CYCLONE: A Platform for Data Intensive Scientific Applications in Heterogeneous Multi-cloud/Multi-provider Environment

Full text

CYCLONE: A Platform for Data Intensive Scientific Applications in Heterogeneous
Multi-cloud/Multi-provider Environment
Yuri Demchenko, Miroslav
Zivkovic, Cees de Laat
University of Amsterdam
{y.demchenko, M.Zivkovic,
C.T.A.M.deLaat}@uva.nl
José Ignacio Aznar Baranda
I2CAT
jose.aznar@i2cat.net
Christophe Blanchet, Mohamed
Bedri, Jean-François Gibrat
CNRS IFB
{christophe.blanchet, mohamed.bedri,
jean-francois.gibrat}@france-
bioinformatique.fr
Oleg Lodygensky
LAL
oleg.lodygensky@lal.in2p3.fr
Mathias Slawick, Ilke Zilci
TU Berlin
{mathias.slawik,
ilke.zilci}@tu-berlin.de
Rob Branchat, Charles
Loomis
SixSq Sàrl
{rob, cal}@sixsq.com
Abstract—This paper presents results of the ongoing
development of the CYCLONE as a platform for scientific
applications in heterogeneous multi-cloud/multi-provider
environment. The paper explains the general use case that
provides a general motivation for the CYCLONE architecture
and provides detailed analysis of the bioinformatics use cases
that define specific requirements to the CYCLONE
infrastructure components. Special attention is given to the
federated access control and security infrastructure that must
provide consistent security and data protection for distributed
bioinformatics data processing infrastructure and distributed
cross-organisations collaborating teams of scientists. The paper
provides information about selected use cases implementation
using SlipStream cloud automation and management platform
with application recipe example. The paper also addresses
requirements for providing dedicated intercloud network
infrastructure which is currently not addressed by cloud
providers (both public and scientific/community).
Keywords- CYCLONE cloud automation platform for
scientific application, SlipStream Cloud Management platform,
Data intensive scientific applications requirements, Intercloud
Architecture Framework (ICAF), Federated Cloud Infrastructure
I. INTRODUCTION
Modern data intensive research require continuously
increasing power of computing resources and storage volume
that are in many cases required on-demand for specific
operations in data lifecycle (e.g. for data collection, extraction,
processing and reporting) and to be elastically scaled. Cloud
Computing [1, 2] provides a platform and environment for
data intensive scientific integration and enables effective use
of Big Data technologies and distributed data resources [3, 4].
Large volumes of data used in modern research, their
distributed nature, and need to support distributed
collaborative researcher groups create new challenges for the
scientific applications engineering.
The presented paper provides information about ongoing
development and implementation of the CYCLONE platform
for data intensive scientific applications development,
integration and automated deployment in heterogeneous
multi-cloud multi-provider environment [5]. The paper refers
to the general use case that motivates the general CYCLONE
platform architecture requirements and discusses in details
specific bioinformatics use cases to define specific
requirements to CYCLONE platform and infrastructure
components.
The CYCLONE platform is based on existing individual
components developed by the authors in previous projects:
SlipStream [6], StratusLab [7], TCTP [8], OpenNaaS [9], and
ensures their integration in a single platform supporting the
whole lifecycle of the scientific applications development.
The paper refers to the general Intercloud Architecture
Framework (ICAF) [10, 11] and the Intercloud Federation
Framework (ICFF) [12] proposed in the earlier authors’ work.
The remainder of the paper is organized as follows.
Section II describes the general use case and CYCLONE
architecture. Section III discusses bioinformatics use cases
and identifies specific requirements to bioinformatics
applications. Section IV provides suggestions for initial use
cases implementation, section V describes the SlipStream
cloud applications management platform, its functionality.
Section VI provides general information about CYCLONE
federated security infrastructure. The paper concludes with the
summary and remarks on future development in section VII.
II.
G
ENERAL USE CASE FOR
I
NTERCLOUD
A
PPLICATIONS
D
EPLOYMENT
P
LATFORM AND
CYCLONE
A
RCHITETCURE
Multiple individual use cases for multi-cloud applications that
require cloud and non-cloud resources integration into one intercloud
infrastructure that executes a single or multiple enterprise or
scientific workflows can be abstracted into general scenario and use
case illustrated in Figure 1. It includes two interacting applications,
that in general can be multi-cloud, that contain both application
related and management components. Application component
interacts with end users, management component is controlled by
application administrator and interacts with the (inter)cloud
management software. The figure also shows Cloud Applications
Deployment and Management Software and Tools as an important
component to support cloud applications deployment and operation
during their whole lifecycle. Intercloud infrastructure should also
provide two other components or services: federated access control
and security, and intercloud network infrastructure that needs to be
provisioned as a part of overall application infrastructure.
Intercloud
applications and infrastructure
may include
multiple
existing
cloud platforms. In the generally distributed

heterogeneous multi-cloud multi-provider environment the problem
of applications integration and management is becoming critical and
require smooth integration with the application workflow and
automation of most of development and operation functions, ideally
integration with the application specific development and operation
(DevOps) tools [13]. Currently widely used cloud automation tools
such as Chef [14], Puppet [15] allow single cloud provider
application deployment. They don’t solve problem of multi-cloud
resources/services integration and provisioning of inter-cloud
network infrastructure.
Figure 1. General use case for multi-cloud applications deployment.
The CYCLONE project attempts to solve this problem by
leveraging original cloud management platform SlipStream and
extending its with necessary functionality and components, in
particular for inter-cloud resources deployment and network
infrastructure provisioning, enabling federated access control for
users and end-to-end security for data transfer, enabling dynamic
trust establishment between cloud and application domains.
Intercloud platforms should deliver
open integration
environment and preferably standardized
APIs,
protocols,
and
data
formats,
allowing for cross-cloud resources interoperability
.
Practical Intercloud platform development should target two major
stakeholders and user communities the
Application
Service
Providers
(ASPs)
as
well
as their
customers to address real life
challenges and problems in a consistent and constructive way.
The effective cloud automation and management platform
should allow dynamic cloud resources allocations depending on
the workload and application workflow. This task can be solved
for single cloud using its native elasticity and load balancing
tools, however in intercloud environment such functionality will
require involving real cloud platform load (including resources
availability) and application monitoring [16].
III. BIOINFORMATICS USE CASES AND REQUIREMENTS
A. Overall description
Bioinformatics deals with the collection and efficient
analysis of biological data, particularly genomic information
from DNA sequencers. The capability of modern sequencers
to produce terabytes of information coupled with low pricing
(less than US$1000 for a human genome) that makes parallel
use of many sequencers feasible causes a "data deluge" that
is being experienced by researchers in this field [17, 18].
Bioinformatics software is characterized by a high degree
of fragmentation: literally hundreds of different software
packages are regularly used for scientific analyses with an
incompatible variety of dependencies and a broad range of
resource requirements. For this reason, the bioinformatics
community has strongly embraced cloud computing with its
ability to provide customized execution environments and
dynamic resource allocation.
The French Institute of Bioinformatics (IFB) [19] consists
of 32 regional bioinformatics platforms (PF) grouped into 6
regional centers spanning the entire country, and a national
hub, the "UMS 3601–IFB-core". The IFB has deployed a
cloud infrastructure on its own premises at IFB-core, and
aims to deploy a federated cloud infrastructure over the
regional PFs devoted to the French life science community,
research and industry, with services for the management and
analysis of life science data.
The CYCLONE project has identified two basic
bioinformatics use cases that aim to address specific
identified limitations while carrying out the analyses starting
from general cloud federation requirements and specific use
cases features to enhance current cloud infrastructure
processes and services provisioning mechanisms.
B. UC1 - Securing human biomedical data
1) Description
Continuous decrease of the genome sequencing costs
(NGS) allows increasing number of clinicians to include
genome analysis and data into their day-to-day diagnostic
practice. Today, most of genomics analyses are realized on
the exome, which is the expressed part (5%) of the genome.
However, the full genome sequencing is being envisaged and
will be soon included in daily medical practices.
It is expected that in the near future, some of the genomic
data processed on the IFB cloud platform will concern human
biomedical data related to patients and thus, will be the
subject to strict personal data protection regulations. To
ensure the data security while carrying out the analysis in a
federated cloud environment, the security of all involved sites
belonging to the federation must be ensured (especially when
involving both public and private cloud infrastructures).
2) Workflow
The use case workflow to ensure data security includes
the following steps (see Figure 2): (1) a biomedical user
connects to the cloud through the IFB web authenticated
dashboard; uses it to (2) run an instance of the appliance
containing the relevant pre-configured analysis pipeline. At
step (3) the VM containing genome applications is deployed
on the cloud (testbed); then (4) the user signs into the web
interface of the VM, (5) uploads the patient’s biomedical
data, and (6) runs the analysis in a secure environment.
Finally, (7) the user gets the results.

The bioinformatics treatment generally relies on a
comparison with the current release of the reference human
genome hg19 (Human Genome version 19 or GRCh37). The
hg19 is a database consisting of many files containing the
public genomics data. It can be used remotely (with sufficient
connectivity) or can be previously deployed by the cloud
providers as a public data set available to all users.
Figure 2: Functional schema of the use case “Securing human
biomedical data”. The figure shows the application components and
describes the different steps of the workflow.
3) Implementation requirements
This use case demonstrates the capability of CYCLONE
infrastructure to provide biomedical staff acting as cloud
users with the deployment of their own cloud infrastructure
for the biomedical analyses consisting of a single virtual
machine (potentially elastically scaled) with a web interface
in a secured environment. The access to this environment will
be based on the identity and authorizations in the federation.
The tests within the project will be carried out only on
non-personally identifiable data (either anonymized
benchmark data or simulated data).
C. UC2 - Cloud pipeline for microbial genomes analysis
1) Description
In the post-NGS research, sequencing bacterial genomes
is very cheap (few hundreds €) what allows researchers to
compare large collections of related genomes (strains). Thus,
this brings requirements to increasing need for automating
the annotation of bacterial genomes.
The IFB-MIGALE platform (one of the bioinformatics
platforms of the IFB) has developed an environment for the
annotation of microbial genomes and a tool for the
visualization of the synteny (local conservation of the gene
order along the genomes). The platform automatically
launches a set of bioinformatics tools (e.g. BLAST,
INTERPROScan) to analyse the data and stores the results of
the tools in a relational database (PostgreSQL). These tools
use several public reference data collections. A web interface
allows the user to consult the results and perform the manual
annotation (manual annotation means adding manually
metadata and biological knowledge to the genome sequence).
Installing the platform requires advanced skills in system
administration and application management. Performing the
analysis of collections of genomes requires large computing
resources that can distributed over several computers,
generally the computing nodes of a cluster.
The proposed CYCLONE cloud federation will allow the
life science researchers to deploy their own comprehensive
annotation platform over one or more cloud infrastructures.
Such deployments can be done with the dynamic allocation
of network resources for the isolation of the VMs inside a
dedicated private cloud including virtual network and
replicated user data.
2) Workflow
As illustrated in Figure 3, a bioinformatician (1) connects
to the cloud web dashboard, uses it to (2) run and (3) deploy
with one click a genomes annotation platform consisting of
many VMs, comprising of a master node of the virtual cluster
that provides also the visualization web-interface, associated
with several computing nodes. Then the user (4) uses secure
communication over SSH to connect to the master and (5)
uploads the raw microbial genomic data (MB) to the cloud
storage. SCP/SFTP protocols are used from a command line
tool or a GUI, to ensure AuthN/Z for the data transfer, and to
overcome the performance issues of HTTP for large datasets.
Still in command line interface, the user (6) runs the
computation to annotate the new microbial genomes. The
first step consists of many data-intensive jobs performing the
comparisons between the new genome and the reference data
Figure 3: Functional schema of the use case “Cloud virtual pipeline for
microbial genomes analysis”. The figure shows the application components
and describes the different steps of the workflow.
The results are stored in a relational database (provided
by a cloud service or a VM deployed within the platform).
Then the scientist (7) signs in the annotated data visualization
environment provided by the Insyght web-interface to 8)
navigate between the abundant homologues, syntenies and
gene functional annotations in bacteria genomes.
3) Implementation requirements
This use case will evaluate the capability of CYCLONE
infrastructure to provide a bioinformatician with a one-click
deployment of a complex application. This deployment needs
to be done in an isolated network for security and
confidentiality reasons. The access to this environment will
be based on the identity and authorizations in the federation.
The application may be deployed over several clouds because
it could require large computing resources not available in
one place or due to functional reasons some data or tools are
only available in certain environments or locations.
IV. IMPLEMENTATION OF USE CASES AND CYCLONE
INFRASTRUCTURE COMPONENTS
A. Deployment of the use cases
The deployment of the bioinformatics use cases was done
in a progressive manner, starting with one VM integrating
Use
case
User
IFB ’s
Cloud
interface
2. run VM
VM
5. upload
data patient
data
Secure
3. deploy
Cloud
hg19
refdata
isolated
network
7. get
results 6. compute
Pipeline
web
interface
4. sign in
1. sign in
Cloud
Use
case
User
web
interface
7. sign in
shared
data worker
worker
worker
worker
worker
worker
worker
worker
worker
worker
isolated
network
master
5. upload data
8. vizualize results
IFB ’s
Cloud
interface
2. run VMs
3. deploy
1. sign in
SSH
4. sign in 6. compute

security features (UC1 Securing human biomedical data).
Afterwards, we then deployed a complex application
requiring the coordinated deployment of several VMs (UC2
Cloud virtual pipeline for microbial genomes analysis).
1) Deployment UC1 Securing human biomedical data
The first deployed bioinformatics use case “Securing
human biomedical data“ is a single-VM application requiring
enhanced security features such as a trusted federated
authentication mode and a deployment done only on certified
(by the French Health Ministry) cloud infrastructure. The
cloud appliance NGS-Unicancer is developed by the
bioinformatics platform of the Centre Léon Bérard (Lyon,
France, www.synergielyoncancer.fr) in the context of the
project NGS-Clinique (INCA - Institut National du Cancer).
It provides a simple web interface to launch the biomedical
genomic analysis pipeline. The appliance was enhanced by
the CYCLONE Federation Provider and is ready for on-
demand deployment on the IFB-core cloud infrastructure.
The user deploys the appliance NGS-Unicancer through the
IFB web interface in “1-click” and uses the CYCLONE
federation provider to get access to the VM web interface
based on its identity in the federation. The user can then easily
upload its data, run the analysis and get the results. In Figure
4, the upper part describes the use case workflow, middle
layer represents the workflow steps that are linked to the
related CYCLONE software components and services. The
bottom part shows the testbed infrastructure components.
Figure 4: Functional relations between the use case “Securing human
biomedical data” and the Cyclone components.
2) Deployment UC2: Cloud virtual pipeline for microbial
genomes analysis
The second bioinformatics use case “Cloud virtual
pipeline for microbial genomes analysis“ is developed by the
platform IFB-MIGALE (Jouy-en-Josas, France,
migale.jouy.inra.fr). This application requires several
components: a user web interface, a relational postgreSQL
database, and a complete computing cluster with a master and
several nodes to perform the data-intensive analyses. This
infrastructure already running in a classical static way on
bare-metal servers in IFB-MIGALE premises was ported to
the cloud and extended with a « 1-click » deployment
features by using SlipStream recipes. The image was
exported from the IFB’s cloud and registered in the
StratusLab Marketplace. Afterwards, IFB-core wrote a
deployment recipe based on SlipStream that instantiates the
complete application with all the required VMs on the
CYCLONE infrastructure.
B. Secure Network Infrastructure Provisioning
Due to the generically distributed Bioinformatics
resources and applications required to be integrated in typical
use cases, there is a strong requirements to provision
dedicated secure virtualised network infrastructure. The
following are specific requirements:
 Dynamically provisioned high-performance network
infrastructure to interconnect multiple data sources
locations and enable high-performance computations
over large volume of data. This functionality is being
developed based on OpenNaaS network provisioning
and monitoring platform [9].
 Guaranteed network performance (QoS) is specifically
important for remote data visualisation such as IGV -
Integrative Genomics Viewer. The QoS (mainly
bandwidth and latency) will require the traffic
engineering mechanisms at the network level.
 Multi-domain Virtual Private Network (VPN) is required
to interconnect distributed data resource and application
components deployed in different clouds. Firewall
configuration is identified as an important part of
intercloud applications integration.
V. SLIPSTREAM: CLOUD APPLICATION MANAGEMENT
PLATFORM
Within CYCLONE, software developers and service
operators manage the complete lifecycle of their cloud
applications with SlipStream, an open source cloud
application management platform. Through its plugin
architecture, SlipStream supports most major cloud service
providers and the primary open source cloud distributions. By
exposing a uniform interface that hides differences between
cloud providers, SlipStream facilitates application portability
across the supported cloud infrastructures.
To take advantage of cloud portability, developers define
“recipes” that transform pre-existing, “base” virtual machines
into the components that they need for their application. By
reusing these base virtual machines, developers can ensure
uniform behaviour of their application components across
clouds without having to deal with the time-consuming and
error-prone transformation of virtual machine images.
Developers bundle the defined components into complete
cloud applications using SlipStream facilities for passing
SW
HW
IFB web
dashboard
SlipStream
StratusLab
/Openstack
Federation
Provider TCTP OpenNaaS
refdata(++)
hg19
DC A
User’s
LAN
Secure
Secure
WAN
academic
+ public
Bioinformatics
appliance
1. 2. 3. 4. 5. 6. 7.
Use
case
User
IFB ’s
Cloud
interface
2. run VM
VM
5. upload
data patient
data
Secure
3. deploy
Cloud
hg19
refdata
isolated
network
7. get
results 6. compute
Pipeline
web
interface
4. sign in
1. sign in

information between components and for coordinating the
configuration of services.
Once a cloud application has been defined, the operator
can deploy the application in “one click”, providing values
for any defined parameters and choosing the cloud
infrastructure to use. With SlipStream, operators may choose
to deploy the components of an application in multiple
clouds, for example, to provide geographic redundancy or to
minimize latencies for clients. To respond to changes in load,
operators may adjust the resources allocated to a running
application by scaling the application horizontally (changing
the number of virtual machines) or vertically (changing the
resources of a virtual machine).
SlipStream combines its deployment engine with an “App
Store” for sharing application definitions with other users and
a “Service Catalog” for finding appropriate cloud service
offers, providing a complete engineering PaaS supporting
DevOps processes. All of the features are available through
its web interface or RESTful API.
A. Functionality used for use cases deployment
The bioinformatics use cases described above principally
used SlipStream’s facilities and tools to define applications
and its deployment engine through the RESTful API.
The definition of an application component actually
consists of a series of recipes that are executed at various
stages in the lifecycle of the application. The main recipes,
in order, are:
 Pre-install: Used principally to configure and initialize
the operating system’s package management.
 Install packages: A list of packages to be installed on
the machine. SlipStream supports the package managers
for the RedHat and Debian families of OS.
 Post-install: Can be used for any software installation
that can not be handled through the package manager.
 Deployment: Used for service configuration and
initialization. This script can take advantage of
SlipStream’s “parameter database” to pass information
between components and to synchronize the
configuration of the components.
 Reporting: Collects files (typically log files) that should
be collected at the end of the deployment and made
available through SlipStream.
There are also a number of recipes that can be defined to
support horizontal and vertical scaling that are not used in the
defined here use cases.
The applications are defined using SlipStream’s web
interface, the bioinformatics portal then triggers the
deployment of these applications using the SlipStream
RESTful API.
B. Example recipes
The application for the bacterial genomics analysis
consisted of a compute cluster based on Sun Grid Engine with
an NFS file system exported from the master node of the
cluster to all of the slave nodes. The master node definition
was combined into a single “deployment” script that
performed the following actions:
1. Initialize the yum package manager.
2. Install bind utilities.
3. Allow SSH access to the master from the slaves.
4. Collect IP addresses for batch system.
5. Configure batch system admin user.
6. Export NFS file systems to slaves.
7. Configure batch system.
8. Indicate that cluster is ready for use.
The deployment script extensively uses the parameter
database that SlipStream maintains for each application to
correctly the configure the master and slaves within the
cluster. A common pattern is the following:
ss-display "Exporting SGE_ROOT_DIR..."
echo -ne "$SGE_ROOT_DIR\t" > $EXPORTS_FILE
for ((i=1; i<=`ss-get
Bacterial_Genomics_Slave:multiplicity`; i++ ));
do
node_host=`ss-get
Bacterial_Genomics_Slave.$i:hostname`
echo -ne $node_host >> $EXPORTS_FILE
echo -ne "(rw,sync,no_root_squash) ">> $EXPORTS_FILE
done
echo "\n" >> $EXPORTS_FILE # last for a newline
exportfs -av
which is used to export one NFS directory. The ss-get
command retrieves a value from the parameter database. In
this case, it determines the number of slaves and then loops
over each one, retrieving each IP address (hostname) and
adding it to the NFS exports file. This command will wait for
a value to be set, allowing it to act as a semaphore and
allowing coordination between the components of the
application. An analogous ss-set command allows the
values of parameters to be set.
A similar pattern is used for the SSH and batch system
configurations in this script.
VI. CYCLONE FEDERATED SECURITY INFRASTRUCTURE
The CYCLONE security infrastructure aims to enable
holistic security functionality in federated multi-cloud
infrastructures by offering a set of ready-to-use components.
The components are selected to best serve the needs of the
distributed bioinformatics data processing infrastructure and
the distributed cross organization collaborating research
teams. The main identified security scenarios include: i)
federated identity management, ii) federated authorization
management, iii) end-to-end secure data management.
Federated identity management is enabled by the
CYCLONE Federation Provider. The CYCLONE Federation
Provider provides federated user identities via the SAML 2.0
web authentication workflow with the local identity
providers. It supports all Identity Providers of eduGAIN [21].
Providing a federated identity includes user attribute
mapping, which means transforming SAML user assertions
into JSON Web Token claims. The current implementation
with Keycloak [22] and SimpleSAMLphp allows browser
based log-in to services which implement OpenID Connect
[23] client side which provides the library support in a variety

of programming languages. The use of OpenID Connect and
JSON Web Tokens add value to the CYCLONE system,
since these are much easier to implement in comparison to
SAML and its tokens for service providers. Moreover, with
extension of the SlipStream authentication service to
implement the OpenID Connect Authorization Code Flow,
eduGAIN users will be able to log in to the SlipStream web
dashboard to configure their deployments.
Federated authorization management addresses the
collaboration between VM developers and bioinformaticians
and the definition of access rules. It will be enabled by lists
of users and groups defined at the pre-deployment stage
which will be configured on the VMs. The user claims can be
used to define a simple configuration with Require
Statements in .htaccess on any Apache-hosted application
with the OpenIdConnect Module enabled.
CYCLONE security infrastructure will address the end-
to-end secure data management with the possible integration
of TCTP [8]. Further security infrastructure will address
dynamic trust infrastructure provisioning and trust
bootstrapping protocol [24, 25].
VII. CONCLUSION AND FUTURE DEVELOPMENT
This paper presents an on-going research and development
of the advanced CYCLONE platform for data intensive
applications development, deployment and management in
heterogeneous multi-cloud multi-provider environment.
The current stage of the development concludes the
development of a general platform components’ architecture,
functional design and pilot implementation of the core
infrastructure provisioning functionality for selected
bioinformatics usecases that represent one of the most
demanding use cases that require high-performance
dynamically configured multi-cloud infrastructure and
complex data protection and access control mechanisms. The
CYCLONE architecture is built around core functionality of
the SlipStream Cloud Management platform and extends its
with new functionalities to provision intercloud network
infrastructure and dynamically created security infrastructure.
ACKNOWLEDGEMENT
The research leading to these results has received funding
from the Horizon2020 project CYCLONE and the French
programs PIA INBS 2012 (CNRS IFB). We are thankful to
our colleagues who collaborate with us to deploy selected
bioinformatics applications in the CYCLONE testbed:
Christian Baudet (Centre Léon Bérard, Lyon, France) and
Thomas Lacroix (IFB-MIGALE, Jouy-en-Josas, France).
REFERENCES
[1] NIST SP 800-145, “A NIST definition of cloud computing”, [online]
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
[2] NIST SP 500-292, Cloud Computing Reference Architecture, v1.0.
[Online] http://www.nist.gov/customcf/get_pdf.cfm?pub_id=909505
[3] NIST Big Data Working Group Documents [online]
http://bigdatawg.nist.gov/V1_output_docs.php
[4] Demchenko, Yuri, Peter Membrey, Cees de Laat, Defining
Architecture Components of the Big Data Ecosystem. Proc. The 2014
Intern Conf. on Collaboration Technologies and Systems (CTS 2014),
May 19-23, 2014, Minneapolis, USA
[5] Slawik, Mathias, Yuri Demchenko, José Ignacio Aznar Baranda,
Robert Branchat, Charles Loomis, Oleg Lodygensky, Christophe
Blanchet,CYCLONE Unified Deployment and Management of
Federated, Multi-Cloud Applications, Proceedings The 8th IEEE/ACM
International Conference on Utility and Cloud Computing (UCC2015),
December 7-10, 2015, Limassol, Cyprus
[6] Slipstream Cloud Management Platform. [online]
http://sixsq.com/products/slipstream/
[7] StratusLab Scientific Cloud [online] http://www.stratuslab.eu/
[8] M.
Slawik,
“The
Trusted Cloud Transfer
Protocol,
”
in Proc.
2013
IEEE 5th
International Conference
on Cloud
Computing
Technology and Science
(CloudCom),
2-5 December 2013, Bristol, UK
[9] OpenNaaS [online] https://github.com/dana-i2cat/opennaas
[10] Demchenko, Y., C.Ngo, M.Makkes, R.Strijkers, C. de Laat, Intercloud
Architecture for Interoperability and Integration. Proc. The 4th IEEE
Conf. on Cloud Computing Technologies and Science
(CloudCom2012), 3 - 6 December 2012, Taipei, Taiwan.
[11] Demchenko, Y., et al, Intercloud Architecture Framework for
Heterogeneous Cloud based Infrastructure Services Provisioning On-
Demand. The 27th IEEE International Conference on Advanced
Information Networking and Applications (AINA2013). 25-28 March
2013. ISBN-13: 978-0-7695-4953-8.
[12] Y.Demchenko, C. Lee, C.Ngo, C. de Laat, Federated Access Control
in Heterogeneous Intercloud Environment: Basic Models and
Architecture Patterns. In Proc IEEE International Conference on Cloud
Engineering (IC2E), March 11, 2014, Boston, USA
[13] Davis, Jennifer; Daniels, Katherine (2015). Effective DevOps.
O'Reilly. ISBN 978-1-4919-2630-7.
[14] Chef: Cloud Automation deployment and DevOps platform [online]
https://www.chef.io/chef/
[15] Puppet: Cloud Automated Provisioning and Management
https://puppetlabs.com/
[16] Dumitru, C. , Oppresscu, AM., Živković, M. , R van der Mei, Grosso
P. , de Laat, C. A queueing theory approach to Pareto optimal bags-
of-tasks scheduling on clouds, Euro-Par 2014 Parallel Processing, pp.
162-173.
[17] Marx, V., Biology: The big challenges of big data.Nature, 2013 vol.
498 (7453) pp. 255-260
[18] Stephens Z.D., Lee S.Y., Faghri F., Campbell R.H., Zhai C., Efron
M.J., et al., Big Data: Astronomical or Genomical? PLoS Biol, 2015
vol. 13 (7): e1002195.
[19] French Institute of Bioinformatics – CNRS IFB UMS3601,
http://www.france-bioinformatique.fr/
[20] Lacroix, T., Loux, V., Gendrault, A., Hoebeke, M. and Gibrat, J.F.
Insyght: navigating amongst abundant homologues, syntenies and gene
functional annotations in bacteria, it's that symbol! Nucl. Acids Res.,
2014 vol. 42 (21): e162.
[21] eduGAIN: Federated Identity Provider infrastructure interconnection
service [online] http://services.geant.net/edugain/Pages/Home.aspx
[22] Keycloak: Integrated SSO and IDM for browser apps and RESTful
web services [online] http://keycloak.jboss.org/
[23] OpenID Connect [online] http://openid.net/connect/
[24] Ngo, C., Y.Demchenko, C. de Laat, Toward a Dynamic Trust
Establishment Approach for Multi-provider Intercloud
EnvironmentThe 4th IEEE Conf. on Cloud Computing Technologies
and Science (CloudCom2012), 3 - 6 December 2012, Taipei, Taiwan
[25] Membrey, P., K.C.C.Chan, C.Ngo, Y.Demchenko, C. de Laat, Trusted
Virtual Infrastructure Bootstrapping for On Demand Services. The 7th
International Conference on Availability, Reliability and Security
(AReS 2012), 20-24 August 2012, Prague.