Participación en las actividades de Convivencia escolar

We examine the client loads, which include the request number, storage size, stored objects, hit rate, etc, to verify that the client-cluster balances the storage and request queries. Table 1 shows a summary of the re- quest number and the sizes of the requested objects. Each client receives roughly the same load, and when the client number increases the load of each client decreases. The properties of DHT-base peer-to-peer protocols account for these findings. For the byte re- quest, we again see the effect of the different sizes of the objects, which we strongly believe account for the performance degradation.

FuturE trEndS

Some peer-to-peer approaches for Web caching appear including our approach (Iyer, Rowstron, & Druschel, 2002; Kim & Park, 2003; Xiao, Zhang, & Xu, 2002; Xu, Hu, & Bhuyan, 2004). These approaches distribute the requests and the loads to the client by using the p2p manner. The only difference of them is the storing method of the Web object. Some approaches store the index of the object on the proxy or the clients to share

the object in the storage of client. The others store the object itself to the storage and find the location with the p2p protocol.

In these days, the Web becomes more complicate because of the dynamic pages, the large objects( au- dio/video ) and so on. Moreover, the p2p networks are very heterogeneous. The computing power, the bandwidth, the storage, etc., of the participant nodes are very various. These facts need the well-adapted p2p support for Web caching.

concLuSIon

In this article, we propose and evaluate peer-to-peer client-cluster, which is used as a backup storage for the proxy cache. The proxy cache with this client-cluster is highly scalable, more efficient, and has low admin- istrative cost. Even if the clients take the load, this load has been verified on a range of real workloads to be low. Moreover, the utility of the client-cluster can be improved by managing objects according to their properties such as size, popularity, and update frequency. We can extend the usage of the client-cluster to other proxy systems. If a proxy performs demanding jobs such as encoding/decoding and complex calculation for many clients, it can use the residual resources of the clients to accomplish these tasks.

rEFErEncES

Chankhunthod, A., Danzig, P. B., Neerdaels, C., Schwartz, M. F., & Worrell, K. J. (1996). A hierarchical Internet object cache. Proceedings of the 1996 Usenix

Technical Conference, San Diego.

Cohen, J., Phadnis, N., Valloppillil, V., & Ross, K. W. (1997). Cache array routing protocol v1.0. Retrieved from http://www.ietf.org/internet-drafts/draft-vinod- carp-v1-03.txt

Client No. Mean Req. Max Req. Dev. Mean Byte Req. (KB) Max Byte Req. (KB) Dev.

100 1024 1369 2.2 13422 316805 11.1

200 602 733 2.4 6711 315158 12.1

300 401 510 2.5 4474 314197 12.9



Efficient and Scalable Client-Clustering for Proxy Cache

Iyer, S., Rowstron, A., & Druschel, P. (2002). Squirrel: A decentralized peer-to-peer Web cache. Proceedings

of the 21st _{ACM SIGACT-SIGOPS Symposium on Prin-}

ciples of Distributed Computing, Monterey.

Kim, K., & Park, D. (2003). Efficient and scalable client clustering for Web proxy cache. IEICE Transaction on

Information and Systems, E86-D(9), 1577-1585.

Ratnasamy, S., Francis, P., Handley, M., Karp, R., & Shenker, S. (2001). A scalable content-addressable network. Proceedings of the 2001 ACM SIGCOMM

Conference, San Diego.

Rodriguez, P., Spanner, C., & Biersack, E. W. (1999). Web caching architectures: Hierarchical and distrib- uted caching. Proceedings of the 4th _{International Web}

Caching Workshop, San Diego.

Rowstron, A., & Druschel, P. (2001). Pastry: Scal- able, decentralized object location, and routing for large-scale peer to-peer systems. Proceedings of the

18th _{International Conference on Distributed Systems}

Platforms, Heidelberg.

Stoica, I., Morris, R., Karger, D., Kaashoek, M. F., & Balakrishnan, H. (2001). Chord: A scalable peer-to-peer lookup service for Internet applications. Proceedings of

the 2001 ACM SIGCOMM Conference, San Diego.

Wolman, A., Voelker, G. M., Sharma, N., Cardwell, N., Karlin, A., & Levy, H. M. (1999). On the scale and per- formance of cooperative Web proxy caching. Proceed-

ings of the 17th _{ACM symposium on Operating Systems}

Principles (pp. 16-31), Kiawah Island Resort.

Xiao, L., Zhang, X., & Xu, Z. (2002). On reliable and scalable peer-to-peer Web document sharing. Proceed-

ings of the 16th _{International Parallel and Distributed}

Processing Symposium (pp. 228), Fort Lauderdale,

IEEE CS.

Xu, Z., Hu, Y., & Bhuyan, L. (2004). Exploiting client cache: A scalable and efficient approach to build large Web. Proceedings of the 18th _{International Parallel and}

Distributed Processing Symposium (pp. 55a), Santa

Fe, IEEE CS.

Zhao, B. Y., Kubiatowicz, J., & Joseph, A. (2001). Tapestry: An infrastructure for fault-tolerant wide- area location and routing. UCB Technical Report

UCB/CSD-01-114.

KEY tErMS

Backward ICP: A communication protocol be-

tween a proxy cache and a client-cluster.

Client-Cluster: A well organized group of clients,

which communicate with the p2p protocol.

Cooperative Proxy Cache: A proxy cache, which

trusts and cooperates with each other proxy cache.

Distributed Hash Table: A hash table, which is

distributed to many other components with the specific mechanism.

Home Node: A node, which is responsible for a

target object key.

Node Key: A unique and hashed value to identify

a node.

Object Key: A unique and hashed value to identify

an object.

Peer-to-Peer: A communication method between

clients. cf. client-to-server

Proxy Cache: A server, which locates on the border

of ISP or institution and helps the original server by intercepting the user request.

Station: An application or daemon to help the com-

E

E-Learning

Gregory R. Gay

University of Toronto, Canada Paola Salomoni

University of Bologna, Italy Silvia Mirri

University of Bologna, Italy

IntroductIon

The evolution of an information society has transformed many activities in our everyday lives, including how we work, communicate, entertain, teach, and learn. More recently widespread Internet connectivity together with the development of new Web-based multimedia tech- nologies, has strongly encouraged educational uses for information and communication technology (ICT).

Activities that require network technologies to de- liver learning and training programs can be considered forms of “e-learning.” In the not too distant past, before the broad use of Internet-based learning technologies be- came commonplace, a variety of technological solutions had been used to support teaching and learning activi- ties, generally referred as computer assisted instruction (CAI). With CAI, people started using a computer as a means of teaching, providing tutorials, simulations, or games. CAI was generally learning activities that did not require network connectivity. E-learning today more often refers to systems and activities that go well beyond interactions with a single computer.

ICT naturally fueled the spread of e-learning, forc- ing the emergence of a society in which economic and social performance are largely judged by a continuous distribution of information and knowledge (Sloman, 2002). After a complex and long experimental phase in the latter decades of the 20th _{century, current e-lear-}

ning technologies support cost effective, just-in-time, customized education that is delivered effectively in a variety of educational scenarios including schools, post secondary education, and professional skills training. Easy access to learning content and learning tech- nologies provides greater opportunity for people and communities to develop new skills and improve their

knowledge, minimizing the effects of socio-economic status, geographic location, and other ethnographic characteristics.

BAcKGround

E-learning currently represents the most widespread form of “distance education,” which generally refers to educational activities that involve teachers and students remotely located both in time and space. Current distance education is based on a wide range of delivery methods, including traditional correspon- dence, as well as books, audio/video tapes, interactive TV, CD-ROM, and DVD, as well as services that can be offered through the Internet (Rosenberg, 2000). More generally “e-learning” can be defined as the delivery of education or training programs through electronic means. Beyond this general definition, are terms or phrases used synonymously with e-learning or its specific forms. One such frequently used term is “computer-based training” (CBT), its main unique distinction being its focus on the computer instead of a more general association with ICT. Another broadly used term is “online learning,” which emphasizes the presence of the Internet as a means for conducting learning activities. An additional specialization is ter- med “Web-based education,” which more specifically refers to the use of Web-based applications that support educational activities.

E-learning activities can be classified according to their time dependence as (Clark & Mayer, 2002): • “Synchronous” e-learning: Where teachers and

0

E-Learning

same time; they have to synchronize their activi- ties; examples include video conferencing, chats, and real-time video lectures

• “Asynchronous” e-learning: In which teachers

and students are involved in learning activities at different times; they “do not” have to synchronize their activities; examples include content delivery, cooperation through a forum, a blog, or a wiki, as well as e-mail communication and file shar- ing; new asynchronous e-learning applications are emerging based on archived podcasting and Webcasting content.

The more common approach is asynchronous e- learning. Its main benefit is its freedom from time and space requirements, thus supporting an “anytime” dimension of e-learning. On the other hand, synchro- nous activities give e-learning more appeal, involving people interacting directly with each other in real-time activities. A typical e-learning scenario might involve components of both.

E-LEArnInG tEcHnoLoGIES

From a technological point of view, today’s e-learning is rooted primarily in Web-based delivery of educa- tional multimedia content, coupled with synchronous and asynchronous communication features that allow students and teachers to interact (Sloman, 2002). There are several types of systems that assist with e-learning activities. Perhaps, the most common type of e-learning application, the so called learning management system (LMS), is devoted to managing learning activities and, more specifically, to keeping track of what learners do and learn, following both their activities within the system, and their progress mastering learning materials. A second category of applications, called a learning

content management systems (LCMS), aim primarily

at managing the delivery of course content, but such systems are frequently extended to include communica- tion tools and user management features often found in an LMS (Jacobsen, 2002).

Often used interchangeably with LMS (and often with LCMS) is the virtual learning environment (VLE), with its focus on the “virtuality” of the learning space and on the idea of a platform that supports the whole range of learning activities as a stand alone integrated “virtual environment.”

A very different role is played by “e-learning con-

tent repositories,” applications that are used to store,

distribute, and share learning content. Often content repositories are linked into an LMS so instructors, and often students, can search and retrieve learning mate- rials, export content from the repository in a standard format, import or link that content into an LMS as part of a course lesson, and often transfer content from an LMS into the repository to be shared with others.

The nature of e-learning content allows it to be reused in many contexts, and to be redistributed as standardized packaged educational materials (Horton & Horton, 2003). Pieces of learning material in a “content package” are often referred to as “learning

objects” (LO). A content package is a collection of

learning objects assembled together with a document, generally XML based, that defines association and sequencing rules used to organize the content within an e-learning system.

E-LEArnInG StAndArdS

A standard description of content structure is needed to make content interoperable across different e-learn- ing platforms. Several interoperability specifications have been developed by international organizations such as:

• The Institute of Electrical and Electronics En- gineers (IEEE): With a specific working group,

the Learning Technology Standards Committee, that is working on e-learning standardization (IEEE, 2006)

• The Instructional Management System Global

Learning Consortium (IMS Global Learning Consortium): A collaboration of government

organizations that are defining specifications to ensure interoperability across e-learning systems (IMS Global Learning Consortium, 2006b) The goal of such standards is to define metadata, data structures, and communication protocols that will make learning content work across platforms, providing guidelines for designing, developing, and delivering electronic learning content.

Another similar project is the advanced distributed learning (ADL) initiative, lead by the U.S. Department of Defence. It has developed the shareable content object

E

reference model (SCORM) standard, one of the more

widely used e-learning specifications. ADL has based its work on that of IEEE and IMS, and has created a more encompassing interoperability standard that takes into consideration recommendations from those and other standards (ADL, 2004). The collected standards can be applied to learning content and to learning platforms (e.g., LMS, LCMS, VLE) with the aim to fully support the reuse of content across systems and standardize the delivery e-learning content. SCORM includes a de-facto standard for defining a sharable content object (SCO). An SCO is a learning resource that can be presented in any SCORM compliant system, displaying and sequenc- ing content, and tracking student progress. Each SCO is made up of one or more assets or resources, which are generally electronic media (e.g., text, images, sound, video), Web pages, or other types of data. SCOs can be described with metadata, and often retrieved from online content repositories by searching for terms in the metadata, thereby encouraging their re-use.

Metadata and structural information about a unit of learning content is usually contained within a “manifest,” an XML file that describes the learning content in a standard manner. A SCORM manifest generally contains the content semantic description (metadata), together with its navigation or structural description (organizations), and the locations of each of the contained assets (resources). The SCORM main specifications are (ADL, 2004):

• The content aggregation model (CAM) that defines the structure or arrangement of learning materials, and describes the content or topics it contains with metadata (based on the IMS content packaging specification)

• The run-time environment (RTE), a JavaScript Application Programming Interface (API) that delivers to a LMS or LCMS real time informa- tion about user actions within a SCO, including exercise solving and tracking through resources • The sequencing and navigation (SN) specification

describes rule-based definitions of possible paths through learning content

In document ÍNDICE CONTEXTO LÍNEA ESTRATÉGICA: CONSOLIDACIÓN DE LA INFRAESTRUCTURA FÍSICA EDUCATIVA... 30 (página 65-0)