Cloud Service Mediation through Brokerage Service

Claudio Giovanoli

Institute of Information Systems, University of Applied Arts and Sciences Northwestern Switzerland, Olten, Switzerland

1 RESEARCH PROBLEM

The advent of cloud computing has introduced an

entirely new way of thinking about IT services and

their role within an organization. It is in many ways

consistent with the general shift in focus to the

outsourcing trend that has become common practice

in modern businesses. Outsourcing lets businesses

concentrate on their core competencies while

simultaneously taking advantage of the core

competencies of their service providers – whether in

accounting services, manufacturing and any number

of other business activities. Cloud computing is the

manifestation of this trend applied to the day to day

operations traditionally delivered by internal IT

departments. As cloud computing services mature, an

increasing number of businesses will choose to move

their IT applications to the cloud (Cambridge

Technology 2011) to take advantage of the benefits it

can deliver.

Common IT services like file storage, e-mail,

databases, web sites, and many others can be pushed

into the cloud – leaving businesses to shift greater

focus and effort to delivering their own products and

services to their customers. The National Institute of

Standards and Technology (Mell and Grance 2011)

identifies the following characteristics associated

with cloud services:

1. On-demand self-service – a consumer is able to

provision computing power, network storage and

other capabilities automatically and as needed

without interaction of a human.

2. Broad Network Access – all services are provided

over the network and are accessible using

standard thin and thick clients such as traditional

workstations, tablet PC’s and even smart phones.

3. Resource Pooling – the service provider’s

computing resources are pooled together in what

is called a multi-tenant model. These resources are

dynamically assigned and reassigned to different

tenants as needed (according to demand).

Examples of such resources include memory,

bandwidth, processing power and storage.

4. Rapid Elasticity – the ability to scale up or down

on demand and often automatically gives the

customer valuable flexibility to meet the needs of

the business.

5. Measure Service – the customer pays only for

what is used. The consumed re-sources are

monitored, controlled and reported upon. This

provides transparency to both the customer and

the service provider. The IT services are

essentially billed to the customer like a traditional

utility company.

Having understood the benefits these characteristics

could potentially deliver, companies are taking notice

and looking to capitalize on this new paradigm of IT

services. To find the right service with the right

quantity and quality so called cloud service brokers

are playing an intermediary role between the

providers and customers.

Companies and organizations planning to use

cloud services are facing today a huge number of

different possible cloud solutions. Because of the

sheer number of possibilities it is hard to orient

oneself and find the optimal solution and offering.

Cloud Brokering companies are offering the

provision of an optimal service to its customers. This

time consuming process stands in opposite to the

cloud paradigms of fast provision and on-demand

self-service of a service. Thus, an automated

brokerage approach could leverage advantages of

cloud computing and increase companies’ agility.

If a company would like to examine if it is able to

use cloud services, different aspects like process

requirements, data security etc., have to be

considered. According to (Xin and Datta 2010) no

best practices or other manageable methods are yet

available to support customers in investigating these

aspects.

The analysis of own requirements can be a very

time intensive task. But also the selection of an

appropriate provider and service are a complex

challenge, especially for small and medium

enterprises with no or low IT expertise.

This research projects picks up these issues with

the goal to simplify the high complexity. Thus the

needed effort for an integrated Cloud Service

Mediation (from requirements analysis, service

Giovanoli, C.

Cloud Service Mediation through Brokerage Service.

In Doctoral Consortium (DCCLOSER 2016), pages 9-13

classification and the appropriate mapping) should be

diminished.

Therefore the thesis statement is defined as

follow: There exists a method to mediate cloud

services based on user requirements and cloud service

quality properties. This method is easy to use for

customers and can deliver automatic decision

support.

2 OUTLINE OF OBJECTIVES

This research project foresees three main objectives,

which can be identified as a part of such an integrated

Mediation Broker approach. The main goal is to

support SME’s using cloud services and to simplify

the due diligence process. From a scientific point of

view, this means that three main aspects have to be

investigated:

 How can customers easily define, collect and

describe their business and functional

requirements. Thus, a requirements evaluation

method has to be developed. Main properties of

such a method should be:

a. Simplicity to understand for business;

defining a language which makes Cloud

necessities understandable for people with a

low technical affinity

b. Easy to use; in this case the project is intended

to find an appropriate form of self-services for

the needs assessment and service selection. A

high degree of self-explanation of the

operation of the system is aimed. The usage of

the self-service should be as simple as possible

to gain access to the required information.

c. Machine readable; whereas results are derived

from the customer analysis, they will be used

for input for the mapping component, this

must also be described and understandable for

the mapping system

 Cloud services must be identified, described and

have to be classified qualitative and quantitative

so that a statement of compliance with the claimed

customer requirements can be made. Areas for

classification criteria could be for example:

a. Contractual aspects

b. Operational processes

c. Properties of Service Level Agreements

d. Training and Support

e. Interfaces and Interoperability

 How can data from I and II be automatically

mapped to derive an optimal match of cloud

service quality and customer requirements.

The following figure should give a short impression

on the entire system, beginning at the requirements

assessment, to the service classification and finally

the mapping of the services to the requirements.

Figure 1: Cloud Mediation.

The three main objectives can be summarized as

follow:

1. Defining and developing a method to collect

customer’s business and functional requirements

with regards to a cloud service. The method has to

be easy to understand and usable for users without

IT background.

2. Defining a classification and weighting scheme

for cloud services with regards to service quality.

3. Evaluating an appropriate technology for

mapping data retrieving from objective 1 and 2.

Based on the chosen technology a proper system

has to be developed to mediate cloud services

easily to the customer.

3 STATE OF THE ART

The functionality and the role of a so-called Cloud

Brokers are discussed at international level for some

time. According to the National Institute for

Standards and Technology (NIST), there exist the

following three forms of brokering services in the

cloud environment (NIST 2013):

 Service-Intermediation: A cloud broker adds a

given service by improving the ability of some

specific services for cloud consumers. Such

added-values can be a management access,

personnel management, performance reporting,

enhanced security, etc be the actual service.

 Service Aggregation: A cloud broker combines

and integrates multiple services into one or more

new services. The broker provides data

DCCLOSER 2016 - Doctoral Consortium on Cloud Computing and Services Science

integration and ensures the secure data movement

between the cloud consumer and multiple cloud

providers.

 Arbitrage-Service: Service arbitrage means that a

broker can choose from different services from

various providers. A cloud broker, for example,

can choose a suitable selection (Cloud Service

Selection) out of various offerings based on

different criteria.

Whereas on the field of intermediation and

aggregation (e.g. on the topic of interoperability)

already a lot of scientific effort takes place

(Sundareswaran et al., 2012) and various solutions are

available (Sun et al., 2013; Gartner, 2013), there is

only limited knowledge and on a high level of

abstraction in the field of arbitration (Kalepu et al.,

2003; Mondal et al., 2010; Buyya et al., 2012). At this

level first scientific efforts were taken of Buyya et al.

(Buyya et al, 2012; Garg et al., 2011). Buyya looks

such a broker as a central role for a market-oriented

approach of cloud services (Garg et al., 2011).

Silas et al., (Silas et al., 2012) propose a service

middleware for efficient service selection. By using

the ELECTRE methodology of the selection process

is approached as a multi on criteria decision problem.

Many criteria, such as response time, service costs,

responsiveness, trust, scalability, performance,

flexibility thereby influence the selection process.

Deng et al. (2011) used for service selection also a

multi-criteria decision-making process, which is

based on the Fuzzy-AHP (Saaty, 1986) and TOPSIS

method.

Furthermore, (Deng-Neng et al., 2011) is using

trust as the sole criterion for the selection of service

providers. Garg et al., (2012) focus their work on the

indexing and classification of the provider, which is

required for a service switching. A so called cloud

service index with parameters such as service type,

price unit, security provider is used. The whole

concept is based on the provision of data by the

respective provider and thus does not provide

independence from the service provider.

The existing activities focus either on the side of

the classification of services or the decision support

methods for the customer. Grag et al., (2012) and

Hussain (2011) also complain about that today's

cloud based service evaluation and selection methods

and in particular functional requirements. For a

successful service selection and service mapping but

also the business needs, as well as non-functional

requirements such as potential regulators,

Performance, Support etc. must be considered and the

service characteristics are com-pared.

The field of Business / IT Alignment (BITA)

seems to be suitable. BITA aims to align, adapt and

integrate business strategy, IT strategy, business

infrastructure and IT infrastructure with each other

(Henderson et al., 1993; Papp, 2001). In this field

there exist various approaches (plugIT, 2012, Wolf et

al., 2011) on how BITA has to be applied and used.

However, these methods are not suitable for the

selection and placement of cloud services, because of

the time consuming processes. Such as experiences

from the plugIT project (2012) have shown in other

areas, the efforts of such frameworks are not practical

for the users.

With regards to the mapping mechanism, the

complexity of such structures have to imply that

corresponding comparison algorithms have to have a

high tolerance for structural deviations without

harming the semantic content of the compared

entities. One way to take this into account are

approximate comparison methods based on

similarity-based reasoning. Among them there exists

a wide range of techniques including Case Based

Reasoning techniques such as example-based

reasoning, instance-based reasoning, memory-based

reasoning, or analogical reasoning (Wolf et al., 2011;

Aamodt, 1994).

4 METHODOLOGY

At this stage of the research project, pragmatism is the

right stance, because it offers the greatest flexibility

within the research project. The disruptive nature of

cloud computing and the resulting complexity within

the vendor landscape might still hold sur-prises along

the way. As research progresses and knowledge for

the development of a mediation broker will mature, it

is crucial in this research to adapt quickly.

As discussed in the previous chapters extensively,

cloud computing evolves at a great pace and the focus

in this market changes constantly. As a consequence

it is getting more difficult to keep the pace and stay

on top of things in the shifting cloud environment.

This rapid evolvement in its early stage makes it even

harder to develop a robust theory to test and validate

(i.e. deductive reasoning), which could withstand the

disruptive nature of cloud computing. However,

literature is available in this area, which would favour

a deductive approach thoroughly. As a pragmatist and

while looking be-yond the horizon, it might be

meaningful to combine both approaches. In the first

stage the inductive approach helps to break down the

complexity and explore the transition in the cloud

vendor market. Then, in a second stage when the

complexity decreased and less debate is excited,

Cloud Service Mediation through Brokerage Service

deductive reasoning can complement research with

more pro-found theories and results. Finally, the

foreseen approach is to collect data first and then

develop a theory out of it while keeping in mind that

deductive reasoning still will be applied in a later

stage.

 Strategy: Design Research, also referred to as

Design Science research (Vaishnavi 2007) is

particularly suited to Information Science

research and is grounded in the types of research

questions that are often asked within this field

(Vaishnavi, 2007). This research will follow the

process steps as described by (Vaishnavi, 2007),

which requires looking at the problem of service

brokerage in a cloud environment and extracting

possible solutions from the existing knowledge.

 Time Horizon: the intended research will follow a

cross-sectional approach as it is per-formed at a

particular point in time.

 Data Collection: with regard to the Design

Science Research phases (Vaishnavi, 2007) the

following data collection methods and partners

are intended:

o Awareness of the Problem: in this phase

literature review will be conducted and

interviews with governmental partners and

SME Associations will be held.

o Suggestion: in the suggestions phase

interviews and workshops with different

responsible roles within SMEs, Expert

Interviews with Consultant e.g. Auditors and

Scientific Experts will be held. First artefacts

will be developed in cooperation with

technical experts

o Development: developing of prototypes with

technical experts from software developing

companies and scientific experts.

o Evaluation and conclusion: for the evaluation

different test cases will be prepared and

afterwards tested with potential customers

from (SMEs and governmental institutions).

Results will be discussed with technical and

scientific experts.

5 EXPECTED OUTCOME

This PhD project is closely coupled with the CLIMB

research project funded by Swiss Commission for

Technology and Innovation. The author is also

CLIMB’s project lead. Goal of this project is to

develop and establish kind of a Service mediation

broker based on the Star Audit Certification. Thus,

the outcome of this PhD project is two folded. On the

one hand there will be a practical outcome. It is

expected to develop a software, which (i) supports

customer to identify their needs regarding cloud

services, which (ii) offers to providers the opportunity

to classify their services and to compare it with other

providers and (iii) which maps these described

services with the claimed client requirements and

gives a valuable recommendation of appropriate

services.

On the other hand it is expected that the project

will bring some scientific contributions like:

 A method to collect customer’s business and

functional requirements with regards to a cloud

service. The method has to be easy to understand

and usable for users without IT background.

 An evaluation method for benchmarking data

mapping methods based on a given mapping

challenge.

 A classification and weighting scheme for cloud

services with regards to service quality.

 A Cloud Service Description method to collect

and compare cloud service properties.

6 STAGE OF THE RESEARCH

6.1 Work in Progress

There are currently several activities in progress.

 With regards to the requirement elicitation part,

there an in depth literature review on requirements

engineering has been executed to gain insights on

current trends and techniques in this field.

Furthermore first expert interviews have been

made to identify the need and gap for a new

requirement evaluation method for cloud services.

 A first identification and analysis on different

mapping candidates like AHP, Fuzzy AHP, Graph

matching, Ontology, Rule Engine, ELECTRE,

TOPSIS, Statistical methods, term-based

matching and preference matrix has been done.

Currently a list of evaluation criteria are setup to

compare and classify these candidates.

 State of the art on cloud brokering approaches. As

result a draft categorisation on different cloud

brokering approaches and cloud service selection

has been made (e.g. service intermediation,

service aggregation, service arbitrage (service

mediation internal vs external)).

DCCLOSER 2016 - Doctoral Consortium on Cloud Computing and Services Science

Figure 2: Cloud Brokering Classification.

6.2 Next Steps

After finishing the current work in progress, there are

several further activities planned as follows:

1. Two to three requirement analysis approaches

will be piloted and tested with a set of potential

customers

2. Six mapping candidates will be short listed and

analyses more in detail. Four candidates of this six

options will be implemented and ranked, based on

test data.

3. Conducting a first study on approaches for service

classification and descriptions. Developing a

method to identify, describe and compare

qualitative and quantitative Cloud Services so that

a statement of compliance with the claimed

customer requirements can be made.

REFERENCES

Cambridge Technology Partners, 2011, Cloud Computing

in Switzerland, A Survey on the Adoption of Cloud

Computing by Swiss Companies, Zurich: CTP-

Consulting

Mell, P., Grance, T., 2011. The NIST Definition of Cloud

Computing. National Institute of Standards and

Technology (NIST). Available from: http://csrc.nist.

gov/publications/nistpubs/800-145/SP800-145.pdf

[Accessed 01 September 2013].

L. Xin and A. Datta, 2010, On trust guided collaboration

among cloud service providers; In Proc. of.

Collaborative Computing: Networking, Applications

and Worksharing (Collaborate-Com), pages 1–8.

IEEE.

NIST Cloud Computing Standards Roadmap:

http://www.nist.gov/itl/cloud/upload/NIST_SP-500-

291_Version-2_2013_June18_FINAL.pdf

S. Sundareswaran et al.; 2012, A Brokerage-Based

Approach for Cloud Service Selection, IEEE Fifth

International Conference on Cloud Computing.

M. Sun et al, 2013, Consumer-Centered Cloud Service

Selection using AHP, IEEE International Conference

on Service Science.

Gartner, 2013, Cloud services brokerages: The dawn of the

next intermediation age.

S. Kalepu, S. Krishnaswamy, and S. Loke, 2003, Verity: a

qos metric for selecting web services and providers. In

Proc. of Web Information Systems Engineering

Workshops, pages 131–139. IEEE

A. Mondal, K. Yadav, and S. Madria, 2010, Ecobroker: An

economic incentive-based brokerage model for

efficient handling multiple-item queries to improve

data availability via replication in mobile-p2p

networks. Databases in Networked Information

Systems, pages 274–283

R. Buyya et al., 2012, Cloudbus Toolkit for Market-

Oriented Cloud Computing, Manjra-soft Pty Ltd,

Melbourne, Australia,.

S. K. Garg, R. Buyya, and C. Vecchiola; “Mandi: a market

exchange for trading utility and cloud computing

services”, 2011.

S. Silas, et al., 2012, Efficient service selection middleware

using ELECTRE methodology for cloud environments,

Information Technology Journal, vol. 11, 01 01

C. Deng-Neng, et al., 2011, Applying Fuzzy AHP on

Product Selection Service in e-Commerce, in Service

Sciences (IJCSS), 2011 International Joint Conference

on, pp. 198-202.

T. L. Saaty, 1986, Axiomatic foundation of the analytic

hierarchy process, Management science, vol. 32, pp.

841-855.

S. K. Garg, S. Versteeg, and R. Buyya, 2012, A framework

for ranking of cloud computing services, Future

Generation Computer Systems, Available online 19

June 2012, ISSN 0167-739X.

F. K. Hussain and O. K. Hussain, 2011, Towards Multi-

criteria Cloud Service Selection in 2011 Fifth

International Conference on Innovative Mobile and

Internet Services in Ubiquitous Computing, pp. 44-48.

Henderson, J.C., Venkatraman, N., 1993; Strategic

Alignment: Leveraging Information Technology for

Transforming Organizations, IBM Systems Journal, 32,

1, pp. 4-16.

Papp, R., 2001; Introduction to Strategic Alignment, in R.

Papp (ed.): Strategic Information Technology:

Opportunities for Competitive Advantage. Idea Group,

Hershey, PA, pp.1-24.

PlugIT, cf. http://plug-it.org/itsocket/ [05.07.2014]

D. Wolff, M. Schaaf, S. Gatziu Grivas, U. Leimstoll, 2011,

Context-aware Website Personalization; Proceedings

of the 15th International Conference on Knowledge-

Based and Intelligent Information & Engineering

Systems (KES 2011).

A. Aamodt, 1994, Case-Based Reasoning:Foundational

Issues, Methodological Variations and System

Approaches, Artificial Intelligence Communications,

7(1), pp.39-59.

Vaishnavi, V., and Kuechler, W., 2007, Design Science

Research Methods and Patterns: Innovating

Information and Communication Technology, Boca

Raton, FL, New York, Auerbach Publications, Taylor

& Francis Group.

Cloud Service Mediation through Brokerage Service