[ieee 16th international conference on artificial reality and telexistence--workshops (icat'06)...

An Application Program and Error Sharing Agent running on Ubiquitous Networks

Eung Nam Ko Kee Chun Bang Dept. of Multimedia, Dept. of Multimedia, Baekseok University Namseoul University

[email protected] [email protected]

Abstract

This paper presents the design and implementation

of an application program an error sharing agent for collaborative multimedia distance education system which is running on RCSM (Reconfigurable Context Sensitive Middleware) for ubiquitous networks. RCSM provides standardized communication protocols to interoperate an application with others under dynamically changing situations. It describes a hybrid software architecture that is running on situation-aware middleware for a web based distance education system which has an object with a various information for each session and it also supports multicasting with this information. 1. Introduction

Advanced information network and multimedia technology are accomplished by combination of educational media through computer, video conference system CSCW(Computer Supported Cooperated Works), environment and interaction between participants of different location with on-screen pictures of each other which are possible to use voice, text and graphic. Distance Education system must be able to support real-time interaction and also support user synchronization including not only temporal synchronization and spatial synchronization but floor control for smooth interaction [1]. A general web-based distance system uses video data and audio data to provide synchronize between teacher and student. In a ubiquitous computing environment, computing anytime, anywhere, any devices, the concept of situation-aware middleware has played very important roles in matching user needs with available computing resources in transparent manner in dynamic environments [2]. There are two approaches to software architecture on which distributed,

collaborative applications are based. Those include CACV (Centralized-Abstraction and Centralized-View) and RARV (Replicated-Abstraction and Replicated-View). We propose an adaptive agent of application program sharing based on a hybrid software architecture which is adopting the advantage of CACV and RARV for situation-aware. 2. Related Work

As shown in Table 1, conventional multimedia distance education systems are Shastra, MERMAID, MMconf, and CECED. You can see the characteristic function of each system function for multimedia distance education.

Table 1. Analysis of Conventional Multimedia

Distance Education System Function Sha-

Stra MER- MAID

MM- conf

CEC-ED

OS UNIX UNIX UNIX UNIX Development Location

PurdueUniv.

USA

NEC, JAPAN

CamBridge USA

SRI, International

Development

Year

1994 1990 1990 1993

Structure Server /client

Server /client

Centralized or Replicated

Repli- cated

protocol TCP/IP TCP/IP TCP/IP TCP/IPmulticast

A proposed main structure is distributed architecture

but for application program sharing, centralized

Proceedings of the 16th International Conference onArtificial Reality and Telexistence--Workshops (ICAT'06)0-7695-2754-X/06 $20.00 © 2006

architecture is used. The problem of rapid increase in communication load due to growth in number of participants was solved by letting only one transmission even with presence of many users, using simultaneous broadcasting. Basically, there are two architectures to implement such collaborative applications; the centralized architecture and replicated architecture, which are in the opposite side of performance spectrum. Because the centralized architecture has to transmit huge amount of view traffic over network medium, its performance is reduced to contaminate the benefits of its simple architecture to share a copy of conventional application program. On the other hand, the replicated architecture guarantees better performance in virtue of its reduced communication costs. However, because the replicated architecture is based on the replication of a copy of application program, it is not suit to use for application sharing realization[3-8]. However, it did not include application sharing agent in the conventional architecture.

3. Our Approach 3.1. Other Services based on RCSM

The Context Toolkit was built based on this conceptual framework. There were five applications that were built to assess the actual benefits of the Context Toolkit. Seminal work has been done by Anind Dey, et al. [18] in defining context-aware computing, identifying what kind of support was required for building context-aware applications and developing a toolkit that enabled rapid prototyping of context-aware applications. They have laid out foundations for the design and development of context-aware applications by proposing a conceptual framework. In the Context Toolkit, a predefined context is acquired and processed in context widgets and then reported to the application through application-initiated queries and callback functions. In this Reconfigurable Context-Sensitive Middleware(RCSM), Stephen S. Yau et al.[9] proposed a new approach in designing their middleware to directly trigger the appropriate actions in an application rather than have the application itself decide which method(or action) to activate based on context. RCSM provides an Object-based framework for supporting context-sensitive applications. Figure 1 shows how all of RCSM’s components are layered inside a device. All of RCSM’s components are layered inside a device. The Object Request Broker of RCSM (R-ORB) assumes the availability of reliable

transport protocols; one R-ORB per device is sufficient.

Fig. 1. RCSM’s integrated components

The number of ADaptive object Containers (ADC)s

depends on the number of context-sensitive objects in the device. ADCs periodically collect the necessary

Situation-Aware Application Objects

O S

<<RCSM- Optional Components>> <Other services>

Sensors Transport Layer Protocols for Ad Hoc Networks

RCSM Ephemeral Group Communication Services

Adaptive Object Containers(ADCs) [Providing awareness of situation]

RCSM Object Request Broker(R-ORB) [Providing transparency over ad hoc

SA-SMA

SA- UIA

FTA AMA

SA- ACCA

MCA


“raw context data” through the R-ORB, which in turn collects the data from sensors and the operating system. Initially, each ADC registers with the R-ORB to express its needs for contexts and to publish the corresponding context-sensitive interface. RCSM is called reconfigurable because it allows addition or deletion of individual ADCs during runtime (to manage new or existing context-sensitive application objects) without affecting other runtime operations inside RCSM. Ubiquitous applications require use of various contexts to adaptively communicate with each other across multiple network environments, such as mobile ad hoc networks, Internet, and mobile phone networks. An example of SmartClassroom is illustrated in [10]. However, it did not include other services support in the architecture. A good example of other services in RCSM is multimedia distance education system. 3.2. Multimedia Distance Education

System

Our proposed model aims at supporting application program sharing agent running RCSM in order to provide ubiquitous, seamless services. An example of situation-aware applications is a multimedia distance education system.

DOORAE(Distance Object Oriented collaboRAtion Environment) is a good example of the framework technology for multimedia distance education system running RCSM. It has primitive service functions. Service functions in DOORAE are implemented with object oriented concept. We call agent layer. As shown in Figure 2, the organization of DOORAE includes 4 layers. The four layers consist of a communication layer, a system layer, a DOORAE agent layer and a multimedia application layer.

The communication network is being presently developed with UDP broadcasting in order to decrease communication rate and TCP/IP on the Ethernet and ATM. Additional packet form has been defined and expanded for realization of DOORAE’s functions. The hardware environment of DOORAE consists of multimedia PCs, a network adapter, keyboard/mouse, image scanner, microphone, video camera, monitor, speaker, printer, video processor and accelerators. The operating system was first developed on windows 98 but presently windows XP and windows 2003 are supporting the development as well. The multimedia application layer includes general application software such as word processors, spreadsheets, presentation tools and so on. DOORAE agent layer includes many agents. They are AMA(Application Management

Agent), IA(Intelligent Agent), SMA(Session Management Agent), ACA(Access Control Agent) and MCA(Media Control Agent). User Application User Application Session Interface Monitor Session Session Data Saver Controller Traffic Video/Audio Daemon Monitor Provider Network Resource Reliable Multipoint Manager Transport Layer Reliable Multipoint Transport Layer Microsoft Windows NT / 2003 / 98 / XP

Physical Network Fig. 2. The organization of DOORAE AMA consists of various subclass modules. These subclass modules provide the basic agent, while AMA supports a mixture of various basic services. AMA includes creation/deletion of shared video window and creation/deletion of shared window. IA which is designed to support interaction between heterogeneous computer is necessary to transfer media data and to notify collaboration status to distant participants. SMA controls the access to the session. This facilities video conference, distance learning, game and development of any software. Session control also enables the access and limits the access to the whole session. It monitors the session start/termination, join/invitation/leave session, and it also permit to join another sessions. ACA decides who has right to speak


and manages distribution of resources along with user status when there is a request for resources. It controls the person who can talk, and the one who can change the information. The mechanism of floor control consists of simultaneous, priority, mediation, token-passing and time-out. MCA manages hardware resources. It controls effective assignment of limited media resources and conclusion of use, and furthermore maintains efficiency in use of resources as well as making sure that media resources are shared between distance users.

SMA consists of GSM(Global Session Manager), Daemon, LSM(Local Session Manager) and PSM(Participant Session Manager). GSM has the function of controlling whole session when a number of sessions are open simultaneously. LSM manages only own session. For example, LSM is a lecture class in distributed multimedia environment. GSM can manage multiple LSM. Daemon is an object with services to create session. Session management can create the sequence below and you can see the message flow in Figure 3. (1) invite participants (2) join message (3) start session A (4) create video/ audio B (5) send lecture note (6) share event and commands

(7) keep interaction history (8) quit session

A: Initial a session,B: Participants

Fig. 3. Session Management

This approach is based on the idea of comparing the expected error type which is generated with the actual error occurred from sites. The sequence of Recovery’s message flow can be shown in Figure 4. If an error is to be recovered, you can create sequences below. It creates a session with initial configuration information. It requests port ids for audio/video servers to build-up a Local Session Manager. It assigns port ids for audio/video servers of an application. It invites to the session and build-up a session instance monitor. It sends invited messages to start build-up of session instance monitor. It builds up Session Instance Monitor using the configuration information from LSM. It sends joint message to the Local Session Manager. It sends session information to Global Session Manager for set-up of GSM table. It begins a session. It exchanges message or command between LSM and

PSM and media data between media server based on interpretation of message handler. <GSM> (8) (2) (3) <Local Daemon> <Remote Daemon> (4) (5) <LSM> (1) (6) <An Application’ <Participant recovery request> (10) Session Manager> (9) (7)

Fig. 4. The relationship between GSM, Daemon and LSM

3.3. Application Program Sharing & Error

Agent

As shown in Fig.5, error and application program sharing windows perform process communication of message form.

Fig. 5. Application Sharing and Error Process

<<E Network view /event view/event

<ES> hook table

<ES>

application

<ES>

Virtual app.

Filter

<ES>

Virtual app.

Filter

Event Distri- buter

Filter function


The roles of application program and error sharing are divided into two main parts; Abstraction and sharing of view generation. Error and application program sharing must take different from each other according to number of replicated application program and an event command. This proposed structure is distributed architecture but for error and application program sharing, centralization architecture is used. 4. Simulation Result

To evaluate the performance of the proposed system,

an error detection method was used to compare the performance of the proposed model against the conventional model by using DEVS formalism. In DEVS, a system has a time base, inputs, states, outputs based on the current states and inputs. DEVS(Discrete Event System Specification) is a formalism of being developed by Bernard P. Zeigler. The structure of atomic model is as follows [11 - 17]. M = < X, S, Y, δint, δext, λ, ta > X: a set of input events. S: a set of sequential states. Y: a set of output events. δint : internal transition function δext : external transition function

λ : output function ta : time advance function

For the structure of software implementation, the CACV(Centralized-Abstraction and Centralized-View) and the RARV(Replicated-Abstraction and Replicated-View) are extreme approaches to design software architecture on which distributed, collaborative applications are based. The CARV(Centralized-Abstraction and Replicated-View) architecture is also comparable with those architecture in terms of performance [17]. Overview of the command serialization overhead is obtained as seen by Figure 7. The figure on the vertical line means relative overheads compared with maximum value on RARV.

<Serialization Overhead of Each Architecture>

0

2

4

6

8

1 3 5 7 9 11 13 15 17

Number of Sites

Late

ncy(

Ratio)

CACV

RARV

CARV

Fig. 6. Serialization Overhead of Each Architecture

Table 2. Comparison for Software Architecture in situation-aware environment

Centrali

-zed Replicat-ed

proposed

Set initial State

Easy hard Medium

Allow Late Comer

Easy hard Easy

Command Serialization

Easy hard Easy

Communication Overhead

High low Low

Probability good bad Good Performance bad good Good Application Copy

one more than

one

More than one

Control Complexity

Low high High

5. Conclusion

The roles of application program sharing are divided into two main parts; abstraction and sharing of view generation. Application program sharing must take different from each other according to number of replicated application program and an event command. There are two different structures. Those are CACV and RARV. In this paper, we discussed a hybrid software architecture which is adopting the advantage of CACV and RARV. The roles of error and application program sharing are divided into two main parts; Abstraction and sharing of view generation. Error and application program sharing must take different from each other according to number of replicated application program and an event command. This proposed structure is distributed architecture but for error and application program sharing, centralization architecture is used. We proposed an adaptive synchronization control agent based on a hybrid software architecture which is adopting the advantage of CACV and RARV for situation-aware middleware. It described a hybrid software architecture that is running on situation-aware ubiquitous computing for a web based distance education system which has an object with an various information for each session and it also supports multicasting with this information. This paper proposed a new model of synchronization control by analyzing the window and attributes of the attributes of the object, and based on this, a mechanism that offers a seamless view without


interfering with error and application program sharing is also suggested. We remain an adaptive agent of error and application program sharing with error elimination function for domino effect based on a hybrid software architecture which is adopting the advantage of CACV and RARV for situation-aware.

References [1] Gil C, Park and Dae J, Hwang, “Design of a multimedia

distance learning system: MIDAS”, Proceedings of the IASTED international conference, Ptiisburgh, USA, Aprl., 1995.

[2]S. S. Yau and F. Karim, "Contention-Sensitive Middleware for Real-time Software in Ubiquitous Computing Environments", Proc. 4th IEEE Int’l Symp. on Object-Oriented Real-time Distributed Computing (ISORC 2001), May 2001, pp. 163-170.

[3] Eung-Nam Ko, A Web Based Multimedia Distance Education System With An URL and Error Synchronization Function, WSEAS Transactions on Computers, Issue 4, Volume 3, October 2004, pp.1142-1146.

[4] ARMANDO B. BARRETO and JING ZHAI, Physiologic Instrumentation for Rea-time Monitoring of Affective State of Computer Users, Proceedings of WSEAS USA Miami Conference, April 2004, #484-220.

[5]A. Anupam and C.L.Bajai, Collaborative Multimedia Scientific Design in Shastra, Proceeding of the ACM Multimedia’93, Aug.1993, pp.447-456.

[6]T. Ohmori and K. Watabe, Distributed Cooperative Control for Application Sharing Based on Multiparty and Multimedia Desktop Conferencing Systems:MERMAID, 4th IEEE ComSoc International Workshop on Multimedia Communications, April 1-4, 1992.

[7]Torrence Crowley and Raymond Tomlinson, MMConf: An Infrastructure for Building Shared Multimedia Applications, CSCW ’90 Proceedings, October 1990.

[8]Earl Craighill and Keith Skinner, CECED: A System For Informal Multimedia Collaboration, Proceedings ACM Multimedia ’93, August 1-6 1993.

[9]S. S. Yau, Y. Wang, D. Huang, and H. In “A Middleware Situation-Aware Contract Specification Language for Ubiquitous Computing”, FTDCS 2003.

[10]Saha, D.; Mukherjee, A.; “Pervasive computing: a paradigm for the 21st century”, IEEE Computer, Volume: 36, Issue:3, March 2003, Page(s): 25-31.

[11]Bernard P.Zeigler, Tae H. Cho, and Jerzy W. Rozenblit, A Knowledge-Based Simulation Environment for Hierarchical Flexible Manufacturing, IEEE Transaction on Systems, Man, and Cybernetics-Part A: System and Humans, Vol. 26, No. 1, January (1996), 81-90.

[12]Tae H. Cho, Bernard P.Zeigler, “Simulation of Intelligent Hierarchical Flexible Manufacturing: Batch Job Routing in Operation Overlapping” , IEEE Transaction on Systems, Man, and Cybernetics-Part A: System and Humans, Vol. 27, No. 1, January (1997), 116-126.

[13]Bernard P.Zeigler, “Object-Oriented Simulation with hierarchical, Modular Models”, Academic Press, (1990).

[14]Bernard P.Zeigler, “Multifacetted Modeling and Discrete Event Simulation”, Orlando, FL: Academic, (1984).

[15]Bernard P.Zeigler, “Theory of Modeling and Simulation”, John Wiley, NY, USA, (1976), reissued by Krieger, Malabar, FL, USA, 1985.

[16]A.I. Conception and B.P. Zeigler, “The DEVS formalism: Hierarchical model development”, IEEE Trans. Software Eng., vol. 14, no.2, (1988), 228-241.

[17]J. C. Lauwers and K. A. Lantz, “Collaboration Awareness in Support of Collaboration Transparency: Requirements for the Next generation of Shared Window Systems”, proc. of ACM CHI’90, April (1990), pp.302-312.

[18]Dey, A.K., et al. “A Conceptual Framework and a Toolkit for Supporting the Rapid Prototyping of Context-Aware Applications”, anchor article of a special issue on Context-Awareness Computing, Human-Computer Interaction HCI Journal, Vol.16, (2001). http://www.cc.gatch.edu/fce/contexttoolkit.


[ieee 16th international conference on artificial reality and telexistence--workshops (icat'06)...

Documents