Dear WG2 member,

Apologizes if you’d receive several copies.

The next meeting in Coimbra proposes WG sessions from Monday 8^th to Wednesday 10^th. A first approx. to the proposed agenda could be:

WG2 meeting in Coimbra (November, 8th)
———————————————–

* WG2 general presentation (15 min)

* Short presentations (working progess & existing collaborations) (1-1,5
hours)

* New collaborations Brainstorming (1 hour)

* WG2 next steps (15 min)

During the WG2 meeting we’ll present the advances about the main actions and related works (I´ll need your help). WG2 is expected to fill approx. one hour (1 hour and half) with presentations and we invite you to participate actively in the workshop discussions and presentations (will be 10-15 minutes, depending on the number of presentations).

Presentations are open to new partners, working progress reports (for existing collaborations) and related existing Projects.

After these presentations, we plan to have an internal (but open to the rest of participants) brainstorming with open discussion about the possibilities to establish new collaborations, reinforce existing works, and specially, to prepare concrete actuations for the next ICT calls.

We want to invite you to indicate your intention to assist to Coimbra. If you’d decide to assist, please provide us, ASAP, the following information (we highly encourage you to present your work to foster collaborations after the colleagues know what the others are working on):

- Name of participant

- Does he/she need an invitation?
- Name of your institution/group

- Do you plan to give a presentation at the WG2 session?

- Please send us the title of your presentation and a short abstract about the content. (Tentative information is also welcome)

- Do you plan to attend the “bainstorming” session?

- In any case, would you be so kind to send me a short summary of the actions that you have done during the last 6 months? A brief description would be enough (5-10 lines)

Please take into account the existing restrictions about invitations for reimbursement by COST/WG2. Send us an e-mail ASAP, before September 28^th, specially if you need travel budget from WG2.

REMEMBER that the invitation covers only the travel, the hotels and the meals for the meeting, plus one night.

Alberto E. Garcia

Karin A. Hummel

Chairs of WG2

(Passau, 2010 April 15th)

Alberto E. Garcia and Karin A. Hummel, as Chairs of the WG2, made the presentation of the WG2 related results during this first period of the Action. Presentation was divided into 2 parts, the first one, during the common meeting, showed some results obtained from the first WG2 deliverable. The second presentation was delayed until the last part of the meeting and it showed some conclusions about the WG2 community and the possible identification of synergies between groups. Meanwhile 9 WG2 groups’ presentations were divided into two batches that were distributed throughout the day.

Actual state of the WG2 related actions

After the contacts between all the interested Laboratories/Groups, each group sent an extended description of itself. All the responses were included into the WG2 1^st deliverable. Actually 19 groups have showed their interest to belong to the WG2 community. 14 of them have included their corresponding profiles into the deliverable.

A first draft was delivered on February (2010) and the first version was published via the Web on April (2010).

First deliverable presentation

It’s titled as “Calibration tools and methodologies to overpass the heterogeneous and dispersed nature of the equipments”.

For each Lab this deliverable includes a general description about the group, its relationship with WG2/IC0804 issues and a concrete specification about the related efforts/researches. For each work, it includes:

• A description of the work
• The related list of references to works and papers (first the own citations and then the related and recommended)
• A list of bibliography with external references
• Each work is auto-classified using a common List of Keywords
• A list of related results, that means, related software, hardware, testbeds, experiments, libraries, etc.

With all, this first deliverable document results as a group’s catalogue, “update-able”.  Each group representative could indicate all the additional information that he could consider as interesting and all the data would be included into the successive versions of the deliverable.

A first evaluation of the 1^st deliverable contents allows identifying the following results obtained inside the groups belonging to the Cost Action community:

• 79 pages with General information about 14 groups (from 19 directly interested on WG2)
• 33 lines of researching (23 directly related with WG2 + 10 related with COST Action)
• 25 interrelated keywords, with 4 main keywords and 4 related general subjects
• 66 related references (publications in journals, conferences, workshops)
• 14 External references
• 19 Identified tools (software, experiments, testbeds, libraries…)

All the results are mainly related with the WG2, but the inclusion of the rest of groups inside the action would be desirable in the future. A general invitation towards all the groups of the COST Action was made during the common meeting. The idea is to include all the groups of the Action and their works and results independently of the WG.

Presentation of the Knowledge Map

To initiate a Web based Knowledge Roadmap was the first objective proposed in Toulouse meeting. It was presented as a common repository where all the information of the groups, projects, publications, etc will be stored and all the partners could access directly. Contents would be based on the 1^st deliverable contents, and it would be updated continuously and directly by each partner.

A first version of this Knowledge Map was presented as an indexed way to observe individually all the numbers resulting from the first deliverable. Proposed solution is developed as a Web + Wiki based information system. All the information of each group is included into the group’s profile and each group representative could edit its own profile.

In this way, updating the group’s related results would be easy and all the rest of the groups could know the evolution of their colleagues. As a desired result, the Knowledge Map will allow to increase the spreading actions and collaborations between groups.

Groups short presentations

During the meeting, 9 presentations were prepared. Lack of time forced to reduce each presentation to five minutes, but the presenters were able to summarize the main features of the studies and results they wanted to share with other attendees.

All the presentations, with the consent of each presenter, will be included as PDF files into the main Website (WG2 File repository + WIKI) and into the WG2 Blog, into the “WG2 Community” part. The list of presentations is included into the Cost 0804 website (see the WG2 WIKI), and all the PDF will be included as soon as the corresponding authors would send their originals (or PDF versions).

All the presenters were invited to send an extended abstract (3-4 pages) about their presentations to be included into a proceedings publication. Perhaps in the future, if this number of contributions follows, the MC might consider the possibility to organize Workshops with a full call of papers (open to external groups).

WG2 organization issues

Karin Anna Hummel directed the second part of WG2 meeting, with the presentation “WG2 – Summary, Synergies, and Next Steps”. The importance of the topics presented invites to the rest of groups to include their own efforts. This is the reason because the list of researching lines and active groups would be always open. As the list is complete, groups can find affinities to establish collaborations with other laboratories. Joint collaborations will be translated to joint works and publications, accomplishing the spreading objectives of the Cost Action.

We further discussed to start a “Survey and Metric Definition” process for energy efficiency in 2010.

From a closer reading of the first deliverable and its comparison with the presentations made during this meeting, we identify four broad areas of study directly related to the research and keywords currently considered:

1. Networking
   1. Mobile networks

1. VANETs/MANETs
2. On-chip for interconnection
3. EE Wireless Mesh Networks
4. Micro producers / prosumers

1. 1. Home networks

1. Architectures
   1. Migration costs of VM/VR
   2. BitTorrent architecture with proxy
   3. FIT4Green
   4. DataCenters
   5. Content Distribution Networks

1. Middleware/data management
   1. Quorum based replication
   2. Mobile data mining

1. Applications
   1. Energy awareness for mobile applications
   2. Applications (indirect measures)
   3. Energy consumption of mobility prediction in hybrid mobile/infrastructure architectures
   4. Home network: applications

Objectives toward the second year

Three main objectives follow the works during the 2nd year of the action:

1. Knowledge roadmap (Web): Four basic continuous actions:
   1. Completing information about existing groups, references (+ programs, testbeds, libraries…)
   2. Invitation to the rest of groups not related with WG2 to include their own profiles.
   3. Updating from ALL the groups
   4. Including information about the researches into the rest of WG’s.
2. Spreading actions and updates:
   1. Initially, this action would be directly completed with the COST results and described by 2nd deliverable: “Cost models for characterizing energy consumption of different components and of a whole system”.
   2. Organization of workshops (as a possibility to be consider by MC), justified before.
3. Following with collaborations
   1. Publications: Actually 66 references were collected but we need to identify all the articles related with collaborations inside the COST environment.
   2. As a good possibility could be proposed a COST position paper titled “Common definitions of energy (efficiency) metrics: Survey and definitions”, using Knowledge Map and Wiki data.
   3. STSM: New proposals during this period
   4. Towards next 7th Frame calls:

1. Some initiatives started with their proposals. Publicity about these actions could be very interesting for the rest of groups, if the implied groups considering it useful for them (for the MC is very interesting, especially to write the annual report!).
2. Building discussion groups (SIG) on specific topics.
3. To identify research questions, potentials for joint project proposals.

During the meeting a request was made: “Please, inform about your advances… updating your group profile could be sufficient!”

Work plan (until November 2010)

The following scheme was proposed:

1. April 2010: Activation of Knowledge Map environment and emailing of user’s accounts to edit each group profile
2. May/June 2010: Identification of researches related with the 2nd deliverable Objectives.
3. 2nd mid of September 2010: Draft version of the 2nd deliverable.
4. October 2010: Updating of 1st deliverable and Knowledge Map contents.
5. November 2010: common WG2 meeting to establish the end of the 2nd year work plan and collaborations.

Proposed presentations

14:15 Working Group 2: Characterization of energy consumption and energy efficiency

Mohamed Bakhouya (GRTC)

Title : Evaluating the Energy Consumption and the Silicon Area of On-Chip Interconnect Architectures

Abstract: Emerging SoCs, such as those for mobile systems, are typically battery-powered systems and have to support a wide range of streaming applications such as video and audio. On-chip interconnect architectures, used in current SoCs to integrate hardware resources such as CPUs, DSPs, Memory and I/O peripherals, are typically based on shared medium or bus systems. For large SoCs, the bus-based schemes become restrictive because they are non-scalable and have higher overheads that adversely impact performance and energy consumption. In this talk, an approach will be presented to evaluate and compare the energy consumption and the area requirements of six common on-chip interconnects using a variety of load and traffic models.

Keywords: System-on-Chip, On-chip interconnects, Energy consumption and area requirements, Load and traffic models, Simulation and evaluation

Georges DaCosta

Title: Energy consumption of applications

Abstract: One way to raise energy awareness for IT systems is to react to instantaneous energy consumption, and then by taking decision that will reduce this consumption. IRIT build autonomous systems able to reduce energy, and University of Vienna focuses on models. Some feedback on energy consumption is necessary to take decisions and we had the hardare limit to only be able to measure it at the computer level.  During this STSM, we worked on how to evaluate the energy consumption of an application based on indirect measurement such as perfcounters or OS process statistics. We put in place a measurement infrastructure that logs around 200 different values each seconds and made some early measurement campaign and a first model.  By instance, using only the number of instructions per second (using performance counters) and read and write bandwidth on the disk we already obtain a model where the total energy consumed is precise at around 1J/s on a computer that consumes between 70 to 130W depending on the load. More precise models are currently worked on, taking into account more values, such as network or memory usage.

Carmela Comito (UNICAL)

Title: Energy-aware mobile data mining

Abstract: The goal of our research is exploiting Mobile-to-Mobile (M2M) technologies to support pervasive and ubiquitous data mining through mobile devices. The project builds around a M2M architecture providing services for distributed data mining algorithms and applications that take into account the energy constraints of mobile devices. This presentation focuses on WG2-related issues, in particular on defining the parameters that will be used to achieve energy efficiency in the proposed system.

Andrea Passarella (PerLab)

Title: Energy-efficient BitTorrent Architecture for Green Internet File Sharing

Abstract: Recent studies have shown that the Internet-related energy consumption represents a significant, and increasing, part of the overall energy consumption of our society. The largest contribution to this energy consumption is due to Internet edge devices (PCs and data centers). As a particularly significant example, users leave their PCs continuously powered on for satisfying connectivity requirements of Peer-to-Peer (P2P) file sharing applications, like BitTorrent (currently the most popular P2P Internet platform). To reduce these energy consumptions, without penalizing the Quality of Service of BitTorrent users, we propose a novel architecture based on the introduction of a BitTorrent proxy. BitTorrent users delegate the download operations to the proxy and, then, power off their PC, while the proxy downloads the requested files. We implemented our solution and validated it in a realistic testbed. Experimental results show that, with respect to the legacy BitTorrent approach, our solution is very effective in reducing the energy consumption without introducing any QoS degradation. Specifically, our results show that the proxy-based solution can provide up to 95% reduction in the energy consumption and, at the same time, a significant reduction in the average file download time.

Robert Basmadjian (UNI-Passau)

Title: Energy Cost Model for Quorum-based replication protocols

Abstract: An energy cost model that computes the energy cost of read and write operations of quorum-based replica control protocols. Also, the presentation will include my current activities with respect to energy efficiency. For this purpose, a small introduction will be done concerning the FIT4Green project. Afterward, one of my colleagues (Florian Niedermeier) will delve into details.

18:00 Working Group 2 specific meeting

Florian Niedermeier (UNI-Passau)

Title: Energy consumption models in the FIT4Green project

Abstract: The FIT4Green project is aiming at reducing the energy consumption in  federated environments. The modeling of data center components is the basis to understand the relevant parameters for an optimization regarding energy efficiency. A critical point is to find the suitable level of abstraction for these models, which on the one hand captures all relevant details while on the other hand, remains easily computable without introducing too much overhead.

Mohamed Bakhouya (GRTC)

Title: Power-aware broadcasting protocols in MANETs

Abstract: Vehicular Ad-hoc Networks (VANETs) appeared as a subclass of MANETs for inter vehicle communication. VANETs have relatively more dynamic nature as compared to MANETs with respect to network topology. However, among numerous solvable issues in VANETs, the design and implementation of an efficient and scalable architecture for information dissemination constitutes one major issue. In the past few years, several broadcasting protocols for information dissemination are proposed specifically in the context of MANETs. In this talk, we will first elaborate some differentiable characteristics of VANETs and then will present a performance analysis of these broadcasting protocols in VANETs environment using NS2.

Athena Vakali (AUTH)

Title: Energy aware p2p CDNs

Abstract: Survey of potential network, traffic and content parameters that affect the overall energy consumption in a Content Distribution Network.

Outline possible forms of heuristic algorithms that utilize these parameters. Suggest software-side only solutions utilizing existing hardware infrastructures.

Karin A. Hummel  (UoV)

Title: Energy efficiency in University of Vienna

Abstract:. The goal of this research is to set up a possible scenario for attacking in a principled way the problem of “power-saving computing”.  We will mainly address three issues: (1) the need for further scientific observations about the relation between power-consumption and computing, (2) the “design of good models” that explain this experimental observations and allow to understand the fundamental limitation and properties of energy vs performance issues; and (3) the development of a “set of canonical algorithmic techniques to make it easier for the new methods to filter into active use”.

WG2 meeting (Toulouse, 9^th November 2009, 9:00-15:00)

Scheduling

List of representatives (19 participants)