1st
European Workshop on Dependable Cloud Computing
EWDCC
2012
In conjunction with 9th European Dependable Computing
Conference - EDCC 2012
May 8th,
2012
OVERVIEW
With the increasing variety of cloud
offerings, ranging from very basic infrastructure services, such as storage and
virtual machines, to more complex platform services (e.g., providing
database-like functionality), to end-user directed solutions such as web-based
office tools, there is an ongoing trend to outsource all kinds of applications
to the cloud.
However, recent studies on cloud
offerings and hardware in general show that clouds are less reliable than
traditional data centers, and hardware faults are
more common than previously assumed. Moreover, the very fact that cloud
services need to be accessed remotely, through the Internet, make their
perceived availability directly dependent from the availability of the
communication between clients and providers. In combination, this basically
inhibits the outsourcing of critical applications (e.g., financial or medical
services) to the cloud.
Besides that, recent failure events
on big cloud providers such as Amazon raised an alert about the dangers of
running critical applications on such third party infrastructures.
The 1st EWDCC workshop aims to
provide a platform for discussion and dissemination of early-stage recent
research efforts that explicitly aim at supporting dependability in the context
of cloud computing. We expect the workshop to foster discussion between
researchers from both industry and academia.
WORKSHOP PROGRAM
The workshop will be composed by two keynotes, two papers (selected among nine submissions) and five cloud-related EU-funded project
presentations. The workshop program is also available on EDCC’12
program booklet.
·
9:00hs
Keynote: The Cosca
PaaS platform: on the way to flexible and dependable
cloud computing. Franz Hauck (Institute of Distributed Systems, University
of Ulm, Germany)
This talk will address our current research projects
around cloud computing. The focus is on COSCA an OSGi-inspired
PaaS system that allows applications to be composed
from components. These can be individually updated and deployed similar to
bundles in OSGi. COSCA further supports sockets for
arbitrary communication. It is able to combine them with elasticity as known
from typical PaaS systems. In the second part of the
talk, we will address dependability issues in this context and how we plan to
connect them to COSCA. We will present multiple building blocks for
fault-tolerant applications that were developed within COSCA's predecessor
projects and are ready to be integrated.
·
10:00hs Bridging the security drawbacks of virtualized
network resource provisioning model. Ayush
Sharma, Volker Fusenig, Ingmar Schen
(Fraunhofer, Germany), Anand
Kannan (KTH, Sweden)
·
10:30-11:00hs
Coffee Break
·
11:00hs Securing Cloud-based Computations against
Malicious Providers. Adil Bouti,
Jorg Keller (FernUniversitat
in Hagen)
·
11:30hs How
to build a reliable mOSAIC of multiple Cloud services
Dana Petcu, West University of Timisoara and
Institute e-Austria Timisoara, Romania
The usage of the infrastructure services of multiple
Clouds, either to easily migrate an application from one Cloud to another, or
to benefit from the special facilities of particular Clouds, requires to reach a certain level of
application portability and technology independence, which are currently hardly
reachable in the absence of well established standards. One approach is to consider a new set of APIs, offering a level of abstraction that ensure
the vendor agnosticity, to follow the main idea of PaaSs, releasing the application developer from the burden
of remote infrastructure settings, and to develop an open-source alternative to
the existing PaaSs, allowing the usage of services
from multiple providers of IaaSs. mOSAIC
(Open-source API and Platform for multiple Clouds) proves the viability of this approach through
its recent software releases and initial test results. Beyond the particular
approach for portability, special features to be reveled
in the workshop presentation  are related to fault-tolerance at
application components level, programmable elasticity of the components, and
the reasons for an event-driven architecture and asynchronous communications.
·
12:00hs Elastic,
scalable and self-tuning data replication in the Cloud-TM platform
Paolo Romano (INESC-ID/IST, Portugal)
Cloud-TM is a
recent EU project focused on the development of an innovative data-centric
platform for the cloud. Cloud-TM leverages on efficient in-memory data
replication mechanisms and on pervarsive self-tuning
techniques aimed at maximizing efficiency in presence of fluctuations of the
incoming workload as well as of elastic re-sizing of the underlying
infrastructure. This talk will overview some recent results achieved by the Cloud-TM project in the area of transactional
data replication, focusing on the issues of elasticity, consistency,
scalability and self-tuning.
·
12:30-14:00hs
Lunch
·
14:00hs TClouds, Trustworthy Clouds Marcelo Pasin (University of Lisbon, Portugal)
TClouds is an
integrated research project funded by the FP7, focused on privacy protection in
cross-border cloud infrastructures and on ensuring resilience against failures
and attacks, while preserving scalability and cost-efficiency. It addresses
legal and business implications, to contribute to create regulations for
privacy-enhanced clouds. It also defines an architecture with enhancements for
securing commodity clouds and provides secure and resilient middleware using
single and multiple of clouds. A prototype involving critical IT-systems is
under construction, including a city lighting system with increased resilience
and fault tolerance, and a patient-centric home healthcare service that
preserves privacy. The talk presents the project goals and structure, and focuses
on the trustworthy architecture defined within the project and some
cloud-of-clouds enhancements.
·
14:30hs Contrail:
A reliable and trustworthy cloud platform for critical applications
Roberto Cascella (INRIA, France)
Guaranteeing
reliability in cloud computing is both a challenging task and an important
feature to make the cloud trustful to run business applications. The support of
dependability calls for ensuring the availability of the computation resources
and for strict guarantees in terms of quality of service (QoS
) and protection (QoP) so that users and
organizations can rely on cloud computing for their critical applications. Contrail develops an integrated approach
to virtualization, offering Infrastructure as a Service services (IaaS), services for federating IaaS
clouds and Contrail Platform as a Service services (ConPaaS)
on top of federated clouds. Contrail technology deploys ConPaaS
services which are self-managed, elastic and scalable, and the reliable cloud
file system GAFS. One of the primary objectives of Contrail is to offer
performance (QoS) and security (QoP)
guarantees via SLA enforcement, data location awareness, and a scalable
management of the computing resources via an interoperable Federation. The
first public release of the Contrail system is forthcoming.
In this talk I will highlight how the results of the Contrail project could
support dependability and the development of trusted and realiable
applications.
·
15:00hs Addressing
Dependability Issues in Storage Clouds: The VISION Cloud approach Dimosthenis Kyriazis (National
Technical University of Athens, Greece)
"It's become the phrase du
jour" according
to the Gartner senior analyst Ben Pring. Cloud
computing offers the potential to dramatically reduce the cost of software
services through the commoditization of IT assets and the efficient adaptation
of resource provisioning according to the
dynamic on-demand usage patterns of Internet users. While current
offerings implement principles such as agility, scalability, elasticity, rapid
self-service provisioning and resources virtualization, key issues related to
dependability have not yet been tackled. The latter is of major importance for
storage clouds, considering the enormous amount of data being produced in
today's rich digital environment (e.g. by smart phones, social networks,
sensors, user generated content, etc). VISION
Cloud aims at providing a flexible cloud environment addressing the
challenge of providing data-intensive storage cloud services through raising
the abstraction level of storage, enabling data mobility across providers,
allowing computational and content-centric access to storage and deploying new
data-oriented mechanisms for Quality of Service and security guarantees. The
presentation will focus on the VISION
Cloud approaches with respect to dependability issues (e.g. durability and
availability of the environment), which are being considered during the full
lifecycle of data-intensive services. Specific mechanisms will be discussed
(e.g. a resiliency manager that automatically restores the proper resiliency
for storage objects in the storage cloud despite failures) aiming at the
provision of quality guarantees for the cloud services.
·
15:30-16:00hs
Coffee Break
·
16:00hs
Keynote: Software Execution Protection
in the Cloud. Miguel Correia (INESC-ID/IST,
Portugal)
Most cloud
computing services execute software on behalf of their users. Many war stories
and several studies suggest that such software execution is threatened by
accidental arbitrary faults and malicious insiders. We present two lines of
work to protect software execution in the cloud. The first replicates tasks to
protect MapReduce executions from the effects of
accidental arbitrary faults. The second uses hardware-based security to protect
software execution from malicious insiders.
·
17:00hs
Workshop closing
WORKSHOP TOPICS
The workshop welcomes submissions on
all aspects of dependability in cloud computing, data center
architectures, scalable systems and virtualization technology. Some of the
topics of interest are:
·
Virtualization-based
mechanism for tolerating faults and intrusions
·
Dependability
issues in virtualization technology
·
Scalability
issues in dependable systems
·
Security and
privacy in the Cloud
·
Data center dependability issues
·
Disaster
recovery for cloud-based and Internet-scale systems
·
Cloud
dependability metrics, models, studies and benchmarks
·
Large scale
problem diagnosis and recovery
·
Intrusion
detection and tolerance in cloud environment
·
Service-oriented
architectures dependability
·
Dependability
issues in "cloud programming models" such as MapReduce
·
Service
Level Agreement (SLA) and Quality of Service in the cloud
More information about the workshop
can be obtained by emailing to ewdcc12_AT_di.fc.ul.pt.
SUBMISSION INSTRUCTIONS
The workshop will accept two formats of papers: regular papers (maximum
6 pages) and position papers (maximum 2 pages). Position papers allow
researchers to present more speculative/futuristic ideas to stimulate
discussion and further work. All papers will be published at the ACM digital
library and thus have to adhere ACM 8.5”x11” two-column camera-ready format.
The LaTeX and Word templates can be found here.
Each paper should be submitted as a
single PDF file through the submission website: https://www.easychair.org/account/signin.cgi?conf=ewdcc2012.
At least one author of an accepted paper must register at the conference
and present the paper at the workshop.
BEST PAPER
Besides the publication of the workshop proceedings at the ACM digital
library, the authors of the best papers of the workshop will be invited to
submit a revised (and maybe extended) version of the papers to SIGOPS Operating
System Review (http://www.sigops.org/osr.html).
IMPORTANT DATES
(EXTENDED) Submission deadline: February
2, 2012
Author notification: March 14, 2012
Final version: March 20, 2012
WORKSHOP ORGANIZERS
Alysson
Bessani, University of Lisbon, Portugal
Rüdiger
Kapitza, TU Braunschweig,
Germany
PROGRAM COMMITTEE
Christian Cachin,
IBM Zurich, Swiss
Brendan Cully, University of British
Columbia, Canada
Sameh Elnikety,
Microsoft Research, USA
Ilir Gashi,
City University London, UK
Henrique Moniz, Microsoft Research,
UK
Kaustubh Joshi, AT&T Labs Research, USA
José
Orlando Pereira, Univesity of
Minho, Portugal
Marcelo Pasin,
University of Lisbon, Portugal
Dana Petcu,
Western University of Timisoara, Romania
Peter Pietzuch,
Imperial College London, UK
Hans P. Reiser,
University of Passau, Germany
Matthias Schunter,
IBM Zurich, Swiss
Marco Serafini,
Yahoo! Research Barcelona, Spain
Johan Tordsson,
Umeå Universitet, Sweden
José Luis Vázquez-Poletti,
Complutense University of Madrid, Spain
Marco Vukolic,
Eurecom, France
SPONSORS
