http://www.cs.duke.edu/nicl/acdc09

June 19^th, 2009, Barcelona, Spain

In conjunction with the 6^th Intl. Conference on Autonomic Computing and Communications (ICAC 2009)

SCOPE:

Large server farms—from enterprise datacenters and giant-scale Internet services to utility and public cloud computing—are a critical part of the execution platform for many of today’s software programs. Interactive web services, business workflows, large-scale batch data processing, and many scientific computations all run on platforms built atop large server farms. While these systems differ in many ways, they share many common challenges, such as the desire to maximize efficiency and performance while maintaining predictable behavior, power efficiency and fairness, and at the same time responding appropriately to environmental and system changes such as hardware failures and changes in workload.

We believe that solutions to these challenges in performance, reliability and scalability will share many similarities across all variants of server farm infrastructures. In particular, we believe automated techniques will play a critical role in reasoning about and orchestrating the behaviors of these large scale systems. The goal of this workshop is to broadly explore the similarities and differences in current and future policies and techniques in the context of these large server farms.

Important Dates:

Submission deadline: February 27^th, 2009

Notification of acceptance: March 27^th, 2009

Camera-ready deadline: April 6^th, 2009

Workshop Organizers:

Program Co-chairs

Jeff Chase, Duke University

Emre Kiciman, Microsoft Research

Rich Wolski, UC Santa Barbara

Program Committee

Katerina Argyraki, EPFL

Franck Capello, INRIA

Renato J. Figueiredo, University of Florida

Marty Humphrey, University of Virginia

Randy Katz, UC Berkeley

Mike Kozuch, Intel Labs Pittsburgh

Milan Milenkovic, Intel

Partha Ranganathan, HP Labs

Douglas Thain, University of Notre Dame

John Wilkes, Google

Topics and Submissions:

We invite authors to submit short position papers or reports of early work related to the automated control of datacenters and clouds. Relevant topics for submission include, but are not limited to:

· Autonomic policies and techniques for improving reliability, performance, scalability and power efficiency of large server farms

· Separation of application concerns from infrastructure, resource, power and other management issues

· Change management and adaptation techniques

· Experience reports on challenges and solutions in large-scale internet services, cloud computing infrastructures, enterprise datacenters and other server farms.

· Experience reports on requirements of high performance data computing, scientific computing, data processing and other applications using large server farms

· Experiences with automatic control strategies, including strategies based on rule systems, strategies economic models, metering-based, control

· Reports on key similarities and difference in autonomic requirements across classes of server farm infrastructures

· Management of network infrastructures within server farms and between server farms

VTDC’09 will be held June 15, 2009, in Barcelona, Spain, in conjunction with ICAC’09 (The International Conference on Autonomic Computing and Communications)

http://www.grid-appliance.org/wiki/index.php/VTDC09

Important dates

• Submission deadline: February 20th, 2009
• Notification of acceptance: March 23rd, 2009
• Final manuscripts due: April 6, 2009
• Workshop: June 15, 2009

Workshop scope

The convergence of virtualization technologies and distributed computing is an exciting development and the subject of much research in both academia and industry, where interest in Cloud Computing environments as a means of provisioning Infrastructure-as-a-Service (IaaS) has grown considerably in the recent past. The VTDC workshop is intended to be a forum for the exchange of ideas and experiences on the use of virtualization technologies in distributed computing, the challenges and opportunities offered by the development of virtual systems themselves, as well as case studies of application of virtualization. The scope of “virtualization technologies” includes techniques and concepts to enable virtual machines, virtual networks, virtual data, virtual storage, virtual applications and virtual instruments. The scope of “distributed computing” includes Grid-computing, cluster computing, peer-to-peer computing and mobile computing.

The VTDC workshop brings together researchers in academia and industry who are involved in research, development and planning activities involving the use of virtualization in the context of distributed systems – such as computing clouds and Grids - where the opportunities and challenges with respect to the management of such virtualized systems is of interest to the ICAC community at large.

Topics

VTDC 2009 topics of interest include, but are not limited to:

• Infrastructure as a service (IaaS)
• Virtualization in data centers and cloud computing infrastructures
• Using virtualization technologies for resource management and QoS assurance
• Security aspects of using virtualization in a distributed environment
• Virtual networks
• Virtual data and storage systems
• Fault tolerance in virtualized environments
• Virtualization in P2P systems
• Virtualization-based adaptive/autonomic systems
• Virtual environment factories and services
• Environment configuration
• Virtual machine management
• Modeling (applications and systems)
• Case studies of applications using virtual technologies
• Deployment studies of virtualization technologies
• Tools relevant to virtualization

Organizers

• General Chair: Kate Keahey, University of Chicago, Argonne National Laboratory
• Program Chair: Renato Figueiredo, University of Florida
• Steering Committee Chair: Jose A. B. Fortes, University of Florida
• Program Committee: TBD

Barcelona, Spain, 15- 19 June 2009

http://icac2009.acis.ufl.edu/

IMPORTANT DATES

Full paper:  11:59 EST, Jan 19, 2009
Workshop proposals: Sep 25, 2008
Demo/Exhibit proposals: Mar 20, 2009

SCOPE

To deal with the increasing complexity of large-scale computing systems, computers and applications must learn to manage themselves in accordance with high-level guidance from humans – a vision that has been referred to as autonomic computing. Meeting the grand challenges of autonomic computing requires scientific and technological advances in a wide variety of fields, as well as new software and system architectures that support the effective integration of the constituent technologies.
The purpose of the 6th International Conference on Autonomic Computing and Communications (ICAC-09) is to bring together researchers and practitioners addressing aspects of self-management in computing systems. In doing so, we hope to develop and nurture a community that can work together to realize the vision of large-scale self-managing systems. The conference builds on previous highly influential meetings in New York, Seattle, Dublin, Jacksonville, and Chicago.
Papers are solicited on a broad array of topics of relevance to autonomic computing, particularly those that bear on connections and relationships among different areas of research or report on prototype systems or experiences. Topics of interest include, but are not limited to:
* Autonomic computing systems or prototype systems that exhibit self-monitoring, self-configuration, self-optimization, self-healing, and/or self-protection.
* Fundamental science of self-managing systems: understanding, controlling, or exploiting emergent behavior, fault-tolerance, machine learning, control theory, predictive methods and their use to automate manual operations and enforce behavior.
* Software engineering principles and architectures for self-managing systems, based on interoperable Grid Services, agent-based systems, Web Services, model-based systems or novel paradigms such as biological, economic or social.
* System-level technologies, middleware or services that entail interactions among two or more components of self-managing systems in standalone, distributed, cluster, and Grid computing environments (e.g., health monitoring, dependency analysis, problem localization or remediation, workload management, and provisioning).
* Toolkits, environments, models, languages, runtime and compiler technologies for building self-managing components, systems or applications.
* Specific self-managing components, such as server, storage, network, data center or specific application elements. Emphasis should be placed on techniques or lessons that may generalize to other components.
* Management topics, such as specification and modeling of service-level agreements, negotiation/conversation support, behavior enforcement, etc., tie in with IT governance, and interaction with legacy systems.
* Interfaces to autonomic systems, including user interfaces, interfaces for monitoring and controlling behavior, techniques for defining, distributing, and understanding policies.
* Experiences with autonomic system or component prototypes: measurements, evaluations, or analyses of system behavior, user studies, experiences with large-scale deployments of self-managing systems or applications.
* Applications of autonomics to real and complex problems in science, engineering, business and society.

ORGANIZATION

GENERAL CO-CHAIRS:
Simon Dobson, UCD Dublin, IE
John Strassner, Waterford Inst. of Tech., IE

STEERING COMMITTE:
Simon Dobson, UCD Dublin, IE
José Fortes, Univ. of Florida, US
Kumar Goswami, HP, USA
Salim Hariri, Univ. of Arizona, US
Jeffrey Kephart, IBM, USA
Manish Parashar, Rutgers Univ., US
Brent Miller, IBM, US
Karsten Schwan, Georgia Tech, US
John Strassner, Waterford Inst. of Tech., IE
John Wilkes, Google, US
Mazin Yousif, Avirtec, US

PROGRAM CO-CHAIRS:
Manish Parashar, Rutgers Univ., USA
Onn Shehory, IBM Research Haifa, IL

PROGRAM COMMITTEE:
Tarek Abdelzaher, Univ. of Illinois U-C, US
Richard Anthony, Univ. of Greenwich, UK
Ozalp Babaoglu, Univ. of Bologna, IT
Liz Bacon, Univ. of Greenwich, UK
Rosa Badia, UPC, ES
Claudio Bartolini, HP Labs, US
Ricardo Biancini, Rutgers Univ., US
Frances Brazier, Vrije Universiteit Amsterdam, NL
David Breitgand, Univ. of Haifa, IL
Marcus Brunner, NEC, DE
Mark Burgess, University College Oslo, NO
Fabián Bustamante, Northwestern Univ., US
Hakima Chaouchi, IT Sud Paris, FR
Luca Console, Universita’ di Torino, IT
Alva Couch, Tufts Univ., US
Michael Covington, Intel, US
Spyros Denazis, U. Patras, GR
Yixin Diao, IBM Corp., US
Peter Dinda, Northwestern Univ., US
Renato Figueiredo, Univ. of Florida, US
Alex Galis, UCL, UK
Holger Giese, HPI/Univ. of Potsdam, DE
Nathan Gnanasambandam, Xerox, US
Richard Golding, IBM Corp., US
Mike Hinchey, Lero, IE
Brendan Jennings, TSSG/WIT, IE
Guofei Jiang, NEC, US
Gail Kaiser, Columbia Univ., US
Carlos A. Kamienski, UFABC, BR
Nagarajan Kandasamy, Drexel Univ., US
Emre Kiciman, Microsoft Research, US
Zhiling Lan, Illinois Inst. of Tech., US
Rogerio de Lemos, Univ. of Kent, UK
Dave Lewis, Trinity College Dublin, IE
Leonardo Mariani, Univ. of Milano Bicocca, IT
Ignacio Martin Llorente, Universidad Complutense de Madrid, ES
Julie McCann, Imperial College, UK
Phillip K. McKinley, Michigan State Univ., US
Nenad Medvidović, Univ. of Southern California, US
Hausi A. Müller, Univ. of Victoria, CA
Jorg Muller, TU Clausthal, DE
Julian Padget, Univ. of Bath, UK
Van Parunak, NewVectors, US
Marcelo Perazolo, IBM Corporation, US
Mauro Pezzè, Univ. of Milano Bicocca, IT
Mohammed A. Razzaque, UCD Dublin, IE
Jacques Sauve, U. F. Campina Grande, BR
Daniel Scheibli, SAP Corp., US
Richard Schlichting, AT&T, US
Giovanna di Marzo Serugendo, Univ. of London, UK
Mikhail Smirnov, Fraunhofer FOKUS, DE
Jordi Torres, UPC, ES
Roy Sterritt, Univ. Of Ulster, UK
Davide Tosi, University of Milan, IT
Giuseppe (Peppo) Valetto, Drexel Univ., US
Sven van den Meer, Waterford IT, IE
Aad van Moorsel, Newcastle Univ., UK
George Vouros, Univ. of the Aegean, GR
Dongyan Xu, Purdue Univ., US
Franco Zambonelli, Universita di Modena e Reggio Emilia, IT
Xiaoyun Zhu, VMware, US

DEMO/EXHIBIT CO-CHAIRS:
Simon Dobson, UCD Dublin, IE
John Strassner, Motorola Labs, US

WORKSHOP CHAIR:
Omer Rana, Cardiff Univ., UK

PUBLICITY CO-CHAIRS:
Masoud Sadjadi, Florida International Univ., US
James Won-Ki Hong, POSTECH, South Korea
Dave Lewis, Trinity, IE

LOCAL ARRANGEMENTS CHAIR:
Joan Serrat, UPC, Barcelona, ES

FINANCE/INDUSTRY CO-CHAIRS:
Brent Miller, IBM Corporation, US
Michael Nunez, Sun Microsystems, US
Jose A. Lozano, Telefónica, ES

HOT TOPICS CHAIR:
Fabián E. Bustamante

DOCTORAL SYMPOSIUM CO-CHAIRS:
Jian Zhang, Microsoft, US
Andres Quiroz Hernandes, Rutgers Univ., US

CYBER CHAIR:
Ming Zhao, Florida International Univ, US

Barcelona, Spain, 15 June 2009
http://autonomic.ac.upc.edu/GRIDmeetsAC/
In association with
the 6th International Conference on Autonomic Computing & Communications
http://icac2009.acis.ufl.edu/

IMPORTANT DATES
Full paper: 23 February 2009
Author notification: 23 March 2009
Final manuscripts: 6 April 2009

SCOPE
Scientific communities worldwide have constructed massive grids that contain several tens of thousands of CPU’s and several PetaBytes of storage. The complexity of grids has increased due to the large range of dynamic resources involved: hardware and middleware, but also software usage rights, sensors, etc.  Grid applications taking advantage of those resources have also increased in complexity.  In addition, grid usage is based on a sharing paradigm, which introduces the collective behaviour of users as a new complicating factor. A major technical and societal challenge for the immediate future is the stabilisation of production grids.  The control and maintenance of these complex systems to achieve that remain a significant operational issue. Application optimization is similarly challenging.
Autonomic Computing has emerged as a specific scientific domain, with the strong involvement of industry. Autonomic computing is highly relevant to grid systems at a time when production grids have become critical scientific infrastructures and must move to sustainable models. Considering the volume and structure of the manpower dedicated to the day-to-day operations, Self-Optimization, Self-Healing and Self-Configuration could provide immediate benefits.
At the core of the autonomic computing is the fact that high-level goals should be exposed by the middleware and should be easily tuned by users and system administrators. The Workshop will contribute to building bridges between grid researchers on one hand and those in the autonomic computing community on the other. The focus of the workshop is to identify key scientific challenges related to the management and evolution of grids as a specific category of complex large-scale systems. The goal of the workshop is to promote community wide discussion of, and collaboration on potentially high-impact ideas that will influence and foster continued research in improving the manageability and reliability of grids.
The workshop will feature invited presentations from major grid projects and Autonomic Computing scientists as well as regular papers and posters. The format will leave ample time for discussion

PAPER/POSTER SUBMISSION AND PUBLICATION
Full papers (a maximum of 8 pages in length) and posters (2 pages) are invited. All manuscripts will be reviewed and judged on merits including correctness, originality, quality of presentation, and relevance to the workshop themes. Please see the conference web site for more information on the submission process. It is expected that the proceedings will be published by ACM.

ORGANISATION
Chair
Cécile Germain-Renaud LRI, France

Program Committee
Mark Burgess Oslo U.,    Norway
Ricard Gavalda UPC, Spain
Charles Loomis LAL, France
Manish Parashar  Rutgers U.., USA
Ron Perrot Queen’s U. Belfast
Omer Rana, Cardiff Univ., UK
Lorenza Saitta UPO, Italy
Jordi Torres UPC, Spain
Ramin Yahyapour Dortmund U., Deutschland

Virtualize the servers is not an end point. There are other parts of the IT infraestructure beyond servers in order to enable a more flexible and efficient IT environment: the network. Nowadays network components are also moving toward being virtualized enabling a dynamic virtual environments.

There are already some products in the market that show how it is not just servers that can be virtualized and illustrates how the days of static infraestructure networks will soon be over. For example HP recently introduced the HP’s Virtual Connect architecture,  that enables the user to add more NICs to each server blade without more hardware and provides the ability to dynamically adjust bandwidth for each network connection on the fly. HP product details are posted here. Some initial steps towards the integration of virtualization of servers and networks  can be found in the Citrix XenServer based virtualization solutions. HP integrated Citrix XenServer with HP ProLiant Virtual Console provides a straightforward deployment and management environment that speeds adoption of flexible, cost-saving virtualization technology.

It is true that this is a great step towards a more Green IT. Let’s wait and see how things develop, but  the way I see it, there is still mising an intelligent layer that could manage the complexity that virtualization poses and use resources efficiently.  For example, those virtualized server (acording their workloads) should be able to talk intelligently and dinamically with the network infrastructure if we what to optimize the use of the resources in the Cloud.  My vision is that the solution for this important challenge of building more sustainable and efficient IT is still open and only could have to come from cross-disciplinary studies over a diverse set of research areas.

La virtualització dels servidors i la  xarxa alhora: És necessària  una capa intel.ligent que gestioni la nova complexitat.

La virtualització no s’acaba en la virtualització dels servidors. Hi ha altres parts de les infraestructures TIC  més enllà dels servidors a tenir en compte per tal de permetre entorns TIC més eficients i eficients, la xarxa per exemple. Avui en dia els components de la xarxa també s’estan movent cap a la virtualització permetent entorns virtuals dinàmics.

En aquest sentit ja podem trobar alguns productes en el mercat que mostra que no només els servidors poden ser virtualitzats i il.lustren com li queden els dies comptats a les infraestructures estàtiques de xarxa. Per exemple l’empresa HP recentment ha presentat la seva Arquitectura virtual de connexió HP,  que permet a l’usuari afegir més NIC (Network Interface Connector, part del hardware que permet connectar a una xarxa) a cada servidor sense cap hardware addicional i disposa de l’habilitat d’ajustar dinàmicament l’ample de banda per cada connexió en temps d’execució. Detalls dels productes HP poden trobar-se postejats aquí. Alguns dels passos inicials cap a la integració de la virtualització de servidors i xarxes alhora poden trobar-se postejats en Citrix XenServer based virtualization solutions. HP ha integrat en Citrix XenServer amb HP ProLiant Virtual Console que permet  una senzilla implantació de la gestió de l’entorn que permet aquesta flexibilitat derivada de la virtualització.

Certament aquest és un pas important cap a una computació més sostenible (Green Computing). Esperem a veure com avancen les coses, però tal com jo ho veig encara cal la intel.ligència de tot plegat, és a dir, cal un component que permeti gestionar la complexitat que la virtualització afegeix per tal de fer un ús eficient dels recursos.  Per exemple, els servidors virtualitzats (en base a la seva càrrega) han de poder comunicar-se de manera intel.ligent i dinàmica amb la infraestructura de xarxa si volem optimitzar l’ús dels recursos del “Cloud”. La meva visió és que la solució per aquest important repte de disposar de sistemems TIC més sostenibles i eficients és troba encara per resoldre i passa forçosament  per trobar el punt de trobada entre diferents arees de recerca.

Green computing (or Green IT) is the study and practice of using computing resources efficiently. This term generally relates to the use of computing resources in conjunction with minimizing environmental impact, maximizing economic viability and ensuring social duties. Some companies uses green computing philosophies mainly to save up on costs rather than save the environment. This green computing concept emerged naturally as businesses find themselves under pressure to maximize resources in order to compete effectively in the market. Aside from minimizing costs, other companies also takes into account other factors such as marketing and branding.

One of the first manifestations of the green computing movement was the launch of the Energy Star program back in 1992. This resulted in the widespread adoption of sleep mode among consumer electronics. Moreover turning off the monitor when it’s not in use or using more energy efficient monitors like LCDs instead of the traditional CRT monitors, as the concept developed, green computing began to encompass a lot of different initiatives at different levels of the IT world.

Nowadays, some of the important green computing practices rely on a more efficient data centers (that uses more efficient hardware, considering the consolidation of servers using virtualization, or using more energy efficient and less noisy cooling systems) and more efficient networks systems (with more efficient components, or increased online security measures through the use of firewalls or antivirus programs to reduce the increasing amount of useless data/work).

Our group is currently doing research in this area contributing in a more efficient data centers (e.g. recent publications [1] and [2] ) and a more efficient networks systems (e.g. recent publications [3] and [4]) base in Autonomic Computing research concepts.

[1] Jordi Torres, David Carrera, Kevin Hogan, Ricard Gavaldà, Vicenç Beltran and Nicolas Poggi. Reducing wasted resources to help achieve green data centers. In The Fourth Workshop on High-Performance, Power-Aware Computing (HPPAC 2008), April 14, 2008.

[2] Jordi Torres, David Carrera, Vicenç Beltran, Nicolás Poggi, Josep Lluis Berral, Ricard Gavaldà, Eduard Ayguadé, Toni Moreno, Jordi Guitart. Tailoring resources: Energy efficient consolidation strategy goes beyond virtualization Proceedings of ICAC 2008, The IEEE International Conference on Autonomic Computing, poster session. Chicago, Illinois, USA, June 2008

[3] N. Poggi, J.L. Berral, T. Moreno, R. Gavaldà, J. Torresé. Automatic Detection and Banning of Content Stealing Bots for E-commerce. NIPS 2007 Workshop on Machine Learning in Adversarial Environments for Computer Security. December 8, 2007. British Columbia, Canada.

[4] Josep L. Berral, Nicolas Poggi, Javier Alonso, Ricard Gavalda, Jordi Torres and Manish Parashar. “Adaptive Distributed Mechanism Against Flooding Network Attacks Based on Machine Learning”. ACM Workshop on Articifial Intelligence on Security 2008, October 27th, 2008, Virginia (USA).

Last week Ramon Nou did the PhD presentation obtaining the maximum mark. The thesis introduces  real-time simulation for self-managed grid middleware.  Part of the work have been done in collaboration with  Samuel Kounev and is framed in the BSC research area.

Congratulation to Ramon for their excellent work.

Here you can find the PDF file of the PhD dissertation (in Spanish).

Last week Vicenç Beltran did the PhD presentation obtaining the maximum mark. The unstoppable growth of the World Wide Web requires a huge amount of computational resources that must be used efficiently. Nowadays, commodity hardware is the preferred platform to run web server systems because it is the most cost-effective solution. The work presented in this thesis aims to improve the efficiency of current web server systems, allowing the web servers to make the most of hardware resources. To this end, we first characterize current web server system and identify the problems that hinder web servers from providing an efficient resource utilization. From the study of web servers in a wide range of situations and environments, we have identified two main issues that prevents web servers systems from efficiently using current hardware resources. The first is the extension of the HTTP protocol to include connection persistence and security, which dramatically impacts the performance and configuration complexity of traditional multi-threaded web servers. The second is the memory-bounded or disk-bounded nature of some web workloads that prevents the full utilization of the abundant CPU resources available on current commodity hardware. We propose two novel techniques to overcome the main issues with current web server systems. Firstly, the Hybrid web server architecture that can be easily implemented in any multi-threaded web server to improve the CPU utilization with better management of client connections. And secondly, a main memory compression technique implemented in the Linux operating system that makes the most of current multiprocessor’s hardware to improve the performance of memory bounded web applications. The thesis work is supported by an exhaustive experimental evaluation that proves the effectiveness and feasibility of our proposals on actual systems. It is worth noting that the main concepts behind the Hybrid architecture have recently been implemented in popular web servers like Apache, Tomcat and Glassfish.

The work was coadvised by Eduard Ayguadé and myself. Congratulation to Vicenç for their excellent work.

Here you can find the PDF file of the PhD dissertation.

It’s early days but I must say that we are looking forward to the next ICAC conference already. This conference has consistently uncovered important discoveries in the area of AC in all of its previous installments (New York, Seattle, Dublin, Jacksonville and Chicago). The next meeting is going to take place in June 15-19 2008 and is going to be happening in Barcelona.

More information can be obtained from the official web site: http://www.acis.ufl.edu/~icac2009/

For those that are interested in the Autonomic Computing research and also members of LinkedIn Javier Alonso thought we’d share a LinkedIn Group.

Last week David Carrera did the PhD presentation obtaining the maximum mark. The thesis contributes to the performance management of a complex application server execution environment using autonomic computing issues. The work was coadvised by Eduard Ayguadé and myself and is framed in a collaboration with one IBM Research Group at Watson Labs (leaded by Malgorzata Steinder) and BSC.

Congratulation to David for their excellent work.

Here you can find the PDF file of the PhD dissertation.

Vas is a Technical lead and the Virtualization Strategic Initiative Manager in the Virtualization Runtime & Tools Department. Vas presented us with an interesting talk about Virtualization technology and how it is driving profound changes in the way large data centers are designed and managed. Moving to a virtualized infrastructure solves many problems such as hardware utilization and power consumption. But it also creates many new problems such as virtual machine image sprawl and a new management layer in the IT stack. The talk included an interesting overview of the emerging trends in the industry by the adoption of virtualization, followed by an IBM Research perspective of what new technical challenges this creates.

Companies are now focusing more than ever on the need to improve energy efficiency. In addition to the cost of energy, a new challenge for them is the increasing social pressure to reduce their carbon footprint. Commercial electricity consumption is a major factor in rising atmospheric CO2 levels and data centers are a significant part of the problem. While energy costs are rising and data center equipment is stressing the power and cooling infrastructure, the main issue is not the current amount of data center emissions, but the fact that data center emissions are increasing faster than any other carbon emission. For this reason nowadays there is a big interest in “Green” data centers and supercomputer centers. In this scenario, the research community is being challenged to rethink data center strategies, and add energy efficiency to a list of critical operating parameters that already includes service ability, reliability and performance.
Consolidation and virtualization can be combined to reduce the management complexity of large data centers as well as to increase the energy efficiency of such a system. But even in a scenario where the resources are consolidated and virtualized, utilizing all the capacity of the components that are switched on (and consuming power) is not always simple. To determine a set of applications to be collocated in a node to perfectly fit and exploit all the resources of the system is a hard problem to solve, especially when tenths or even hundreds of nodes and applications can be found in a data center. Furthermore, the fact that the demand associated with each system resource for a given application may not be related in any way to its demand for other resources (i.e. an application with a large memory footprint may not be very demanding in terms of CPU power) which creates a structural problem requiring constraints to be relaxed in order to overcome it.

To solve this problem we have a poster in the ICAC conference held in Chicago last June.
In this paper we present how these two simple and wellknown techniques can be combined to dramatically increase the energy efficiency of a virtualized and consolidated data center. Increased energy efficiency is obtained through the introduction of a new approach to the consolidation strategy by combining: memory compression and request discrimination. Combining these techniques enables an important reduction in the amount of active nodes required to process a web workload by dynamically classifying and shaping the workload, without degrading the offered service level. Furthermore, when the system eventually gets overloaded and no energy can be saved without loosing performance, we show how request discrimination can still improve the overall value obtained from the workload. The two techniques were separately studied and validated in a previous work to be now combined in a joint effort. Memory compression is used to convert CPU power into extra memory capacity to overcome system underutilization scenarios caused by memory constraints. Request discrimination is used to characterize web clients by predicting their class and the value they will have on the system. Our experiments performed on a real workload, obtained from a top national travel service exemplify the dramatic improvement these thechniques offer in energy and performance efficiency.
The main contribution of this article is to demonstrate that the consolidation of dynamic workloads does not end with virtualization, but there is even more to consolidate when energy-efficiency is pursued. We will present two alternatives to rescue resources that consolidation does not currently capitalize on while using virtualization.

link to the poster - link to the extended version of the paper.

A new paper in cooperation with a research team from IBM Research at Watson Lab in NY have been accepted in one of the important conferences in the area: IEEE/IFIP Network Operations and Management Symposium (NOMS 2008). The paper is related with the energy efficient computing, and study the problem of dynamic resource allocation to clustered Web applications.

Utility-based Placement of DynamicWeb Applications with Fairness Goals
David Carrera, Malgorzata Steinder, Ian Whalley, Jordi Torres, and Eduard Ayguadé

Abstract. We study the problem of dynamic resource allocation to clustered Web applications. We extend application server middleware with the ability to automatically decide the size of application clusters and their placement on physical machines. Unlike existing solutions, which focus on maximizing resource utilization and may unfairly treat some applications, the approach introduced in this paper considers the satisfaction of each application with a particular resource allocation and attempts to at least equally satisfy all applications. We model satisfaction using utility functions, mapping CPU resource allocation to the performance of an application relative to its objective. The demonstrated online placement technique aims at equalizing the utility value across all applications while also satisfying operational constraints, preventing the over-allocation of memory, and minimizing the number of placement changes. We have implemented our technique in a leading commercial middleware product. Using this real-life testbed and a simulation we demonstrate the benefit of the utility-driven technique as compared to other state-of-the-art techniques.

Two weeks ago, Luis Moura Silva and Javier Alonso were in Boston (USA) to assist in the 6th IEEE International Symposium on Network Computing and Applications (IEEE NCA07). They presented the paper “Using Virtualization to improve the Software Rejuvenation” (link: paper_ieee_nca_camera_ready.pdf ) written in collaboration with Prof. Luis Silva (Univ. Coimbra, Portugal) and Dr. Artur Andrzejak (ZIB, Germany). Furthermore, this paper was awarded the “Best paper award on Network Applications”. The paper presents an approach for software rejuvenation based on automated self-healing techniques that can be easily applied to off-the-shelf Application Servers and Internet sites. This self-healing scheme is meant to be the least disruptive as possible for running the service and results in zero downtime for most of the cases. In our scheme, we exploit the usage of virtualization to optimize the self-recovery actions. The results demonstrate that virtualization can be extremely helpful for software rejuvenation and fail-over in the occurrence of transient application failures and software aging. During the conference they have had the opportunity to meet different researchers and some potential collaboration can be opened in the short/middle term.

Last week Prof. Manish Parashar (Professor of Electrical and Computer Engineering at Rutgers University , New Jersey) was in Barcelona invited by our research group.

During his stay we discussed various themes of research in the area of Autonomic Computing and we have detected some possible areas for collaboration that should be explored from now on. On top of that, I’m thankful for the opportunity that Professor Parashar has given us to compare and contrast with him what the future of Autonomic Computing will be. We agreed that the discipline is cross-sectional and that in the near future it will become a part of many areas of current research.

During his stay Professor Parashar also gave a course, as a continuation of Omer F. Rana’s, in the area of Autonomic Computing. In the same way as the course from Omer F. Rana, it was taught as part of the new Masters of Research, “Computer Architecture, Networks and Systems” (CANS), from the Computer Architecture department in UPC. The title of the course was “Autonomic Computing: Concepts, Infrastructures and Applications” and was divided in three parts: (a) Overview, Motivations, Concepts (b) Approaches and Infrastructures (c) Realizations and Applications.

Like the course from Omer F. Rana, this course was impressive and the students enjoyed it very much with him. Whatsmore the students could contrast with Parashar some of his ideas on their respective doctorate works.

If you want more information about Professor Manish Parashar you can visit his web page at http://www.caip.rutgers.edu/~parashar/

During last week Omer F. Rana from Cardiff University (U.K.) was hosted by our research group. Prof. Rana also gave a course in autonomic computing during his stay. The course was taught as part of the new Masters of Research, “Computer Architecture, Networks and Systems” (CANS), from the Computer Architecture department in UPC. The course was exciting and it’s safe to say the students enjoyed the excellent presentation and examples that Prof. Rana presented. For example, do you know what “Stigmergetic Behaviour” is? Or do you know the waggle and tremble dances of bees? Incredible course. If you have the opportunity to attend future editions don’t miss it! If you want more information about Professor Rana you could visit his web page at http://users.cs.cf.ac.uk/O.F.Rana.

Today Malgorzata Steinder from IBM T. J. Watson Research Center gived an excellent talk “Reducing systems management complexity using server virtualization technology”. Margorzata Steinder is leading the team at IBM T.J. Watson Research Center that is collaborating with our research group at BSC under the Adaptive Systems BSC-IBM SOW. We are very happy to have Gosia with us in Barcelona during this two days.

Gosia presented us how server virtualization opens up a range of new possibilities for datacenter management, through the availability of new automation mechanisms that can be exploited to manage software within virtual machines. Server virtualization permits innovative ways of software distribution inside pre-installed, pre-wired, and pre-configured freeze-dried disk images. Server virtualization also enables powerful and flexible autonomic control, through management software that maintains the system in a desired state in the face of changing workload and demand.

Nowadays, business and scientific organizations have a large amount of critical workflow processes that depend upon a set of heterogeneous applications. This set of applications can range from transactional applications, with databases, to non-interactive applications such as those that perform scientific CPU-intensive computing, document indexing or intensive I/O. In order to efficiently execute all these kinds of applications, a new execution environment is required to manage the available resources and simplify the development and integration of different types of applications and services. These middleware components are of crucial importance for new systems and architectures because of their impact on the quality of the service offered by the application.

The EEDC 2007 workshop provides an open forum for students to present discuss issues, directions, and results that will shape the future of execution environments for High-Performance Distributed Systems.

The EEDC 2007 workshop takes a broad view of the Execution Environments for High-Performance Distributed Systems area and solicits contributions from many fields, including, but not limited to, new middleware architectures, management and control of self-* middleware, replication and file management, benchmarking and performance measurement, distributed technologies, interaction of hardware and software development. We particularly encourage contributions containing highly original ideas, new approaches, and/or groundbreaking results.

Proceedings: First Workshop on Execution Environments for Distributed Computing. 68 pages. Lulu Ed. 2007. ISBN: 978-1-84799-878-1

During this April two important researchers in this research areas have been hosted by our group: Dr. Mark Baker from Reading University and Samuel Kounev from Cambridge University. Both of them gave a course in topics related with this research area. The overall goal of Mark Baker course was to introduce Web Services and it related technologies. In the other hand Samuel Kounev taught the course “Performance Engineering of Distributed E-Business Systems”.  Both ourses were taught inside the new Master of Research “Computer Architecture, Networks and Systems” (CANS) from the Computer Architecture department of the UPC. Both courses are coordinated with the course (Execution Environments for Distributed Computing - EEDC) that our research group give in the same master. Our research group are committed to give to the students the oportunity of learn the knowledge related to this amazing area giving a wide sight of subjects that will cover it.

The majority of posts in this category are to help my students and introduce them to the topics related with the Masters course “Execution environment for Distributed Computing ” of CANS Master at UPC.

This course begins the new Master of Research in Computer Architecture, Networks and Systems - CANS as a result of the reform of university courses at UPC within the framework of the European Higher Education Area (EHEA). Our research group participates activelly in the research course Execution Environments for Distributed Computing. The course focuses on conceptual and practical aspects related with the performance of the middleware that support the execution environments of current High-Performance Distributed Systems.

Congratulations to the organizers (Simon Dobson, Fabrice Saffre and their team), of ICAC 06 in Dublin (June 12th-15th) for a smoothly run conference. Also congratulations to Mazin Yousif and Omer F. Rana that have efficiently produced an outstanding program for us. And also thanks to Jeffrey Kephart and Manish Parashar for their good job as general co-chairs. There were a great selection of interesting papers, lots of interaction and discussion throughout.

And most surprisingly of all the weather stayed sunny for the whole week!

Here are some pictures of the demo stand set up by BSC for ICAC 06 in Dublin. We presented some of the work being carried out in our group: understanding and creating the core of autonomic middleware. There were some good comments and suggestions from the participants that we really appreciated.