Identity Management in Cloud Platforms using VOMS and SPID
Francesco De Angelis, Fausto Marcantoni, Alberto Polzonetti and Samuele Rilli
Computer Science Division, University of Camerino, Palazzo Battibocca, Via del Bastione, 62032, Camerino (MC), Italy
Keywords: Cloud Computing, Identity Management, VOMS, SPID, Cloud Foundry, Openstack, SAML,
Authentication, Authorization.
Abstract: Cloud computing is being adopted more and more in recent years. It offers several benefits, such as high
elasticity, availability and cost reduction, but yet presents some issues. Among the most important, the
potential lack of security can affect the spreading of this technology. As cloud computing is pushing
forward to the digital era, where users can have their own digital identity to access restricted resources or
services, a reliable authentication and authorization system would attract more users to get involved in such
process. This paper proposes an integration of the VOMS (Virtual Organization Membership Service)
system for authorization and SPID (Sistema Pubblico per la gestione dell'Identità Digitale) system for
authentication, within Cloud Foundry PaaS (Platform as a Service) model. Considerations, differences and
interoperability matters will be addressed in order to provide a comprehensive scheme.
1 INTRODUCTION
Cloud computing, or simply cloud, is a word that is
getting more and more used in the last few years.
Although it is a novel technology, it is changing the
way we are used to thinking about programs and
computing. Cloud computing, in fact, aims to "shift
IT services away from local computers to the
Internet or, generally speaking, in networks. The
network will be the computer" (T-Systems
Enterprise¸ 2010). This means, many next
generation applications and resources will be
available directly on the Internet using a browser,
rather than on the local machine. Nevertheless,
issues affecting cloud computing are still slowing
down this revolution: among the most important, we
can point out the lack of standards, that affects the
interoperability among different providers, as well as
the security mechanisms for managing digital
identities and the authentication process.
In this paper we study the integration of Cloud
Foundry, a PaaS system, and OpenStack, an IaaS
system, with authorization and authentication
frameworks named VOMS (Virtual Organization
Membership Service) and SPID (Sistema Pubblico
per la gestione dell'Identità Digitale) respectively.
The former is a tool used mainly in grid computing
since several years. The latter is a system, being
developed by the Italian government, for federated
strong authentication; it is scheduled to enter service
in April 2015, and complies with eIDAS European
regulation, thus provides a framework interoperable
not only in the Italian territory.
The remainder of the paper is organized as
follow. Section 2 defines the IaaS and PaaS systems,
including their main features. The concepts of
identity and the relative identity management are
introduced in Section 3, in particular the definition
of authentication and authorization, and the
presentation of VOMS and SPID models. In Section
4 we consider the implementation of a model that
incorporates VOMS as authorization and SPID as
authentication systems.
2 IAAS AND PAAS TOOL FOR
CLOUD
A more technical definition of cloud computing is
given by (Armbrust et al., 2010) as "both the
applications delivered as services over the Internet,
and the hardware and systems software in the data
centers that provide those services".
Indeed, cloud providers supply IT resources in
terms of infrastructure, software, storage and
bandwidths, somehow like water and electricity are
daily supplied in our houses. Computing capability
is offered on demand to those companies that need
flexible IT resources, and then they pay only for the
time resources have been used, potentially leading to
cost savings.
96
De Angelis F., Marcantoni F., Polzonetti A. and Rilli S..
Identity Management in Cloud Platforms using VOMS and SPID.
DOI: 10.5220/0005450800960103
In Proceedings of the 11th International Conference on Web Information Systems and Technologies (WEBIST-2015), pages 96-103
ISBN: 978-989-758-106-9
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)
2.1 Cloud Models
Since cloud is a complex technology, it has been
layered by NIST (NIST, 2011) in three main models.
SaaS (Software as a Service): supplies the user
with the provider's application, and users do not
manage or control the underlying cloud
infrastructure including network, operating system,
and application capabilities. An example of SaaS
application is the e-mail service such as Google
mail, where the customer simply uses the
application, but doesn't know the software structure
and is not responsible for its maintenance.
PaaS (Platform as a Service): the provider
supports users with a framework for application
development, so that they can build and deploy their
own software. Users do not manage or control the
underlying cloud infrastructure including network,
operating system, and application capabilities, but
have control over the deployed applications and
possibly environment configurations.
IaaS (Infrastructure as a Service): user is
provided with fundamental computing resources that
he can use to deploy and run arbitrary software.
Users don't manage or control the underlying cloud
infrastructure but have control over operating
system, storage, deployed applications, and possibly
limited control over some networking components.
Several other "as a Service" are emerging
recently, most notable are Database as a Service
(DBaaS) and Backup as a Service (BaaS). The trend
is to shift any functionality to a service provided
through the Internet, that will bring to the dawn of
the XaaS, "Anything as a Service" (Dixon, 2014).
Some examples of PaaS software are OpenShift
(OpenShift, 2014), Salesforce (Salesforce, 2014),
AppScale (AppScale, 2014), CloudControl
(CloudControl, 2014), Cloud Foundry (Cloud
Foundry, 2014), Microsoft Azure Web Sites (Azure,
2014). Regarding the IaaS, instances of well known
systems are Amazon Elastic Compute Cloud
(Amazon EC2, 2014), OpenStack (OpenStack,
2014), Apache Hadoop (Apache Hadoop, 2014),
OpenNebula (OpenNebula, 2014).
Among all these options, we chose to use Cloud
Foundry as PaaS and OpenStack as IaaS, because
they are currently the leading projects in their
category. Both are open source, developed by
several different contributors, and supported by a
broad community; more information are provided in
Section 2.3.
2.2 Benefits and Downsides
Such technology can bring several benefits to users.
At first, a cost saving factor for a company. A user
can rent IT resources from a cloud instead of buying
a physical machine, allowing to start using few
resources at first and increase only when there is a
further need. In this way the company avoids the
initial hardware costs.
Then, the pay-as-you-go offer lets consumers
pay only for the used computing resources on a
short-term basis (for example, processors or storage
resources by hours), and release them when are not
needed anymore. The result is an optimization of
resource usage, instead of having an own server idle
during low workload periods, thus preventing the
overprovisioning issue. Another aspect is the
flexibility provided by the virtually infinite
computing capacity that cloud makes available; this
relieves the customer from the task to foresee near
future resource needs and relative resource provision
plans, avoiding the underprovisioning issue. Finally,
the cloud data center itself manages the underlying
infrastructure and technical problems, so that
customers do not have to concern about IT
maintenance nor to acquire a wide hardware
competence.
However, significant challenges are yet to be
addressed. Cloud is a young computing model and
many systems are still in a development state: a lack
of standard Application Public Interfaces (API)
brings providers to use proprietary interfaces in their
services, thus restricting the ability for consumers to
move from a provider to another.
2.3 Cloud Foundry and OpenStack
Cloud Foundry is an open source cloud computing
Platform as a Service (PaaS) being developed by
several contributors, among which EMC, IBM,
Rackspace, and VMware can be mentioned. Cloud
Foundry is designed to support application
development with high productivity, taking into
account SaaS integration: it brings innovation in
application services and at the same times
incorporates heterogeneous cloud deployment
options to facilitate migration. As overviewed by
(Heller, 2014), Cloud Foundry provides easy to
install frameworks that support languages such as
Java, Node.js, Ruby and Go. Once an application
has been deployed, Cloud Foundry stores it in an
image which then is run within a container.
OpenStack is an open source system that
provides Infrastructure as a Service (IaaS) functiona-
lities. It comprises a series of components such as:
Keystone for identity management, Nova for
computing, Neutron for networking, Glace for image
IdentityManagementinCloudPlatformsusingVOMSandSPID
97
storing, Horizon for dashboard, Swift for storage,
Ceilometers for telemetry and billing accountability,
Heat for orchestration. As an IaaS system,
OpenStack can be installed over the bare hardware;
it allows to create, launch, monitor and terminate
instances of virtual machines, connect them over
virtual networks, and mount virtual storage drives.
Cloud Foundry is one of those instances that can be
deployed over OpenStack.
3 IDENTITY MANAGEMENT
Another arguable aspect is the security of cloud.
Many of the applications available on the internet
usually require an authentication from the user. How
a service manages and stores the customers identity
is a very concerning issue. Privacy and data
protection are matters that cannot be taken lightly, as
users may need to provide their own personal
information to access a service, and the system must
assure the confidentiality of such data.
3.1 Authentication Vs Authorization
In a scenario where a user needs to authenticate, the
idea is to set up a centralized system, where digital
identities are managed by an external entity called
identity provider (IdP), instead of entrusting the
resource provider (the entity which supplies
resources) itself with this functionality. The users
identities are stored in the IdP, and a user can
authenticate against the IdP to obtain his own digital
identity. Additionally, including the IdP in a
federated environment, is it possible to extend the
centralized authentication to each cloud in the
federation. For the users, this means they can own a
single digital identity, and they can authenticate with
that digital identity in any cloud belonging to the
federation. Regarding resource providers, they won't
have to store, secure and assure the privacy of the
users information, as this work is delegated to IdPs.
In such context, we can point out the concepts of
authentication and authorization. The former deals
with the user identity: it permits to verify a person as
he login to an application. The latter allows to grant
permissions to carry out a given action, and applies
both to users and to processes that must access a
protected resource (Alfieri et al., 2005).
Although these two operations may feel very
similar, actually they are pretty different. A user
authenticates against a website to access to his
account. Once logged, the user gains full control and
thus can perform any action enabled for that
account. In contrast, the user can authorize a third
party application or website to use some
functionalities or to act as behalf of the user. In this
case, the third party application is not provided with
user credentials, but instead with a token it can use
to request access to resources or functionalities. The
above approach is a step forward to improve
security, and it allows a user to have a single digital
identity instead of creating a second account in the
third party website.
This means more control for the user, as he can
select and restrict the allowed operations for the
third party app, and at the same time the third party
collects only user's token instead of his credentials.
If the third party gets hacked, the hacker will be able
to perform only the operations permitted by the user.
to the third party. To solve this, the user can login to
the main site and revoke access granted to the third
party application, and no other action on the original
website, such as password reset, is needed.
Authentication process can be performed using
different kind of information: a password
(something the user knows); smart cards that contain
the user identifier (something the user has); a
physical trait or characteristic that acts as user
identifier, such as fingerprints and voice recognition
(something the user is). Currently the most used
techniques are the former (usually with username
and password) and the second (typically a certificate
contained in a smart card, or a OTP token).
In recent years the use of double factor
authentication is increasing, since it offers a stronger
security: smart cards are more difficult to be cloned
than a password to be stolen. However this solution
requires the user to always bring with him the smart
card containing the identifier token. On the other
hand, the password-base authentication remains the
wider adopted solution, and reason is the simplicity
of the process: users only need to remember their
credentials, without the need of having any
additional card that provides a certificate or device
to receive an OTP. Furthermore, this is the most
technologically neutral solution, as it do not require
any specific device to read user’s smart card.
Some examples of authentication systems used in
cloud are: OpenID (OpenID 2015), User Account
and Authentication (UAA, 2015), DIGIPASS as a
Service (DIGIPASS, 2015), PowerBroker Open
(PowerBroker Open, 2015). Instances of
authorization tools are Conjur (Conjur, 2015), UAA,
and OAuth 2.0 (OAuth 2.0, 2012).
3.2 VOMS
VOMS (Virtual Organization Membership Service)
(VOMS, 2014) is defined by (Venturi et al., 2008) as
WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies
98
a tool that allows to define a dynamic collection of
individuals, institutions, and resources, in order to
flexible, secure, coordinated resource sharing across
dynamic, multi-institutional collaborations. It is
based on the concept of Virtual Organization (VO)
that defines virtual collections, and adds
functionalities to manage these abstract entities. It
permits a resource owner to define a set of rules for
sharing his resources, which can be used to drive
authorization decisions. Such characteristic fits well
in the grid computing system, an indeed VOMS is
currently the de-facto standard for VO management:
it is already being used by Lightweight Middleware
for Grid Computing (gLite) now developed by EMI
(EMI, 2014), and the Virtual Data ToolKit (VDT)
(VDT, 2014) grid infrastructures.
3.2.1 VOMS Architecture
In order to provide its functionalities, the VOMS
supplies several interfaces with different features
(Alfieri et al., 2004):
User client: sends a request to the User server
containing user credentials (typically a X.509
certificate) and obtains a list of groups, roles
and capabilities of the user;
User server (we will refer to it more generally
as VOMS server): receives requests from a
user client and returns information about the
user signed with the server X.509 certificate;
Administration Client: is the tool used by VO
administrators to manage users, groups, roles,
etc;
Administration Server: receives the requests
from administration clients and accordingly
updates the data in the system.
The server is essentially a front-end to an
RDBMS, where all the information about users is
stored. It acts as an Attribute Authority; the user
sends to the VOMS server a request for attributes,
and authenticates using his certificate. The VOMS
server creates signed assertions containing the user’s
requested VO attributes, according to the groups and
roles he has been assigned. The user, once received
the information, creates a proxy certificate including
assertions returned by the VOMS server. This proxy
certificate is presented to the resource provider,
which uses those information to decide whether
granting resources or not. The user can request
certificates from more than one server.
More in detail, the authorization process is
accomplished following the steps below (Alfieri et
al., 2005):
1. The user and the VOMS server authenticate
each other using their certificates;
2. The user creates a request, signs it with his
certificate, and sends it to the VOMS server;
3. The VOMS server verifies the user’s identity
and checks the syntactic correctness of the
request;
4. The VOMS server creates a response
containing the required information, signs it
with its certificate, and sends it back to the
user;
5. The user checks the validity of the information
received;
6. The user optionally repeats this process for
other VOMS servers;
7. The user creates the proxy certificate
containing all the information received from
the VOMS server(s); information is included in
an Attribute Certificate signed by the VOMS
server itself;
8. The user presents the proxy certificate to the
resource provider, which decides if granting
the access to resources.
3.3 Spid
SPID (Sistema Pubblico per la gestione dell'Identità
Digitale) (SPID, 2014) stands for "Public System for
the management of the Digital Identity"; it is a
project lead by the AgID (Agenzia per l'Italia
Digitale), the Italian agency in charge of the
realization and diffusion of information and
communication technology.
The goal is to provide a federated and centralized
authentication service, by managing the registration
and provisioning tools; the strong authentication
implies that a certain digital identity corresponds
unambiguously to a specific person. As this system
has been designed primarily to facilitate the work of
public administrations (but not only), the user that
performs the authentication must be accurately
identified. Trust is enforced by AgID, which
maintains a list of entities complying with SPID
specification and performs the necessary monitoring
activities.
3.3.1 SPID Architecture
SPID establishes a federated environment formed by
Identity Providers (IdP), Resource Providers (RP)
called also Service Providers (SP), users and
Attribute Authorities (AA). An IdP is an entity
which stores the users digital identities, and allows
users to authenticate; on the other hand, an RP is an
IdentityManagementinCloudPlatformsusingVOMSandSPID
99
entity that provides services and resources. An AA
releases users qualified attributes, generally
requested by RP to compute permission decisions. In
a federated environment, SPID allows a user to
authenticate against one IdP to obtain the strictly
necessary information regarding his identity, needed
to access a protected service or resource offered by
an RP. In this scenario authentication is delegated to
the IdP rather than to the RP itself.
SPID provides three levels of authentication due
to the need to comply with currently most adopted
technologies as well as to keep up with the newest
ones; each level complies with the international
ISO/IEC DIS 29115 (ISO/IEC 29115, 2014). The
first level requires username and password; the
second level implies multi-factor authentication
based on OTP token; the third level requires multi-
factor authentication based on digital certificates.
To perform authentication, SPID can incorporate
several IdPs, and each of them is an independent
entity complying with SPID specifications. Thus a
user digital identity is not replicated on each IdP, but
is rather stored in one location. Then, when the user
authenticates against the IdP that provides his
identity, that authentication is accepted by any RP in
the system, because RPs do not discriminate IdPs.
The correct interaction sequence follows these steps:
1. the user requests access to a resource on a RP
through a User Agent (UA, generally a
browser);
2. the RP sends to UA an authentication request;
3. the UA is redirected to the appropriate IdP;
4. the IdP starts a challenge for credentials with
the user;
5. the user provides his credentials;
6. the IdP checks for credentials correctness then
returns assertions, signed by IdP itself,
containing the authentication statements for the
RP to the UA, otherwise if they are incorrect
continues with the challenge.
7. the UA forwards the assertion created by the
IdP to the RP;
8. the RP checks the validity of assertions,
identifies the user and delivers the service.
4 CONSIDERATION OF
IMPLEMENTATION OF
IDENTITY MANAGEMENT IN
THE CLOUD PLATFORM
In order to improve privacy and data protection on
cloud computing, we are pioneering and studying a
model in which, in order to access a resource or
service supplied by a resource provider,
authentication is delegated to a federated identity
provider, and authorization is performed by a
federated attribute authority. Specifically, we are
proposing an integration of VOMS for authorization
ad SPID for authentication, in the Cloud Foundry
PaaS system. Such integration aims to exploit the
federation environment provided by SPID, and to
incorporate the qualified attribute release performed
by VOMS.
4.1 VOMS Vs SPID
Although VOMS and SPID may seem similar, these
are two deeply different concepts. VOMS role is to
supply authorization information regarding a user in
a certain domain. It provides a single service that
can be executed on a single machine. SPID is a more
complex system, comprises several entities that
cooperate and communicate using protocols adopted
by SPID. It orchestrates the interaction of IdPs,
users, RPs and AAs, and also specifies the technical
standards for interoperability. In this context, VOMS
is one of these entities: it is essentially an AA that
operates performing authorization service by issuing
tokens on demand; on the other hand, SPID includes
IdPs which provide authentication service.
Another relevant difference lies in the policy
matters. A VOMS service do not require any
accreditation, thus can be set up and made run
autonomously. SPID is a system that the Italian
government aims to adopt in a large scale, in order
to spread the use of digital identities, thus entities
must be accredited. Finally, VOMS is being used,
mainly in grid computing, since several years,
whereas SPID is a novel scheme targeted for the
management of digital identities, thus supplies a
wider and more complete view over this issues.
4.2 Integration of VOMS and SPID in
Cloud Foundry
For the integration to be successful, each system
must be analyzed in order to understand how it
works and interacts with users or other components.
In particular, we analyzed the input and output of the
involved systems. The goal is to create a structure
composed by SPID and VOMS that can operate and
replace the Cloud Foundry build-in login system,
called UAA; to accomplish this, first we need to
understand its mechanisms.
The UAA (User Account and Authentication
Service) is the identity management service of Cloud
WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies
100
Foundry. As stated in (Features of the UAA, 2012),
it "is responsible for securing the platform services
and providing a single sign on for web applications",
and offers several features such as Centralized
Identity Management, Delegation Access to
Services, User Account Management and Client
Application Registration.
When a user performs the login into a Cloud
Foundry Resource Provider, the authentication and
authorization processes are handled by this
component. (UAA Server, 2012) specifies that the
UAA acts both as Login Server and an OAuth 2.0
Authorization Server (OAuth 2.0, 2012), managing
resources access by granting tokens which are
delivered to the client application. These tokens
contain, among the other, user information such as
the 'user_id', the 'user_name', and the 'scope' that
specifies the level of access granted by the user to
the token.
Figure 1 reports and compares the input and
output of each system we analyzed. This step is
crucial to understand if VOMS and SPID can be
integrated in Cloud Foundry, and which gaps are
still to be filled.
Figure 1: Comparison of the input and output requirements
of UAA, VOMS and SPID.
When accessing resources on the internet,
security must be ensured, especially when an
authentication process is performed. SPID issues this
problem by using standard SAML (SAML 2.0,
2005) message format.
SAML is an XML-based data format protocol
used for authentication and authorization process.
The SAML specification, accordingly with SPID
requirements, defines three roles: the user, the IdP,
and the RP. It specifies three components:
assertions, protocol, and binding. There are three
assertion types: authentication (validates the user's
identity), attribute (contains specific information
about the user), and authorization (specifies the
user's authorizations). Then, the protocol defines the
format for sending and receiving assertions, whereas
binding defines how SAML messages can be
mapped to SOAP messages.
On the other hand, VOMS uses certificates to
address the identification of the users, needed to
release the user qualified attributes; however
(Venturi et al., 2008) shows how VOMS can be
adapted to exchange SAML messages. VOMS runs
in an own server rather than living inside the
authentication server. If a HTTP server is set up,
VOMS can be used to perform SAML authorization,
as SAML can transport both authorization and
authentication information; such SAML-based
VOMS can therefore accept SAML requests
containing the user identifier, for instance the user
certificate, and return a SAML message composed
of assertions containing user authorizations.
In this context, SPID acts as authentication
service, allowing a user to provide his credential in
order to authenticate in the system through an IdP,
and VOMS operates as an external AA, performing
authorization service.
Figure 2: SPID and VOMS integration in Cloud Foundry.
Whenever a user wants to access a resource on a
Cloud Foundry RP, the system performs the
following steps, depicted in Figure 2, in this order:
1. the user requests access to a resource on a RP
through a User Agent (UA);
2. the RP sends to UA an authentication request;
3. the UA is redirected to the corresponding IdP;
4. the IdP starts a challenge for credentials with
the user;
IdentityManagementinCloudPlatformsusingVOMSandSPID
101
5. the user provides his credentials;
6. the IdP checks for credentials correctness then
returns assertions, signed by IdP itself,
containing the authentication statements for
the RP to the UA, whereas if credentials are
incorrect continues with the challenge.
7. the UA forwards the assertions created by the
IdP to the RP, the RP checks for validity of
the assertions and identifies the user;
8. the RP sends to the VOMS an authorization
request, including assertions of the
authenticated user;
9. the VOMS checks for validity of assertions,
and then returns a response, which includes
assertions containing user authorization
information, to the RP;
10. the RP checks for validity of assertions,
checks if the user is authorized to access the
resource and delivers it.
As already mentioned, during the communication
SAML messages containing or requesting user
information are sent. SAML messages are encoded
using the Base64 format and compressed with a
DEFLATE algorithm. In accordance with SPID
guideline (SPID specifications, 2014), a SAML
message, which is XML-based, must include
specific tags. Among these, the most relevant is the
tag that contains the assertions. An assertion is
defined by the <AttributeStatement> tag in the
following format:
<saml:AttributeStatement ...>
<saml:Attribute...>
<saml:AttributeValue ...>
myAttributeValue
</saml:AttributeValue>
</saml:Attribute>
</saml:AttributeStatement>
Beside the benefits mentioned above,
authentication (not limited to the cloud environment)
still presents risks regarding security. Identity theft
is one of the most concerning matters in a federated
centralized environment: if a user credentials are
stolen, the hacker can access all the services with the
same credentials. This problem is issued by using
strong authentication protocols, where information
are sent over secure connections.
Moreover, a large scale deployment of such
architecture requires specific software to be
deployed on each entity. RPs must be able to create
SAML requests, redirect users to their IdP, and
receive SAML responses. IdPs must be able to
receive SAML requests, perform authentication, and
return SAML responses redirecting users back to the
RP. SPID specifies that IdP must use non-
proprietary solutions (in terms of software and
hardware), in order to foster interoperability and
avoid vendors technological lock-in.
Setting up a working federation based on SPID
will require time to complete the specifications, test
the interactions between entities and address any
further technical issue. The diffusion of this
framework is going to be gradual: all the Italian
public administrations are expected to become SPID
RP compliant within 24 months since the
accreditation of the first IdP. Private entities may
intentionally join the federation, once they are
become SPID compliant. In addition, the number of
digital identities will increase over time, together
with the spreading of this technology among the
territory, only if this solution will prove to be
reliable, efficient and user-friendly. SPID aims to
simplify and secure the user access to services, but
may not succeed if such features, that are
fundamental for a wide adoption of digital identities
in a federation, are not ensured.
5 CONCLUSIONS
In this paper we studied a cloud computing system
handling identity management issues. First we
presented the state of the art of cloud technology,
describing the principal advantages and downsides
affecting it. The main goal was to improve the
identity management in the cloud PaaS model, in
particular Cloud Foundry. For this purpose, VOMS
as authorization and SPID as authentication systems
were chosen to be integrated. Regarding the
communication, SAML was chosen as it is a
protocol created to transport authentication and
authorization statements. Such solution can benefit
in terms of trust (since SPID provides a federate
environment), usability proof (since VOMS is being
used in grid since many years), and interoperability
(since SAML is a open-standard format based on
international XML format).
A future work may consist in setting up a system
as discussed above. However, SPID is currently in a
development phase, and it has not been deployed in
Italian territory yet.
REFERENCES
T-Systems Enterprise¸ 2010, White Paper Cloud
Computing. Alternative sourcing strategy for business
ICT. T-Systems Enterprise.
WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies
102
Armbrust, M., Fox, A., Griffith, R., Joseph, A.D., Katz,
R., Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,
Stoica, I., Zaharia, M., 2010. A View of Cloud
Computing, ACM Digital Library.
NIST, 2011. Cloud Computing Reference Architecture.
http://www.nist.gov/customcf/get_pdf.cfm?pub_id=90
9505. Accessed: 2015/03/17.
Dixon, J., 2014. X as a service (XaaS):What the future of
cloud computing will bring. http://www.cloud
computing-news.net/news/2014/aug/18/x-as-a-service-
xaas-what-the-future-of-cloud-computing-will-bring.
Accessed: 2015/03/17.
OpenShift, 2014. OpenShift Online. https://www.
openshift.com/products/online. Accessed: 2015/03/17.
Salesforce, 2014. http://www.salesforce.com. Accessed:
2015/03/17.
AppScale, 2014. http://www.appscale.com. Accessed:
2015/03/17.
CloudControl, 2014. https://www.cloudcontrol.com.
Accessed: 2015/03/17.
Cloud Foundry, 2014. http://cloudfoundry.org/index.html.
Accessed: 2015/03/17.
Azure, 2014. http://azure.microsoft.com/en-us/services/
websites. Accessed: 2015/03/17.
Amazon EC2, 2014. https://aws.amazon.com/ec2.
Accessed: 2015/03/17.
OpenStack, 2014. http://www.openstack.org. Accessed:
2015/03/17.
Apache Hadoop, 2014. http://hadoop.apache.org.
Accessed: 2015/03/17.
OpenNebula, 2014. http://opennebula.org. Accessed:
2015/03/17.
Heller, M., 2014. Review: Cloud Foundry brings power
and polish to PaaS. http://www.infoworld.com/
article/2608299/cloud-computing/review--cloud-
foundry-brings-power-and-polish-to-paas.html.
Accessed: 2015/03/17.
Alfieri, R., Cecchini, R., Ciaschini, V., Dell’Agnello, L.,
Frohner, A., Lorentey, K., Spataro, F., 2005. From
gridmap-file to VOMS managing authorization in a
Grid environment. FGCS.
OpendID, 2015. http://openid.net/. Accessed: 2015/03/17.
Cloud Foundry UAA, 2012. Introducing the UAA and
Security for Cloud Foundry. http://blog.cloudfoundry.
org/2012/07/23/introducing-the-uaa-and-security-for-
cloud-foundry/. Accessed: 2015/03/17.
DIGIPASS, 2015. DIGIPASS as a Service - Cloud based
Authentication.
https://www.vasco.com/products/managed_services/da
s/digipass_as_a_service.aspx. Accessed: 2015/03/17.
PowerBroker Open, 2015. http://www.powerbrokeropen.
org/. Accessed: 2015/03/17.
Conjur, 2015. What Is Conjur?. http://www.conjur.net/
what-is-conjur/. Accessed: 2015/03/17.
OAuth 2.0, 2012. An open protocol to allow secure
authorization in a simple and standard method from
web, mobile and desktop applications.
http://oauth.net/2/. Accessed: 2015/03/17.
VOMS, 2014. VOMS: Virtual Organization Membership
Service.
http://toolkit.globus.org/grid_software/security/voms.p
hp. Accessed: 2015/03/17.
Venturi, V., Riedel, M., Memon, Shi., MemonSha.,
Stagni, F., Schuller, B., Mallmann, D., Tweddell, B.,
Gianoli, A., Van denBerghe, S. et al., 2008. Using
SAML-Based VOMS for Authorizationwithin Web
Services-Based UNICORE Grids. Springer.
EMI, 2014. http://www.eu-emi.eu/. Accessed: 2015/03/17.
VDT, 2014. http://wlcg.web.cern.ch/virtual-data-toolkit.
Accessed: 2015/03/17.
Alfieri, R., Cecchini, R., Ciaschini, V., Dell’Agnello, L.,
Frohner, A., Gianoli, A., Lorentey, K., Spataro, F.,
2004. VOMS, an Authorization System for Virtual
Organizations. Springer.
SPID, 2014. Sistema Pubblico per la gestione dell'Identità
Digitale - SPID. http://www.agid.gov.it/agenda-
digitale/infrastrutture-architetture/spid. Accessed:
2015/03/17.
ISO/IEC 29115, 2011. ITU-T Recommendation X.1254 |
International Standard ISO/IEC DIS 29115.
https://www.oasis-
open.org/committees/download.php/44751/285-
17Attach1.pdf. Accessed: 2015/03/17.
Features of the UAA, 2012. High Level Features of the
UAA. http://blog.cloudfoundry.org/2012/07/24/high-
level-features-of-the-uaa-
2/#centralized_identity_management. Accessed:
2015/03/17.
UAA Server, 2012. User Account and Authentication
(UAA) Server.
http://docs.cloudfoundry.org/concepts/architecture/uaa
.html. Accessed: 2015/03/17.
SAML 2.0, 2005. SAML V2.0. http://saml.xml.org/saml-
specifications. Accessed: 2015/03/17.
SPID specifications, 2014. SPID regole tecniche e
modalità attuative.
http://www.agid.gov.it/sites/default/files/regole_tecnic
he/spid_regole_tecniche_v0_1.pdf. Accessed:
2015/03/17.
IdentityManagementinCloudPlatformsusingVOMSandSPID
103