming languages are often available, although
normally with some restrictions based on security and scalability concerns.
Additional support for programming and confi g uration tasks is often available
from a developer community speci fi c to the PaaS offering.
5.9.1 Portability
P ortability is a concern to many organisations considering
developing PaaS applications due to the differing requirements of the various
vendors, so, for example, if you build an application using Google App Engine,
there may be signifi c ant work involved if you later decide to move to
Microsoft Azure. Using an established language like Java goes some way to
ameliorate concerns about application portability but does not necessarily
eliminate them as the code will often need to be adjusted to fi t in with the vendor’s particular
requirements and it is quite likely that the program will only work within the
specifi c environment. The way in which
data is persisted is especially prone to quite distinct strategies between the
vendors.
5.9.2 Simple Cloud API
T he Simple Cloud aims to provide a set of common interfaces
for fi l e storage services, document
storage services, simple queue services and infrastructure services, or in their
own words, ‘the Simple Cloud API brings cloud technologies to PHP and the PHP
philosophy to the cloud’ ( http://simplecloud.org/ ). Zend has invited
the open source community and software vendors of all sizes to participate.
IBM, Microsoft, Rackspace, Nirvanix, and GoGrid have already joined the project
as contributors.
5.9.3 Java
M any programming languages are available on the various PaaS
offerings, but Java would currently seem to be the most popular. Although Java
is sometimes referred to somewhat mockingly by developers as ‘the COBOL of the
twenty-fi r st century’, Java is nevertheless a powerful object-oriented
language which is freely available and offers a number of attractive features
such as automatic memory management. Perhaps most importantly there is a huge
pool of developers with Java expertise: currently a reported ten million Java
users. The language has particular appeal to those building client–server web
applications.
Java applications are typically compiled to
form a Java source fi le to Java bytecode. Once in this format the code can
run on any Java Virtual Machine (JVM) regardless of computer architecture. One
of the main goals of the language is summarised in the famous ‘write once, run
anywhere’ adage, meaning that code that runs on one platform (e.g. Windows)
does not need to be edited to run on a different one (e.g. Linux). Another of
the reasons for its appeal is the large number of supporting tools, libraries
and frameworks also freely available and generally open source. These include
the widely used open source and freely available Eclipse Integrated Development
Environment (IDE) which supports rapid development and debugging. The JUnit library
(which can be easily integrated into Eclipse) supports test-driven development
and an agile approach to software production. The Web application ARchive (WAR)
is a standard format used to package and distribute resources used for Java Web
applications which aims to facilitate development, testing and deployment.
A
number of vendors such as Amazon, Google and VMForce have developed ‘plug-ins’
for Eclipse so that their Java PaaS offerings can be developed, tested,
debugged and deployed in an environment already familiar to many developers.
The WAR fi le format is also frequently
used by a number of PaaS vendors, again supporting a familiar style of
development and reducing portability concerns.
5.9.4 Google App Engine
A few years ago before
cloud really hit the headlines, Google famously gave away stickers saying ‘My
other computer is a data center’. Google App Engine (GAE) is a good example of
PaaS and allows users to write and deploy their applications on the same
infrastructure which powers Google’s products. We are going to give a little
more time here to discuss GAE as this is the product we will use at the end of
this chapter when you will develop your own GAE cloud application.
G oogle, of course, has great fl e
xibility in terms of dynamic reallocation of resources required to meet
changing needs and demands. GAE applications are not hosted on a single server
but will run in a distributed environment (see Chap. 4 ) even whilst the developer is
completely unaware of the production con fi guration of the data centres and is
totally free from capacity management, server maintenance and load-balancing
tasks. As the number of requests for the web application increases, GAE
automatically allocates more resources without the developer needing to change
their code.
G AE applications can currently be built
using three different programming languages:
1. The
Python programming language was the fi r st to appear when GAE was initially
made available to developers. Python System Development Kit (SDK) is still
available and widely used.
2. Java
has been added together with integration with Eclipse. Using the Eclipse
plug-in developers can run and test their Java applications locally and deploy
to the cloud with a single click. Google uses the Java Virtual Machine with the
Jetty servlet engine and a standard WAR fi l e structure. Any programming
language which can run on a JVM-based interpreter, such as JRuby, Groovy,
JavaScript (Rhino) and Scala, can also be run on GAE although there may be a
little more work in terms of initial con fi guration. GAE also supports many of
the Java standards and frameworks such as servlets, the Spring Framework and
Apache Struts.
3. Most
recently an ‘experimental’ environment has been added for the new ‘Go’ ( http://golang.org/
) concurrent programming language developed by Google.
Java, Python and Go runtime environments are
provided with APIs (application program interface) to interact with Google’s
runtime environment. Google App Engine provides developers with a simulated
environment to build and test applications locally with any operating system
that supports a suitable version of the Python, Java or Go language
environments. However, there are a number of important restrictions. Developers
cannot write to the fi le system, Java
applications cannot create new threads, and only a subset of the standard Java
classes and packages are available (you can view the list of supported classes
at http://code.google.com/ appengine/docs/java/jrewhitelist.html
).
R estrictions such as these have led to
portability concerns, but recently a number of projects such as AppScale have
been able to run Python, Java and Go GAE applications on EC2 and other cloud
vendors.
The GAE platform also allows developers to
write code and integrate customdesigned applications with other Google
services. GAE also has a number of composable supporting services including:
• Integrated
web services
• Authentication
using Google Accounts
• Scalable
non-relational, schema-less storage using standard Java persistence models
• Fast
in-memory storage using a key-value cache (memcache)
• Task
queues and scheduling
GAE allows applications to be served from the
developers own domain name although by default this will be the name the
developer assigns to their application at appspot.com, for example, http://myApp.appspot.com .
The URL can be shared publicly or selectively.
The code required for an application can be stored on Google code which can be
managed by various open source version control systems such as SVN ( http://subversion.apache.org/
).
A
great advantage, particularly to new developers, is that you can begin
using GAE for free with no credit card required before continuing onto a
graduated fee based on usage (this is one of the reasons we choose a tutorial using
GAE). At the time of writing, each developer can host up to ten free
applications. An application on a free account can use up to 1 GB of storage
and up to fi ve million page views a
month.
T o persist data using GAE, the
application must use the provided app engine services. The data store is a
key-value storage system similar to Amazon SimpleDB and Windows Azure Table
Service (see below). GAE provides high availability by data replication and
synchronisation. The memcache service is used to improve performance by using a
short term, in memory local cache rather than going to the data store. Although
data cannot be read from the fi l e system, it is possible to read fi les that are packages as part of a WAR.
5.9.5 Google Web Toolkit
Google Web Toolkit
(GWT) is a set of open source libraries under Apache 2.0 licence, which allows
web developers to write AJAX web applications entirely in Java. GWT will
compile the Java to optimised JavaScript code which will run in the user’s
browser whilst allowing for mature Java testing strategies, for example, using
JUnit. GWT aims to improve developer productivity and web application
maintenance issues and includes support for asynchronous communication (AJAX),
history management, bookmarking, internationalisation and cross-browser
portability. A GWT application using HTML5 can have separate views for tablets
and mobile phones. There is also a GWT plug-in to the eclipse IDE and a
straightforward procedure for deploying GWT applications to GAE.
5.9.6 Microsoft Azure
The Azure Services
Platform provides a wide array of Windows-based services for developing and
deploying Windows-based applications on the cloud. Whilst Azure is primarily
designed as a PaaS offering where users can build host and scale web applications
on Microsoft’s global data centres, it also includes Infrastructure as a
Service (IaaS) features. The platform consists of compute, storage and fabric
(network resources). Azure provides scaling, load balancing and a number of
support services such as identity and authentication, message encryption,
monitoring and management.
Three main brands
have been developed:
1. The
Windows Azure scalable cloud operating system
2. SQL
Azure which is a cloud-based scalable version of Microsoft SQL Server
3. AppFabric
which is a collection of services supporting the cloud service
A zure applications are typically written
either as a ‘web role’, a ‘worker role’ or both. Web role is used for typical
web applications that are hosted on servers running Microsoft’s IIS web server,
whilst the worker role is used for applications that require compute intensive
operations. The virtual machine role is currently in beta but enables the
developer to deploy a custom-built Windows Server machine image to Windows
Azure potentially saving signi fi cant time on installation-related tasks.
A zure uses a combination of standard web
service technologies REST, HTTP and XML and integrates with Microsoft Visual
Studio. By using the Microsoft Azure SDK, developers can create services that
leverage the .NET Framework. The aim is to allow developers to be able to
choose the language or framework and to integrate public cloud applications
with existing applications. These applications have to be uploaded through the
Microsoft Azure portal in order to be executed on top of WindowsAzure.
Additional services are available, such as work fl ow execution and management, web services
orchestration and access to SQL data stores. Microsoft also offers system
development kits for other languages such as Java, PHP and Ruby.
5.9.7 Force.com
F orce.com from SalesForce.com is an approach to PaaS for
developing CRM systems and applications for social enterprise with built-in
support for social and mobile functionality. The design of its platform and the
runtime environment is based on Java technology. The platform uses a
proprietary programming language and environment called Apex Code, which has a
reputation for simplicity in learning and rapid development and execution.
Users of Force.com can build their own applications which are either
stand-alone or integrated with Salesforce.com.
5.9.8 VMForce
Salesfoce.com and
VMware have partnered to introduce VMForce, one of the fi rst enterprise cloud platforms for Java
developers. VMForce uses Eclipse and allows for development with standard Java
Web technology such as JSPs and servlets. Once developed applications can be
deployed to the cloud with a single click.
5.9.9 Heroku
Heroku is a cloud
platform for instant deployment of Ruby on Rails web applications. Ruby on Rails provides a
complete framework for web application development including a template and
rendering system and methods to persist and retrieve data. Ruby on Rails is
popular amongst developers, has a good track record for rapid web application production
and sits particularly well with agile methods of software production. Heroku
servers are invisibly managed and applications automatically scaled and
balanced.
5.9.10 Cloud Foundry
Cloud Foundry claims to
be the fi rst open source scalable PaaS
offering, using industry-standard frameworks such as Spring (Java) and Scala
and a choice of application infrastructure services. The project was initiated
by VMware but has gained broad industry support. The Cloud Foundry aims to
minimise restrictions on development and deployment in the cloud so that
developers need not be concerned with middleware or vendor infrastructure.
5.10 Infrastructure as a Service (IaaS)
5.10 Infrastructure as a Service (IaaS)
The capability provided to the
consumer is to provision processing, storage, networks, and other fundamental
computing resources where the consumer is able to deploy and run arbitrary
software, which can include operating systems and applications. The consumer
does not manage or control the underlying cloud infrastructure but has control
over operating systems, storage, and deployed applications; and possibly
limited control of select networking components (e.g., host fi rewalls). NIST (
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf )
IaaS
provides developers with on-demand infrastructure resources such as c ompute,
storage and communication as virtualised services in the cloud. The provider
actually manages the entire infrastructure and operates data centres large
enough to provide seemingly unlimited resources. The client is responsible for
all other aspects of deployment which can include the operating system itself,
together with programming languages, web servers and applications. IaaS
normally employs a pay-as-you-go model with vendors typically charging by the
hour.
O nce connected the developers work with
the resources as if they owned them. IaaS has been largely facilitated by the
advances in operating system virtualisation which enables a level of
indirection or abstraction with regard to direct hardware usage (see Chap. 4 ). The virtual machine (VM) is
the most common form for providing computational resources, and users normally
get super-user access to their virtual machines.
Virtualised forms of fundamental resources
such as computing power, storage or network bandwidth are provided and can be
composed in order to construct new cloud software environments or applications.
Virtualisation enables the IaaS provider to control and manage the effi c ient
utilisation of the physical resources and allows users unprecedented fl exibility in con fi guration whilst protecting
the physical infrastructure of the data centre. The IaaS model allows for
existing applications to be directly migrated from an organisation’s servers to
the cloud supplier’s hardware, potentially with minimal or no changes to the
software.
5.10.1 Virtual Appliances
A virtual appliance is
a virtual machine image which is designed to run on a particular virtualised
platform, typically supplied via IaaS. The virtual appliance may have software
such as an operating system, compilers, web server, database system and any required
applications already installed, con fi gured and optimised to solve the
particular problem for which it was designed. In practice many virtual
appliances are often used to run a single application. We can think of an
appliance as an application combined with everything else it needs to actually
run. Often the operating system is a stripped down version of its all-purpose
counterpart, a process sometimes referred to as ‘just-enough-operating-system’
or ‘Juice’. Virtual appliances are normally controlled using a web interface
and can be used as a starting point for building a more complex server.
A
number of online marketplaces have been set up to allow the exchange of
ready-made appliances containing a wide range of operating system and software
combinations. VMWare has the largest selection of appliances (available at http:// www.vmware.com/appliances/directory/) which can be readily deployed to the VMWare
hypervisor. In Chap. 4 tutorial,
we used VMWare to create an Ubuntu virtual machine and then installed Java and
Eclipse to our machine. An alternative approach would have been to download the
appropriate appliance with the software we required already installed.
Amazon Machine Image (AMI) is used to store
copies of particular virtual machines which can run on Amazon’s Elastic Compute
Cloud (EC2) and other IaaS platforms which support the format. Amazon Web
Services offer hundreds of readymade AMIs, which Amazon refers to as ‘pre-built
solutions’ containing common combinations of operating systems and application
software.
5.10.2 Amazon Web Services
A mazon Web Services (AWS) is one of the most successful
cloud-based businesses based on SOA standards and includes a whole family of
related services. The computational
needs of an organisation are of course extremely variable, and some
applications are likely to be compute intensive, whilst others will emphasise
storage. EC2 is renowned for its ability to provide scalable solutions for
diverse requirements.
5.10.3 Amazon Elastic Compute Cloud (EC2)
EC2 provides
virtualised computing power as a service and is considered the central
application of the AWS offering and is often referred to as the ‘model example’
of IaaS. Users can rent computing power on Amazon’s infrastructure and pay with
an hourly rate using EC2. EC2 can be customised in a similar way to a physical
server and can run a variety of operating systems including Linux, Solaris and
Windows server. The users gain administrator privileges and can perform
numerous activities on the server such as starting and stopping, con fi
guration and installation of any database system, programming language or
software package.
T he computing resources available on EC2
instances consist of combinations of computing power and other resources such
as memory. EC2 is offered as a metered service, and an EC2 Compute Unit (CU) is
used as a standard measure of computing power. EC2 users can select from a
number of different instance types to best match particular computing needs.
Amazon Auto Scaling is a set of command line tools that allows scaling EC2
capacity up or down automatically and according to conditions speci fi ed by
the end user.
Standard instances are currently offered with
the following resources as illustrated in Table
5.1 .
There are many other instance types designed
to meet nonstandard requirements. EC2 also offers users the ability to select
from a number of regions and within each
5.10 Infrastructure as a Service (IaaS) Table 5.1 Standard EC2 instance types
|
Instance type
|
Memory
(GB)
|
C ompute
|
Instance storage
(GB)
|
Platform
|
|
Small
|
1.7
|
1 EC2 Compute
Unit (1 virtual core with 1 EC2 Compute Unit)
|
160
|
32-bit
|
|
Large
|
7.5
|
4 EC2 Compute Units (2 virtual cores with 2
EC2 Compute Units each)
|
850
|
64-bit
|
|
Extra large
|
15
|
8 EC2 Compute Units (4 virtual cores with 2
EC2 Compute Units each)
|
1,690
|
64-bit
|
region a number of availability zones where the virtual data
centres on which the EC2 instances will be running.
5.10.4 Amazon Storage Services
Various forms of data
storage are also provided by Amazon ‘as-a-service’. Amazon’s Simple Storage
Service (S3) is a pure object-based storage system which can be accessed via a
REST-based web interface. S3 does not provide the ability to run programs and does
not even provide any fi le system or
indexing, but has the advantages of reliability, availability and scalability.
The user can allocate a block of storage up to 5 TB in size that has a unique
user-assigned identifi e r and can then read or write to the block. S3 is often
cited as a good example of a perfectly focused cloud service which does exactly
one thing in a very narrow way. The service can then be composed with others to
provide the resources required for development.
T he S3 user is charged according to the
volume of data and network bandwidth used when storing and retrieving data. S3
is a popular choice amongst web application developers and can easily be used
to replace existing web hosting infrastructure. The Apache Hadoop fi le system (see
Chap. 4)
can also be hosted on S3 for running MapReduce jobs. At the time of
writing Amazon S3 is reported to store more than 762 billion objects. Dropbox
and Ubuntu One are examples of online backup and synchronisation serviceswhich
use S3 for storage and transfer. Amazons SimpleDB is a key-value NoSQL data
store (see Chap. 6)
and Amazon Relational Database Service provides a MySQL database
instance in the cloud.
A s mentioned above Amazon S3 works over
HTTP. Amazon also offers block storage systems where resources appear to be
physical discs on the EC2 server. This provides improved performance over S3
and is more suitable to the case where multiple disc operations are required.
Instance storage is attached to a particular EC2 instance, can only be accessed
by that particular EC2 and will be lost as soon as the EC2 instance stops. The
Elastic Block Storage (EBS) provides block storage for EC2 but is independent
of a particular instance and can be shared between EC2 instances.
EC2 provides the Elastic Load Balancer which
as the name suggests automatically
balances the load across multiple servers. The load balancer also scales the
number of servers up or down depending on the requirements of the particular
load.
5.10.5 Amazon Elastic Beanstalk
AWS Elastic Beanstalk
is currently in beta but promises to provide ‘an even easier way for you to
quickly deploy and manage applications in the AWS cloud’. The developer need
only upload their applications, and the Elastic Beanstalk will handle the
deployment details, load balancing and scaling. In some ways this makes the
product more like a PaaS offering although in this case the developer can
control the underlying infrastructure if required (Beanstalk will use resources
such as EC2 and S3 when deploying applications). Amazon clearly states that the
aim is to offer the developer the best of both the PaaS and IaaS worlds.
The current release of Elastic Beanstalk is
built for Java developers, and the Elastic beanstalk application should be
packaged using the standard Java WAR structure. An AWS plug-in is available for
Eclipse which also allows you to test and deploy your application to one or
more EC2 instances running the Apache Tomcat web server.
5.10.6 FlexiScale
FlexiScale ( http://www. fl
exiscale.com/ ) offers on-demand hosting and operates
in a somewhat similar manner to Amazon EC2. FlexiScale supports both Windows
and Linux servers and uses the Xen hypervisor.
5.10.7 GoGrid
G oGrid (h ttp://www.gogrid.com/ ) allows
users to utilise a range of premade windows and Linux images in a range of fi xed instance sizes and also offers stacks
on top for applications such as high-volume web applications, e-commerce and
databases.
5.10.8 Eucalyptus
(‘Elastic Utility Computing Architecture
for Linking Your Programs to Useful Systems’)
Eucalyptus ( http://www.eucalyptus.com/ ) is an
open source Linux-based software platform for creating cloud computing IaaS
systems based on computer clusters. The project has an interface that can be
connected to Amazon’s compute and storage cloud system (EC2 and S3), and users
can interact with Eucalyptus clouds using the same tools used on Amazon.
Eucalyptus supports multiple hypervisor technologies within the same cloud and
provides a range of cloud administration
tools for system management and accounting. Eucalyptus is
currently supported by most Linux distributions. Ubuntu enterprise cloud is a
new initiative to make it easier to provide, deploy, con fi gure and use cloud
infrastructures based on Eucalyptus.
5.10.9 Rackspace
Rackspace (
http://www.rackspace.com/ ) offers
an IaaS solution fi x ed size instances in the cloud. Rackspace cloud servers
offer a range of Linux-based premade images, and users can request different
sized images.
5.11 Chapter Summary
I n this chapter, we have looked at how using web services and
service-oriented architectures has contributed to the rise of cloud computing.
We have reviewed the SPI model and examined how it can be useful to view cloud
computing as consisting of various layers which are most easily identifi e d by
considering the particular responsibilities of vendors, developers and end
users. We should note that although the SPI model is extremely useful in
categorising and understanding cloud-based systems, the layers are rather
blurred as, for example, some SaaS offerings can be programmed and some PaaS
products offer the developers the option of direct control of the underlying
hardware.
5.11.1 End of Chapter Exercises
In this tutorial we are
going to work with Google’s cloud offering to create a simple web application.
This choice is due partly to the availability of Google’s free tier as we
appreciate that many students of cloud computing may be put off by having to
purchase cloud resources or supply credit card details.
Create a Google App Engine account (free) by
clicking the Sign up link under Getting Started on http://code.google.com/appengine/ and follow the instructions. No credit card is
required for the free tier, but you will need to be able to receive an SMS text
message.
There are currently three programming
languages which you can use on GAE, but we are going to use Java. One advantage
of this is that we can use the GAE plugin to the freely available eclipse IDE
which we installed to our VM previously.
5.11.2 Task 1: Prepare Eclipse and Install GAE Plug-In
We need to install some
add-ons to eclipse. Whilst the install takes place, you will need to agree to
run unsigned content and agree to licence conditions. You will also need to
restart eclipse when requested (if at any point you hit a problem, you should
restart eclipse and try again).
1. Open
your VM and go to the Ubuntu software centre. Type in eclipse and install.
2. Open
eclipse and select the default Workspace location and then go to Help/ Install
New Software. Click on the ‘Available Software Sites’ link and enable the entry
whose location is http://download.eclipse.org/releases/helios
3. After
the list populates (this may take some time depending on your Internet
connection speed), in the text box below ‘work with’, type ‘xml’ to fi lter the entries and then select ‘Eclipse
XML Editors and Tools’ which should be available in the ‘programming
languages’.
4. Select
‘next’ twice, agree the licence and ‘ fi nish’ (again this may take some time
to complete).
5. Restart
eclipse when prompted.
6. You
are now ready to install the plug-in for Google App Engine. Again, open eclipse
and go to Help/Install New Software and enter this URL: http://dl.google. com/eclipse/plugin/3.7
7. Select
the Google plug-in for eclipse and the SDKs as shown below.
8. Select
‘next’ twice, agree the licence and ‘ fi nish’ (again this may take some time
to complete) (Fig. 5.2 ).
5.11.3 Task 2: Create the First Web Application
1. Click
on the Google icon and select ‘new web application project’ (Fig. 5.3 ). We are
not going to use GWT (Google Web Toolkit), so make sure this option is
unchecked. Fill in a project name and a package name and make sure the ‘Use
Google App Engine’ is selected (Fig. 5.4 ).
2. You
can then select ‘Finish’. This will create a new eclipse project with a
structure similar to that shown below (once you have expanded the elements)
(Fig. 5.5)
. As you can see GAE uses a standard Java web archive folder (WAR).
3. Open
the sample Java servlet code which has been generated for you (again you may
need to expand by clicking the ‘+’ sign) (Fig.
5.6 ).
You can run the project locally by right clicking the
project node in the package explorer (GAE_isbn) and then selecting ‘run as’ and
then selecting ‘web application’. If
successful you will see the message appear in the console window:
INFO: The server is
running at
http://localhost:8888/_ah/admin
4. Open
a browser session and open http://localhost:8888/ . Hopefully
you will see a page similar to that shown below in Fig. 5.7 .
I f you click on the servlet, you should
see the ‘hello world’ message returned by the sample servlet. Even if you are
not familiar with Java, take a look at the code for the servlet and see if you
can change the text of the message sent from the server.
W e are now going to try and create a simple ISBN lookup
service on GAE. This will be a limited application and really is only intended
to give you a fl a vour of the type of thing you can do with app engine.
Fig. 5.2 GAE Install Dialogue
5.11.4 Task 3: ISBN App
Our GAE application is
going to prompt the user for an ISBN number. Once the user submits the number,
we will perform some very basic validation and then return the title, author
and publisher of the book.
We
are going to achieve this by composition. We will use the REST-based service
offered by http://isbndb.com . The service is
free, although you will need to complete the simple sign up procedure to obtain
an access key. Once you have the key, you can test the service by pasting a URL
similar to the following into your browser but replacing the value for key with
your own key:
Fig. 5.3 New Web Application dialogue
If all goes well you
should see an XML fi le similar to the
following:
<?xml version=”1.0” encoding=”UTF-8”?>
<ISBNdb
server_time=”2012-02-22T13:19:46Z”>
<BookList total_results=”1”
shown_results=”1”> <BookData
book_id=”alan_turing_a04” isbn=”0045100608” isbn13=”9780045100606”> <Title>Alan Turing</Title>
<TitleLong>Alan Turing: the enigma of intelli-gence</
TitleLong>
<AuthorsText>Andrew Hodges</AuthorsText>
<PublisherText
publish-er_id=”unwin_paperbacks”>
London : Unwin Paperbacks,
1985.</PublisherText>
</BookData>
</BookList>
</ISBNdb>
Now return to eclipse and the GAE project you
created earlier where we have three updates to make:
1. We need to update the index.html fi
l e (in the WAR folder) which is the default starting place for the web
application. Replace the existing contents with the following:
<html>
<head>
<script type=”text/javascript”> var xmlhttp;
Fig. 5.4 New Web Application Project dialogue
function lookupISBN(isbn) { xmlhttp = null; if (window.XMLHttpRequest) { xmlhttp = new XMLHttpRequest();
}
else if (window.ActiveXObject){//code for IE6,IE5 xmlhttp = new
ActiveXObject(“Microsoft.XMLHTTP”);
}
if (xmlhttp != null) { xmlhttp.onreadystatechange =
state_Change; var url =
“/isbnservice?isbn=” + isbn;
xmlhttp.open(“GET”, url, true);
xmlhttp.send(null);
Fig. 5.5 Eclipse
Package
Explorer
Fig. 5.6 Servlet code
} else { alert(“Your browser
does not support XMLHTTP.”);
} }
function state_Change() { if (xmlhttp.readyState == 4) {// 4 =
“loaded” if (xmlhttp.status == 200)
{// 200 = “OK” docment.getElementById(
‘ISBNServiceResponse’).innerHTML =
xmlhttp.responseText;
} else {
Fig. 5.7 Hello App Engine 
alert(“Problem looking up ISBN Service :” + xmlhttp.statusText);
}
}
}
</script>
</head>
< title>ISBN Lookup</title>
< body>
<h2>ISBN Lookup</h2>
<hr />
<h4>
(Powered by ISBNdb
<a href=”http://isbndb.com/”>ISBN
Service</a>)
</h4>
<hr />
<p>
<b>Lookup ISBN:</b><input type=”text” id=”isbn”></input>
</p>
<p>
<b>Book details :</b> <br />
<span id=”ISBNServiceResponse”></span>
</p>
<button
onclick=”lookupISBN(isbn.value)”>
Lookup ISBN</button>
</body>
</html>
The above includes html and JavaScript. The
JavaScript function called lookupISBN
will use AJAX to update the page with the information returned by
the isbnservice .
2. We now need to update the servlet code.
The code below might look complex but in fact is simply using some standard
libraries to make a call to the ISBNdb URL, navigate the XML returned and write
out the data from the selected elements. Update the servlet code with the
following: package
com.cloudtest.isbn; import
java.io.BufferedReader; import
java.io.IOException; import
java.io.InputStreamReader; import java.io.StringReader; import java.net.URL; import javax.servlet.ServletException; import javax.servlet.http.HttpServlet; import
javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse; import javax.xml.parsers.DocumentBuilder; import
javax.xml.parsers.DocumentBuilderFactory;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.Node; import
org.w3c.dom.NodeList; import
org.xml.sax.InputSource; @SuppressWarnings(“serial”) public class ISBNService extends HttpServlet
{ public void doGet(HttpServletRequest
req, HttpServletResponse resp) throws IOException { String strCallResult = “”; resp.setContentType(“text/plain”); try {
// Extract out the user entered ISBN number
String strISBN =
req.getParameter(“isbn”); // Do basic
validation - could be much more thorough
if (strISBN == null) throw new
Exception(“ISBN fie ld cannot be
empty.”);
// Trim
strISBN = strISBN.trim(); if
(strISBN.length() == 0)
throw new Exception(“ISBN fie ld
empty.”);
String strISBNServiceCall =
“http://isbndb.com/api/books.xml?access_key=123&index1=
isbn&value1=”; strISBNServiceCall +=
strISBN;
URL url = new URL(strISBNServiceCall); BufferedReader reader = new BufferedReader(new
InputStreamReader( url.openStream()));
StringBuffer response = new
StringBuffer(); String line; while ((line = reader.readLine()) != null)
{ response.append(line);
}
reader.close();
strCallResult = response.toString(); DocumentBuilderFactory builderFactory =
DocumentBuilderFactory.newInstance();
DocumentBuilder builder =
builderFactory.newDocumentBuilder();
Document doc = builder.parse(new
InputSource(new StringReader(
strCallResult.toString())));
doc.getDocumentElement().normalize();
NodeList nList =
doc.getElementsByTagName(“BookData”);
if (nList.getLength() < 1)
throw new Exception(“no books: check the ISBN.”); for (int temp = 0; temp <
nList.getLength(); temp++)
{
Node nNode = nList.item(temp);
if (nNode.getNodeType() == Node.ELEMENT_NODE)
{ Element eElement = (Element)
nNode; resp.getWriter().println( “<br />Title : “+ getTagValue(“TitleLong”, eElement)); resp.getWriter().println(“<br />Author
: “ + getTagValue(“AuthorsText”,
eElement)); resp.getWriter().println(“<br
/>Publisher : “+
getTagValue(“PublisherText”, eElement));
}
resp.getWriter().println(strCallResult);
}
} catch (Exception ex) { strCallResult = “Error: “ +
ex.getMessage();
resp.getWriter().println(strCallResult);
}
}
@Override public void doPost(HttpServletRequest req,
Fig. 5.8 ISBN Lookup
HttpServletResponse resp) throws ServletException, IOException { doGet(req, resp);
}
private static String getTagValue(String sTag,
Element eElement) { NodeList nlList = eEle-ment.getElementsByTagName(sTag).item(0).
getChildNodes();
Node nValue = (Node) nlList.item(0);
return nValue.getNodeValue();
}
}
3. Change
the package statement (if required) at the top so that it refl e cts the name
of the package you created when setting up your GAE project.
4. Update
the value of the string for the strISBNServiceCall so that your access code for
isbndb.com is used.
References
5. Finally,
you need to update the web.xml (also in the WAR folder) fi le such that it includes the following:
<servlet>
<servlet-name>ISBNService</servlet-name>
<servlet-class>com.cloudtest.isbn.ISBNService</
servlet-class> </servlet>
<servlet-mapping>
<servlet-name>ISBNService</servlet-name>
<url-pattern>/isbnservice</url-pattern>
</servlet-mapping>
Y ou need to ensure that the
packages specifi e d for the servlet-class matches your own package. Once you
have saved these updates, you can test the application locally. Run the web
application and open in the browser. Hopefully now the page will look something
like that shown below in Fig. 5.8 .
Test the application
by entering ISBN numbers (e.g. 1884133320).
Finally, we are ready to deploy the
application. Click on the Google icon in eclipse and select the deploy option.
You will need to give a name for your application that has not been used before.
Once deployed you should be able to open and test the application in your
browser using a URL such as http://lazaruscloud.appspot. com/isbn.html
References
Chu-Carrol
C.: Code in the Cloud: Programming Google App Engine. Pragmatic Programmers
(2011)
Marcel, C.B., Marcel, K., Mimis, J., Tai, S.: Cloud
Computing: Web-based Dynamic IT Services.
Springer, Heidelberg (2011)
0 komentar:
Posting Komentar