dc.description.abstract |
Cloud computing provides scalable, elastic and on-demand services to various
individual consumers, organizations and IT industry. It o ers optimum
resource utilization, broad network access, availability and measured services
to all the consumers. The next evolutionary step of this technology is the
Cloud federation, which brings remarkable advancement in Cloud services by
extending the existing infrastructure for sharing of resources. It allows Cloud
Service Providers (CSPs) to share their computing resources to ful l the consumer's
requests when the demand exceeds the limitations at certain time.
One Cloud provider (home CSP) may rent in the resources of other provider
(foreign CSP) who is available with its vacant capacity of infrastructure.
Besides various advantages, the Cloud federation has many challenges
which mainly include optimum resource discovery, interoperable services, resource
migration, establishment of trust between the participating Cloud
providers and dynamic resource provisioning. Trust is one of the challenging
issues that limit the adoption of Cloud federation by service providers.
During the federation, the complete virtual machine images or partial data
objects are migrated from home to foreign CSP domain which brings security
and privacy concerns for Cloud consumers. In order to assure the security of
data at foreign CSP platform, there is need to evaluate and establish trust
between both participating Cloud providers.
Considerable work has been proposed over the last few years on trust
evaluation and establishment in Cloud computing; however trust issues still
emerge in Cloud federation with more advance risks. Existing trust evaluation
techniques, methodologies and mechanisms are known as trust models
in literature. These trust models are used to evaluate the trustworthiness of
CSP based on certain factors and establish trust between Cloud consumers
and CSPs. However, these modes are not adequate to evaluate trust between
two CSPs participating in federation. There is need for more e ective trust
models that are based on multiple factors for trust evaluation and dynamically
establish the trust in Cloud federation.
In this thesis we have carried out research in two major folds, where
2 Abstract
one direction includes the comprehensive analysis of existing trust models
in Cloud computing. After the detailed literature survey, we have proposed
a benchmark for assessment of trust models in Cloud domain comprising of
most essential trust evaluation parameters. Furthermore, we have proposed
foremost taxonomies for functional and non-functional features provided by
the existing trust models. First, we have explored and identi ed the obligatory
functional features of trust models and presented them in form of comprehensive
taxonomy. Our second taxonomy comprises of non-functional features
which mainly covers the security, control, deployment and performance
perspectives of trust models along with their mechanism of realization. We
have analyzed existing trust models using our taxonomies to nd the differences
among them, in terms of providing these essential functional and
non-functional features.
In second fold of research, we have proposed and implemented a trust
evaluation system for establishment of trust between home and foreign CSPs
participating in Cloud federation. The proposed system is based on two signi
cant factors for trust evaluation namely the feedback provided by Cloud
consumers and Service Level Agreements (SLAs) of participating service
providers. After the evaluation of trust for CSPs, the Security Assertion
Markup Language (SAML) is used to exchange the trust credentials between
both the service providers. We have extended the SAML protocol by
introducing a new pro le \Trust Credential Exchange" which includes the
new trust assertion and request response protocol. We have evaluated our
work using two important veri cation tools which are Netlogo and Scyther.
The veri cation of trust evaluation system is performed in Netlogo platform
whereas the exchange of trust credentials using SAML protocol is veri ed
through Scyther tool. |
en_US |