2
1 Introduction
In the IT world recently emerged a new buzzword called Cloud Computing,
characterized by several definitions and different viewpoints. The birth of the term
Cloud Computing is related to the common stereotype representing the data transport
across carrier backbones to an endpoint on the other side of the cloud.
Figure 1 - Classical network diagram
The first signs of these concepts date back to 1961, when John McCarthy
1
, a well-
known authority in artificial intelligence, expressed his vision of the future, where
computing power and even specific applications might be sold through a utility-type
business model. This vision was revived in 1969 by J.C.R. Licklider
2
, with his vision of a
wide computer network where people would be globally connected.
By the mid-70s this idea faded away when it became clear that the IT-related
technologies of the time were unable to sustain this futuristic computing model.
However, with the new millennium, the concept has once again been revived and the
term “Cloud Computing” began to emerge in technology circles (Rittinghouse and
Ransome 2009, Hoefer and Karagiannis 2010, Wilson 2009, IDC 2011).
1
John McCarthy (born September 4, 1927, in Boston, Massachusetts), is an American computer scientist
who received the Turing Award in 1971 for his major contributions to the field of Artificial Intelligence
(AI). He was responsible for the coining of the term "Artificial Intelligence" in his 1955 proposal for the
1956 Dartmouth Conference and is the inventor of the Lisp programming language.
2
Joseph Carl Robnett Licklider (March 11, 1915 – June 26, 1990), known simply as J.C.R. or "Lick" was an
American computer scientist, considered one of the most important figures in computer science and
general computing history. He is particularly remembered as an Internet pioneer, with an early vision of
a world-wide computer network long before it was built.
3
The technological evolution with the increased computer systems use for daily
activities are transforming the ICT resources into mass services, trading and providing
them as “traditional” resources like water, electricity, gas and telephony, always
according to the pay per use paradigm (Buyya et al. 2009).
Although it is more complex and advanced than traditional resources, IT has all the
features of an infrastructure suitable for rapid commoditization as shown in Figure 2
(Carr 2003).
In fact, IT can be considered as a transport mechanism for digital information just as
railroads carry goods and power grids provide electricity, and it adds value when data
are shared.
Figure 2 - Infrastructures evolution and commodization (Carr 2003)
This new technological approach is also known as utility computing. This term has been
4
a term coined in the 70s by Leonard Kleinrock
3
, one of the main contributors of
ARPANET
4
, who laid the foundation for the development of Internet: “As of now,
computer networks are still in their infancy, but as they grow up and become
sophisticated, 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” (Norman 2010).
The IT commoditization, that became essential to any business, was accelerated by the
internet spread. Let us consider this commoditization progress through the evolution
stages towards Cloud Computing, as shown in Figure 3.
Figure 3 - ISP Evolution - Forrester Research Inc.
In a first stage, ISPs
5
were providing internet connectivity by physical access points. The
3
Leonard Kleinrock (born 1934) is an American engineer and computer scientist who played an
important role in the development of the ARPANET , the precursor to the Internet, at UCLA.
4
The Advanced Research Projects Agency Network (ARPANET), was the world's first operational packet
switching network and the core network of a set that came to compose the global Internet. The network
was created by a small research team at the Massachusetts Institute of Technology and the Defense
Advanced Research Projects Agency (DARPA) of the United States Department of Defense.
5
An Internet service provider (ISP) is a company that provides access to the Internet. Access ISPs
connect customers to the Internet using copper, wireless or fiber connections. Hosting ISPs lease server
space for smaller businesses and host other people servers (colocation). Transit ISP (Internet Services
5
next stage offered hosting capabilities. Furthermore, the request of servers for private
hosting with customized operating systems led to rack services, that ware able to
accommodate a whole Line of Business (LoB)
6
.
The next stage saw the spread of Application Service Provider (ASP)
7
, form which SaaS
(Software as a Service) stemmed. The ASP can be considered as a ISP level 4 service.
Eventually, the next and current stage is characterized by a whole portfolio of Cloud
Computing services, that includes on demand infrastructure (IaaS) and on demand
platform (PaaS). The Cloud Computing services spread is driven by the business request
to avoid the non-core capabilities costs, such as data center, security or deployment
management. However, as hosting and housing, non-cloud solutions are still very
popular in small and medium business.
A cloud services main driver is the elasticity of infrastructure (meaning you use what
you need). Typically, a traditional data center, in order to be safe to service users, will
conservatively oversize the resources, resulting then in an additional idle resources
costs. Conversely, a Cloud Computing solution will allocate resources on demand,
reflecting the real time requirements.
The cost model is also different. Users pay for the resources allocated within a specific
period. Therefore, the fixed costs of hardware and software (Capex
8
) become
Providers) provide large tubes to connect hosting ISPs to access ISPs.
6
Line of business (LOB) is a general term which often refers to a set of one or more highly related
products which service a particular customer transaction or business need.
7
An application service provider (ASP) is a business that provides computer-based services to customers
over a network. Software offered using an ASP model is also sometimes called On-demand software or
software as a service (SaaS). The most limited sense of this business is that of providing access to a
particular application program (such as customer relationship management) using a standard protocol
such as HTTP .
8
Capital expenditures (Capex) are expenditures creating future benefits. A capital expenditure is
incurred when a business spends money either to buy fixed assets or to add to the value of an existing
fixed asset with a useful life that extends beyond the taxable year. Capex are used by a company to
acquire or upgrade physical assets such as equipment, property, or industrial buildings. In accounting, a
capital expenditure is added to an asset account ("capitalized"), thus increasing the asset's basis (the
cost or value of an asset as adjusted for tax purposes). Capex is commonly found on the cash flow
statement as "Investment in Plant Property and Equipment" or something similar in the Investing
6
operational costs (Opex
9
) which reflect the real resources utilization.
Cloud Computing provides not only flexibility but also scalability. This means that if in
In a traditional environment, the request for a new additional resource in a peak time
may result into a service disruption, in a cloud environment the new additional
resource is dynamically allocated. Of course, as load drops to a “normal” situation, new
additional resources are reallocated elsewhere.
In general, thanks to the scope and the scale economies cloud providers can be more
efficient. Let us shortly consider why.
First of all, a cloud service provider will focus only on technical issues about IT
infrastructure size and costs. By contrast, a bank typical Chief Executive Officer (CIO)
10
has to think about a very wide range of organizational issues, that include support to
new bank products and user requests on refitting applications. Thus, the CIO can
devote just a fraction of his mind for both IT service and IT infrastructure cost and size.
So, the CIO’s IT infrastructure will be less smart and less efficient than the cloud
provider’s one.
Second, the vast majority of applications follow “Gaussian like” pattern as it is shown
by Figure 4 a CRM transactions in a sample of Utility companies (Bracchi, Francalanci
and Motta 2010).
In a “on premises” where a company sustains the investment costs resources, systems
should service the peak (here on hour 3 = 10 am). Therefore the processing capacity
will be almost idle in 22 hours and totally idle during 17 hours.
subsection.
9
An operating expense, operating expenditure, operational expense, operational expenditure or OPEX is
an ongoing cost for running a product, business, or system . Its counterpart, a capital expenditure, is the
cost of developing or providing non-consumable parts for the product or system. For example, the
purchase of a photocopier involves CAPEX, and the annual paper, toner, power and maintenance costs
represents Opex. For larger systems like businesses, OPEX may also include the cost of workers and
facility expenses such as rent and utilities.
10
A chief executive officer (CEO, American English), managing director (MD, British English),or chief
executive is the highest-ranking corporate officer (executive) or administrator in charge of total
management of an organization. An individual appointed as a CEO of a corporation, company,
organization, or agency typically reports to the board of directors.
7
It is clear that a cloud outsourcers serving customers with different time zones can
have a cost that is a fraction, no more than 1/3, of the “on premise” infrastructure,
and, at the same time, can provide a much higher space capacity. Finally, a very large
outsourcer can enjoy the mass purchasing advantages, with lower prices. So the cloud
outsourcer business comes from selling the same infrastructure to different customers
and from the mass purchasing advantage.
Figure 4 - Hourly distribution of calls
In addition, the resources concentration and the reallocation advantage fosters a green
computing approach, with also a higher energy efficiency.
All these services shall ensure high levels of availability. For this, Service Level
Agreement (SLA) is a key element for cloud governance (e.g. providers usually promise
a 99% of service uptime).
The technology used by the cloud providers is based on capabilities that enhances
virtual machines (VMs) performances. That assures the resources “isolation” offered by
the physical hardware and moreover some possible issues prevent performance and
security issues among users (Kesavan et al. 2008).
Virtualization technologies also allow dynamic resources allocation that maximizes
capacity use trough assigning one same resource to multiple users.
However, some factors may limit the opportunity of using cloud services:
Governance: companies shall balance the dependency on cloud with a
sophisticated governance structure, that will include SLA and Service Level
8
Management systems.
Cloud providers must guarantee data privacy: some Governments require that
sensitive data are protected and, in European Union, are stored within the
borders of a member State.
In Government, despite the excellent performances and uptime, the most
popular solution deploys in-house and runs all operations, without relying on
outsourcers.
1.1 Thesis Structure
The main purpose of this section is to provide the reader a high-level overview of this
thesis structure. The topics progression within the chapters is well defined to attempt
to properly guide the reader in the complex and wide “world of Cloud Computing”.
Chapter 1: After a brief introduction about Cloud Computing this section aims
to specify the analysis boundaries, the approach methodology and the problem
formulation about the cloud environment difficulties in measuring the service
levels;
Chapter 2: This section is entirely reserved to an accurate and systematic
literature review aiming on one hand to provide an effective Cloud Computing
overview and on the other one to highlight the research trends;
Chapter 3: The chapter proposes a statistical analysis based on the results of an
online survey intended only to companies. This inquiry is intended to deepen
the business approach to Cloud Computing, pointing out the deployment,
delivery and methodologies typologies used for measuring service levels;
Chapter 4: This section examines the main framework of the Web applications
for service level management by studying its application to Cloud Computing
environments;
Chapter 5: In this chapter, based on the results obtained from the former
frameworks analysis, is presented a prototype for service level management in
a Cloud Computing private environment;
Chapter 6: in this chapter is proposed a performance test and the application of
9
the prototyped framework on a Cloud Computing environment realized in the
department of Computer Engineering and Systems Science of University of
Pavia;
Chapter 7: in this last section are presented general comments on the thesis
and the results obtained from it. There are also new ideas and improvements
for future studies.
1.2 Problem formulation and purpose
As described by the summary above, the Cloud Computing appears to be a
revolutionary IT approach for the companies. The possibility of a dynamic resources
calibration and a payment based on the effective use has highlighted the strategic
importance of using Cloud Computing systems. These characteristics have driven
companies to adopt these systems and transform Cloud Computing in a new
technological buzzword. Consequently, this race for innovation has led providers and
vendors to supply products and services for Cloud Computing.
The frenzy created around this topic, however, has generated a strong disinformation.
In fact, as evidenced by Simon Wardley in a brilliant presentation at OSCON ‘09
11
,
through a Google search, you can find 67 definitions about Cloud Computing.
Even the academic world has shown great fervor around this issue by producing an
ever increasing number of scientific publications.
For this reason, a purpose of this thesis is to clearly outline what really the Cloud
Computing really is.
To avoid being influenced by the vendors perspective, which define the cloud that is
based on their products characteristics, this part of our research is based primarily on
the study of scientific publications.
The main thesis purpose, however, was determined by a six months study of related to
the tests of some open source Cloud Computing environments. Through these systems
use we have immediately observed that environments flexibility and scalability could
11
The O'Reilly Open Source Convention (OSCON) is an annual convention for the discussion of free and
open source software. It is organized by the publisher O'Reilly Media and is held each summer in the
United States.
10
greatly complicate the service levels control. Furthermore, the different ways of
delivery and deployment present characteristics and problems different from each
other. Starting from these assumptions, our research has been focused on the study of
the main web services frameworks with the final aim of being able to define and
prototype service level management framework in Cloud Computing environments.
1.3 Scope
The research question presented above is too much broad to be approached without
defining its precise limits. For this reason this paragraph aims to further clarify the
boundaries and the scope of this thesis research.
The causes of complex service level management in a Cloud Computing environment
are both technical and non- technical. This thesis however focuses just on the technical
issues examining the approach of the existing frameworks for measuring
performances.
In particular this topic is treated differently than the current research trends. In fact,
our focus of this thesis is on the services measurement provided to the end user,
leaving aside (or considering less important) the performance measurement of the
infrastructure. This approach allows, in our view, greater accuracy in the definition of
service levels and the subsequent drawing of the SLA.
Ricevi informazioni sui nostri servizi, sulle offerte e non perdere news e consigli su università e lavoro.
