=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

CALL FOR P= APERS

9th Worksh= op on Virtualization in High-Performance Cloud Computing (VHPC '14)

held i= n conjunction with Euro-Par 2014, August 25-29, Porto, Portugal

(Sprin= ger LNCS)

= =3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

Date: = August 26, 2014

= Workshop U= RL: http:/= /vhpc.org

Paper = Submission Deadline: June 9, 2014 (extended)

Confir= med Keynote Speakers:

= Ron Bright= well, Sandia National Laboratory

= Hobbes: Us= ing Virtualization to Enable Exascale Applications

= and=

= Helge Mein= hard, CERN

CA= LL FOR PAPERS

Virtua= lization technologies constitute a key enabling factor for flexible resourc= e

= management= in modern data centers, and particularly in cloud environments.

= Cloud prov= iders need to dynamically manage complex infrastructures in a

= seamless f= ashion for varying workloads and hosted applications, independently of

= the custom= ers deploying software or users submitting highly dynamic and

= heterogene= ous workloads. Thanks to virtualization, we have the ability to manage

= vast compu= ting and networking resources dynamically and close to the marginal<= /p>

= cost of pr= oviding the services, which is unprecedented in the history of scientific

= and commer= cial computing.

Variou= s virtualization technologies contribute to the overall picture in differen= t

= ways: mach= ine virtualization, with its capability to enable consolidation of multiple=

= under-util= ized servers with heterogeneous software and operating systems (OSes),

= and its ca= pability to live-migrate a fully operating virtual machine (VM) with a very=

= short down= time, enables novel and dynamic ways to manage physical servers;
OS-= level virtualization, with its capability to isolate multiple user-space

= environmen= ts and to allow for their co-existence within the same OS kernel,

= promises t= o provide many of the advantages of machine virtualization with high=
lev= els of responsiveness and performance; I/O Virtualization allows physical
NIC= s/HBAs to take traffic from multiple VMs; network virtualization, with its<= /span>
cap= ability to create logical network overlays that are independent of the

= underlying= physical topology and IP addressing, provides the fundamental

= ground on = top of which evolved network services can be realized with an

= unpreceden= ted level of dynamicity and flexibility; the increasingly adopted

= paradigm o= f Software-Defined Networking (SDN) promises to extend this

= flexibilit= y to the control and data planes of network paths. =C2=A0These technologies=

= have to be= inter-mixed and integrated in an intelligent way, to support

= workloads = that are increasingly demanding in terms of absolute performance,

= responsive= ness and interactivity, and have to respect well-specified Service-<= /p>

= Level Agre= ements (SLAs), as needed for industrial-grade provided services.

= Indeed, am= ong emerging and increasingly interesting application domains

= for virtua= lization, we can find big-data application workloads in cloud

= infrastruc= tures, interactive and real-time multimedia services in the cloud,

= including = real-time big-data streaming platforms such as used in real-time

= analytics = supporting nowadays a plethora of application domains. Distributed

= cloud infr= astructures promise to offer unprecedented responsiveness levels for=

= hosted app= lications, but that is only possible if the underlying virtualization

= technologi= es can overcome most of the latency impairments typical of current

= virtualize= d infrastructures (e.g., far worse tail-latency). What is more, in data

= communicat= ions Network Function Virtualization (NFV) is becoming a key

= technology= enabling a shift from supplying hardware-based network functions,

= to providi= ng them in a software-based and elastic way. In conjunction with

= (public an= d private) cloud technologies, NFV may be used for constructing the<= /p>

= foundation= for cost-effective network functions that can easily and seamlessly=

= adapt to d= emand, still keeping their major carrier-grade characteristics in terms

= of QoS and= reliability.

The Wo= rkshop on Virtualization in High-Performance Cloud Computing (VHPC)<= /p>

= aims to br= ing together researchers and industrial practitioners facing the challenges=

= posed by v= irtualization in order to foster discussion, collaboration, mutual exchange=

= of knowled= ge and experience, enabling research to ultimately provide novel

= solutions = for virtualized computing systems of tomorrow.

The wo= rkshop will be one day in length, composed of 20 min paper presentations,

= each follo= wed by 10 min discussion sections, and lightning talks, limited to 5=

= minutes. P= resentations may be accompanied by interactive demonstrations.

TOPICS=

Topics= of interest include, but are not limited to:

- Mana= gement, deployment and monitoring of virtualized environments

= - Language= -process virtual machines

= - Performa= nce monitoring for virtualized/cloud workloads

= - Virtual = machine monitor platforms

= - Topology= management and optimization for distributed virtualized applications

= - Paravirt= ualized I/O

= - Improvin= g I/O and network virtualization including use of RDMA, Infiniband, PCIe

= - Improvin= g performance in VM access to GPUs, GPU clusters, GP-GPUs

= - HPC stor= age virtualization

= - Virtuali= zed systems for big-data and analytics workloads

= - Optimiza= tions and enhancements to OS virtualization support

= - Improvin= g OS-level virtualization and its integration within cloud management

= - Performa= nce modelling for virtualized/cloud applications

= - Heteroge= neous virtualized environments

= - Parallel= virtualized - virtualization aware file systems

= - Network = virtualization

= - Software= defined networking

= - Network = function virtualization

= - Hypervis= or and network virtualization QoS and SLAs

= - Cloudbur= sting

= - Evolved = European grid architectures including such based on network virtualization<= /span>

= - Workload= characterization for VM-based environments

= - Optimize= d communication libraries/protocols in the cloud

= - System a= nd process/bytecode VM convergence

= - Cloud fr= ameworks and APIs

= - Checkpoi= nting/migration of VM-based large compute jobs

= - Job sche= duling/control/policy with VMs

= - Instrume= ntation interfaces and languages

= - VMM perf= ormance (auto-)tuning on various load types

= - Cloud re= liability, fault-tolerance, and security

= - Research= , industrial and educational use cases

= - Virtuali= zation in cloud, cluster and grid environments

= - Cross-la= yer VM optimizations

= - Cloud HP= C use cases including optimizations

= - Services= in cloud HPC

= - Hypervis= or extensions and tools for cluster and grid computing

= - Cluster = provisioning in the cloud

= - Performa= nce and cost modelling

= - Language= s for describing highly-distributed compute jobs

= - VM cloud= and cluster distribution algorithms, load balancing

= - Instrume= ntation interfaces and languages

= - Energy-a= ware virtualization

Import= ant Dates

Rollin= g Paper registration

= June 9, 20= 14 - Full paper submission (extended)

= July 4, 20= 14 - Acceptance notification

= October 3,= 2014 - Camera-ready version due

August= 26, 2014 - Workshop Date

TP= C

CHAIR<= /span>

Michae= l Alexander (chair), TU Wien, Austria

= Anastassio= s Nanos (co-chair), NTUA, Greece

= Tommaso Cu= cinotta (co-chair), Bell Labs, Dublin, Ireland

PROGRA= M COMMITTEE

= Costas Bek= as, IBM

= Jakob Blom= er, CERN

= Roberto Ca= nonico, University of Napoli Federico II, Italy

= Piero Cast= oldi, Sant'Anna School of Advanced Studies

= Paolo Cost= a, MS Research Cambridge, England

= Jorge Ejar= que Artigas, Barcelona Supercomputing Center, Spain

= William Ga= rdner, University of Guelph, USA

= Balazs Ger= ofi, University of Tokyo, Japan

= Krishna Ka= nt, Temple University, USA

= Romeo Kinz= ler, IBM

= Nectarios = Koziris, National Technical University of Athens, Greece

= Giuseppe L= ettieri, University of Pisa, Italy

= Jean-Marc = Menaud, Ecole des Mines de Nantes, France

= Christine = Morin, INRIA, France

= Dimitrios = Nikolopoulos, Queen's University of Belfast, UK

= Herbert Po= etzl, VServer, Austria

= Luigi Rizz= o, University of Pisa, Italy

= Josh Simon= s, VMware, USA

= Borja Soto= mayor, University of Chicago, USA

= Vangelis T= asoulas, Simula Research Lab, Norway

= Yoshio Tur= ner, HP Labs, USA

= Kurt Tutsc= hku, Blekinge Institute of Technology, Sweden

= Chao-Tung = Yang, Tunghai University, Taiwan

PA= PER SUBMISSION-PUBLICATION

Papers= submitted to the workshop will be reviewed by at least two

= members of= the program committee and external reviewers. Submissions

= should inc= lude abstract, key words, the e-mail address of the

= correspond= ing author, and must not exceed 10 pages, including tables

= and figure= s at a main font size no smaller than 11 point. Submission

= of a paper= should be regarded as a commitment that, should the paper

= be accepte= d, at least one of the authors will register and attend the

= conference= to present the work.

Accept= ed papers will be published in the Springer LNCS series - the

= format mus= t be according to the Springer LNCS Style. Initial

= submission= s are in PDF; authors of accepted papers will be requested

= to provide= source files.

Format= Guidelines:

= http://www.springer.de/comp= /lncs/authors.html

EasyCh= air Abstract Submission Link:

= https://www.e= asychair.org/conferences/?conf=3Deuropar2014ws

GENERA= L INFORMATION

The wo= rkshop is one day in length and will be held in conjunction with
Eur= o-Par 2014, 25-29 August, Porto, Portugal