David Chess, Colin Harrison, Aaron Kershenbaum "Mobile Agents: Are They a Good Idea?" IBM Research Report RC  Home/Search   Document Details and Download   Summary   Related Articles   Check

....uncertain data. Mobile Agents An agent needs to interact with its host system and other agents in order to be useful. Mobile agents are defined as programs which may be dispatched from a client computer and transported to a remote server computer for execution and interaction with other agents [11]. For certain applications, it is useful for agents to be able to move within heterogeneous networks of computers. This is possible only if there is common framework for agent operations across the entire network. One approach is to provide a framework which has an agent infrastructure, which is ....

C.G.Harrison, D.M.Chess, A.Kershenbaum, "Mobile Agents: Are they a good idea?", IBM Research Report, T.J.Watson Research Center, NY, 1995.

....Magic introduced Telescript in 1994 [14] agent based computing, and the use of mobile agents in particular, has yet to gain widespread acceptance. This is despite the many potential advantages that are often argued in its favor, including increased performance, fault tolerance, and security [2, 4, 6, 11, 15]. This problem is not due to a lack of agent support systems, as some have now existed for many years and are attaining good levels of maturity, such as Aglets [9] Ara [10] D Agents [5] Mole [13] and TACOMA [8] amongst many others. These middleware software systems support the appropriate ....

C. Harrison, D. Chess, A. Kershenbaum, "Mobile Agents: Are They a Good Idea?" IBM Research Report, Mar. 1995.

....is synchronous. In contrast, mobile agents can support dynamic configuration, automated versioning and remote installation, dynamical addition of new functionality to remote computational facilities, increased asynchrony and robustness, disconnected operations and reduced network use [12]. But if we use mobile agents only to enhance the maintenance and versioning of the ABIS components, the resulting system will in principle, is the same as the one with purely static agents. The single difference is that the packages of the agents now can be remotely installed. 3.2 Arguments for ....

Harrison C.G., Chess D.M., Kershenbaum, A., (1995), "Mobile Agents: Are They a Good Idea?", Research Report, IBM Research.

....captures both the data and moving behavior of mobile transactions and it is defined as a general model where it can provide mobile transaction processing in a heterogeneous, multidatabase environment. The model can deal with both short lived and long lived transactions. The mobile agents concept [3] for multi node processing of a KT can be used when the user requests new subtransactions based on the results of earlier ones. This idea is discussed in [6] as pointing out that there will be no need to keep status table and log files in the base stations DAA. In this case, agent infrastructure ....

D.Chess, C.Harrison, A.Kershenbaum, "Mobile Agents: Are they a good idea?", IBM Research Report, T.J.Watson Research Center, NY, 1995.

....network functionality and could replace current network protocols such as Remote Procedure Calling (RPC) and messaging. In fact, if all the functionality provided by a dynamic agent system is taken together, there is no single alternative that can provide that same level of functionality [2]. Advantages of using dynamic agents include [2] 7] 11] 1) They can reduce communication and bandwidth costs 2) They can be hardware and operating system independent depending on the mobile agent platform they execute on 10 3) They can be fault tolerant 4) They can maintain an optimal ....

....current network protocols such as Remote Procedure Calling (RPC) and messaging. In fact, if all the functionality provided by a dynamic agent system is taken together, there is no single alternative that can provide that same level of functionality [2] Advantages of using dynamic agents include [2][7] 11] 1) They can reduce communication and bandwidth costs 2) They can be hardware and operating system independent depending on the mobile agent platform they execute on 10 3) They can be fault tolerant 4) They can maintain an optimal configuration However, there are three main reasons ....

[Article contains additional citation context not shown here]

Chess, D., C. Harrison, and A. Kershenbaum, "Mobile Agents: Are They a Good Idea?," Mobile Object Systems. Towards the Programmable Internet. Second International Workshop, MOS '9, 1996. Linz, Austria: Springer-Verlag.

....agents offer a new paradigm for large scale distributed heterogeneous applications. The paradigm focuses on the interactions of autonomous, cooperating processes which can adapt to human other agents [12] Mobility is an orthogonal characteristic which many, but not all, consider important [13]. Intelligence is always a desirable characteristic but is not strictly required by the paradigm. The paradigm is still forming. The roots of agent oriented (AO) methodology attain from OOP methodology [14] 15] 16] and AI studies [17] hence, an intelligent agent may be defined as a decision ....

C.G.Harrison, D.M. Chess, A.Kershenbaum "Mobile agents: are they a good idea?", IBM Research Division, T. J. Watson Research Center, Yorktown Heights, NY 10598

....unchanged for decades, while more recent approaches such as COM and CORBA [8] continue to be actively developed. However, even the most sophisticated of these technologies have roots in the fundamental techniques of messaging, simple datagrams, sockets, remote procedure calls, and conversations [18]. In general, these methods can be split into two categories synchronous and asynchronous protocols. Synchronous data communication requires that each end of an exchange of communication respond in turn without initiating a new communication. Asynchronous communication pertains to processes ....

....discovered. For example, the use of several remote procedure calls to perform a client server transaction may use more network bandwidth than sending a more complicated query to a server, performing necessary computation or accessing of databases locally, and returning the results to the client [18]. Initial attempts used the concept of process migration in an attempt to save bandwidth and increase performance. However, movement of an entire address space from one machine to another, as utilized by this technique, makes it difficult to return the results to the client without returning the ....

[Article contains additional citation context not shown here]

C. G. Harrison, D. M. Chess, A. Kershenbaum, "Mobile Agents: Are They a Good Idea?", Technical report, IBM T.J. Watson Research Center, Mar. 1995.

....example CORBA, only the data is transferred as a message. Advantages of Mobile Agents Some of the advantages of Mobile Agents include the simplicity of the installed server base, server flexibility, local and real time interaction. Further analysis of these and other advantages can be found in [5][6] Mobile agent frameworks also tend to support the use of cloning, persistent storage and multicast messaging [8] Most of the program logic usually encapsulated by current distribution paradigms in static clients or servers has been moved into the mobile agent, ensuring the installed server ....

Chess, D., Harrison, C., Kershenbaum, A., "Mobile Agents: Are They a Good Idea?", in [2], also available at http://www.research.ibm.com/massive/ mobag.ps

....be moved (such as a particular device) or 2) a resource whose movement would cause more traffic than migrating the component (such as a large database) When the component no longer requires the resource it can be migrated back to the original host. This is similar to data collection agents[2] in effect, with the exception that control of the component s movement is maintained outside of the mobile entity. Decisions such as when to migrate a particular component and to which host are system policy decisions that are outside the scope of our work. We need to establish an understanding ....

Harrison, C., Chess, D. & Kershenbaum A., "Mobile Agents: Are they a good idea?", Technical Report, IBM T.J. Watson, 1995.

..... Particularly in highly urban spots, where the location areas are relatively small, the statistics have shown that the location updating rate can be 10 times that of the calling rate. In future wireless systems it is even expected to augment dramatically since the LA sizes tend to be decreasing [2]. 2.1 Reducing the number of location updates with mobile agents As we have seen, the part of registering or, in other words, tracking the user while roaming is causing a high load in the network because it requires quite a number of signaling messages to be exchanged for every location update. ....

.... is a program that helps a user to perform some task (or a set of tasks) possibly by maintaining persistent state and communicating with its owner, other agents or its environment in general [14] Developing applications using mobile agents is an effective choice for many reasons as stated in [2,12]. One interesting application area for mobile agents is mobile computing. Mobile devices, such as portable computers, are only intermittently connected to a network and have only intermittent access to a server. The mobile devices are getting smaller, but the wireless access to the services will ....

Colin G. Harrison, David M. Chess, Aaron Kershenbaum, "Mobile Agents: Are they a good idea?", Technical Report, IBM T.J. Watson Research Center, 1995

.... To minimize network administrator intervention in a component s configuration. These objectives can be met with a use of current and proposed features of CORBA [10] 11] and mobile code. Other approaches to PnP we have explored include mobile agents [2] 12] and Jini [14] Software agents [4] and, specifically, mobile agents 2 [3] 5] have also been shown to be widely applicable to network management tasks [1] We propose a Plug and Play (PnP) framework for networks that is composed of smart, intelligent service provider components. After some manual steps that are required to ....

Chess, D., Harrison, C. and Kershenbaum, A.; "Mobile Agents: Are they a good idea?"; IBM Research Report; RC 19887 (88465); Intelligent Agents Project at IBM T.J. Watson Research Center; available at URL: http://www.research.ibm.com/iagents/paps/mobile_idea.ps.

....promising in application domains like information retrieval in widely distributed heterogeneous open environments (e.g. the WWW) network management, electronic commerce, or mobile computing. The question what real advantages mobile agents offer has been subject of various papers (e.g. [HaChKe95], 2 Our Agent Model 2 [BaTsVi96] RoHoRa97] and ongoing discussions in mobile agent mailing lists. The results of these investigations and discussions show that we are far from a common understanding concerning the pros and cons of mobile agent technology. But all agree on the following: To ....

....and Future Work 22 The next problems we will investigate are group models, a performance model and the area of commercial applications. It is often argued that the advantage of agent migration lies in the reduction of (expensive) global communication costs by moving the computation to the data [GenMag96, HaChKe95]. Although this argument is understandable from an intuitive point of view, not much work has yet been done to evaluate the performance of migration on a quantitative basis. A performance model could provide help to identify situations in which agent migration is advantageous compared to remote ....

C. Harrison, D. Chess, A. Kershenbaum. "Mobile Agents: Are they a good idea?", IBM Research Report, IBM T.J. Watson Research Center, 1995.

....PICCO, AND VIGNA: UNDERSTANDING CODE MOBILITY 3 mobile code concepts and technologies. A first problem is the unclear distinction between implementation technologies, specific applications, and paradigms used to design these applications. In an early and yet valuable assessment of code mobility [12], the authors analyze and compare issues and concepts that belong to different abstraction levels. Similarly, in a recentwork about autonomous objects [13] mechanisms like REV [14] and RPC [15] are compared to the Echo distributed algorithms [16] to applications like intelligent e mail and Web ....

....In [23] Knabe lists the essential characteristics of a mobile code language. They include support for manipulating, transmitting, receiving, and executing code containing objects . However, there is no discussion about how to manage the state of mobile components. Other contributions [24] [12] consider only the support for mobility of both code and state, without mentioning weaker forms of code mobility involving code migration alone as we discuss later on in the paper. Certainly, confusion and disagreement are typical of a new and still immature research field. Nevertheless, ....

C.G. Harrison, D.M. Chess, and A. Kershenbaum, "Mobile Agents: Are they a good idea?," In Vitek and Tschudin [73], pp. 25--47, Also available as IBM Technical Report.

.... x Asynchronous task execution allowed by the autonomy of mobile agents x Reduction in network traffic by performing local interaction x Reduction in dependence regarding network availability by migrating from one network node to another Beside these advantages, some problems still exist [6,19]: x The security aspect or how to protect the network nodes from malicious agents. x The network load could be higher and not lower if there are no strategies established under which circumstances an agent would migrate. x The ownership and charging policies are other very important issues. ....

....session. The mobile devices are getting smaller, but the wireless access to the services will remain slow due the relatively low bandwidth connection. The mobile agent can perform information retrieval and filtering at the server, and then return only the relevant information to the client [6]. In our case the opposite direction can be used as well (i.e. client server) The mobile agent can be sent to a cellular phone to extend the phone s capabilities with the latest software features or fulfill certain administrative tasks for the service providers. In the area of cellular phones ....

Colin G. Harrison, David M. Chess, Aaron Kershenbaum, "Mobile Agents: Are they a good idea?", Technical Report, IBM T.J. Watson Research Center, 1995.

.... that can freely roam the network and act on behalf of their users [1] A slightly more technical definition suggests that mobile agents are programs, typically written in a script language, which may be dispatched from a client computer and transported to a remote server computer for execution [2]. From a historical perspective, the work on distributed computing led to mobile agents following a route through the problem of process migration and the concept of mobile objects [1] The distinguishing characteristic, in terms of being agents , of such migrating processes or mobile ....

C. G. Harrison, D. Chess, and A. Kershenbaum, "Mobile Agents: Are they a good idea?," T.J. Watson Research Center, Yorktown Heights, NY, Research Report 1994.

....by offering the possibility of a widely deployed application that could use mobile agent technology. The research community visualizes mobile agents launched via web browsers to gather information and interact with any node in the network. IBM and General Magic were early pioneers of this vision, [CHES95, HARR95]. Concurrent with this effort, ARPA sponsored a Knowledge Sharing program. The KQML language [FINI94] was developed under this program and remains one of the viable Agent Communication Languages (ACLs) This research area was reformulated in the 95 96 time frame when Java was released by Sun ....

Harrison, C.G., D.M. Chess, A. Kershenbaum, "Mobile Agents: Are they a good idea?," IBM Research Report, March 1995.

....of the agents i.e. their ability to migrate from one place to another is the basic difference from other approaches for distributed systems. Major advantages of mobile agents are seen in the possibility of reducing (expensive) global communication costs by moving the computation to the data [Chess et al. 1997] and in the possibility to easily distribute complex computations onto several, possibly heterogenous, hosts. Straer and Schwehm 1997] deals with the first of this two advantages and presents a performance model regarding network load and execution time which can help to identify situations for ....

Chess, D. and Harrison, C. and Kershenbaum, A. (1997), "Mobile Agents: Are They a Good Idea?", In: Vitek, J. and Tschudin, C. (ed) Mobile Object Systems. Towards the Programmable Internet. Second International Workshop, MO'96. Selected Presentations and Invited Papers, Springer, Berlin, Germany, pp.

....here. The agents are designed to respond with human like intelligence [Hermans96] or human like emotions such as happiness, anger, frustration etc. Emotional agents such as those in the Oz project at CMU [Bates92] are targeted primarily for entertainment. 15 2.4.4. Mobile Agents A mobile agent [HCK95] is a piece of software that either resides in a mobile hardware platform (robots) or a piece of software that can move from computer to computer across a network. Mobility provides the agents the ability to explore physical and virtual spaces. In addition, mobility is also desired for software ....

.... Open to new agents Message Passing Lots of Negotiation Assist collaboration between many humans Human computer interaction Anthropomorphic Agents [Bates92] High Autonomy Asynchronous Not Mobile Very Complex Usually single agent Message Passing Lots of Negotiation Entertainment Mobile Agents [Brooks91a, HCK95] High Autonomy Asynchronous Mobile Fairly Complex Usually Flat Varied agent types Open to new agents Message Passing Variable Negotiation Exploration of physical and virtual spaces. Information Gathering Actor Systems [Agha86] High Autonomy Synchronous Not mobile Complex Flat Single type of ....

Colin G. Harrison, David M. Chess, Aaron Kershenbaum, "Mobile Agents: Are they a good idea?," Technical Report, IBM Research Division, T.J.Watson Research Center, NY, March, 1995,

No context found.

David Chess, Colin Harrison, Aaron Kershenbaum "Mobile Agents: Are They a Good Idea?" IBM Research Report RC

No context found.

C.G. Harrison, D.M. Chess, and A. Kershenbaum, "Mobile Agents: Are They a Good Idea?" Mobile Object Systems: Toward the Programmable Internet, Vol. 1222 of Lecture Notes in Computer Science, J. Vitek and C. Tschudin, eds., Springer-Verlag, New York, 1997, pp. 25-47.

No context found.

Colin G. Harrison, David M.chess, Aaron Kershenbaum, "Mobile Agents: Are they a good idea?", IBM T.J Watson Research Center, 28 March 1995.

No context found.

D. Chess and C. Harrison, A. Kershenbaum, "Mobile Agents: Are They a Good Idea?", In Lecture Notes in Computer Science, Vol. 1222, p. 25-47, 1997.

No context found.

HCK95 C.G.Harrison, D.M.Chess, A.Kershenbaum: "Mobile Agents: Are they a good idea?". IBM Research Divsion, Watson Research Center, 1995.

No context found.

Harrison, Colin G. "Mobile Agents: Are They a Good Idea?" IBM Research Report, Watson Research Center, Yorktown Heights, NY, March, 1995.

No context found.

Chess, D., Harrison, C., Kershenbaum, A., (1997), "Mobile Agents: Are they a Good Idea?" in Vitec (1997).

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC