R. Gray, D. Rus, and D. Kotz, Transportable Information Agents, Technical Report PCS-TR96-278, Department of Computer-Science, Dartmouth College, 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....logic; Stratified programs; Agent development environments 1 Introduction Over the last few years, there has been tremendous interest in the area of intelligent software agents. Such agents provide a wide range of services, ranging from data mediation agents [5, 6, 12, 14, 26] to mobile agents [34], to personalized visualization agents [13, 23] to agents that monitor newspapers, prioritize mail buffers and the like [18, 21, 36, 38] Most such existing work on agents subscribes to the view that agents should be autonomous, and that such autonomous agents should behave according to a ....

D. Rus, R. Gray, and D. Kotz. Transportable information agents. In M. Huhns and M. Singh, editors, Readings in Agents, pages 283--291. Morgan Kaufmann, 1997. 56

....that it should have control over its own actions and internal state [11] In the context of mobile agents, it is even more difficult to define autonomy. Having investigated some Web applications such as Internet information retrieval and mobile computation for which mobile agents are suitable [6][20], we identify the necessary features that an autonomous mobile agent should have. By autonomy, a mobile agent is self contained in navigation, computation and communication when transporting through the underlying heterogeneous and dynamic networks. An autonomous mobile agent should be aware of ....

Rus, D., Gray, R., S., and Kotz, D. (1997): Transportable Information Agents. In Proc. of International Conference on Autonomous Agents, 228-236.

....the focus of much attention in the AI and distributed systems communities. Mobile agents are program segments sent across a network which execute on host machines (very much like a friendly virus) Their aim is to perform some task on behalf of the user with a certain degree of autonomy (see [29] for an overview) Proposed uses include bartering, negotiating, entertainment, monitoring, data selection and ltration, searching, and distributed processing. Current suggestions for payment schemes are not well adapted to use with mobile agents: if an agent carries digital cash, for instance, ....

D. Rus, R. Gray and D. Kotz, "Transportable Information Agents", 1st Intl. Conf. Autonomous Agents, 1997.

....Other uses of agents for monitoring, such as in JAT[4] and WHERE[3] do not support the migration of agents between processes of the computation. Thus, these agents lack the ability react locally and to track non local properties as easily. General purpose mobile agent systems, such as AgentTcl[7], are not wellsuited for the tasks of monitoring because they provide high level services such as name services, authentication, and or network references that are not necessary for monitoring and incur extra perturbation. General purpose systems are typically designed to serve as stand alone ....

Rus, D., Gray, R., Kotz, D.: Transportable Information Agents. Readings in Agents. (1998) 283-291

....be performed at a process or an object level. The mobile agent model is a generalisation of the autonomous agent model. A mobile agent is capable of migrating through some programmed itinerary or it can design its own itinerary, dependent on dynamic properties. Examples of this category include [72, 113, 115, 156]. Mobile agents are inherently not migration transparent due to their autonomous nature. The code on demand model is used in systems where code repositories or libraries are essential to the system s execution. These libraries can be dynamically updated, thereby producing adaptive systems. Code ....

Daniela Rus, Robert Gray and David Kotz. Transportable Information Agents. In International Conference on Autonomous Agents, pages 228--236. ACM Press, February 1997.

....in weak mobility the code is restarted on arrival at the remote destination. Java only supports weak mobility, because it does not allow access to the execution stack. Nonetheless, it has been used by many mobile agent platforms such as Voyager [23] Aglets [32] HIVE [35] MOLE [51] D Agents [45], and many others (see [26] for a more comprehensive list) Strong mobility can be provided by modifying the Java Virtual Machine (JVM) which has been done in the D Agents platform [45] 2 Load balancing with Mobile Agents Mobile agents are promoted as infrastructure for information agents [13, ....

.... has been used by many mobile agent platforms such as Voyager [23] Aglets [32] HIVE [35] MOLE [51] D Agents [45] and many others (see [26] for a more comprehensive list) Strong mobility can be provided by modifying the Java Virtual Machine (JVM) which has been done in the D Agents platform [45]. 2 Load balancing with Mobile Agents Mobile agents are promoted as infrastructure for information agents [13, 53, 44, 20, 10] as they potentially decrease the data, which has to be transmitted, and enable asynchronous computing, in that not all infrastructure has to be available all the time. ....

[Article contains additional citation context not shown here]

D. Rus, R. Gray, and D. Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9, 1997.

....its owner [GHNCSE97] In this paper, we are concerned with the mobile software agents. The mobile agents research is being concentrated in areas such as (telecommunication) network management, electronic commerce, load balancing, fault tolerance and mobile computing. In this context, Rus et al. [RGK97] defines a mobile agent as an object that migrates through many hosts in a heterogeneous network, under its own control, in order to perform tasks using resources of these hosts. 5 2.2.2 Mobile Agents versus Client Server Mobile agents are autonomous entities, independent from the applications ....

....[CHK94; KT98, RGK97] describes many applications that can benefit from the use of the mobile agent paradigm. Some examples are described as follows: Mobile computing: In applications evolving mobility, the presence of a network connection is intermittent, or has variable bandwidth rates [RGK97]. Mobile agents can migrate to mobile hosts, perform their activities and move out when the network connection allows to. Fault tolerance and Load Balancing: Tasks and processes can be split in small subprocesses in order to perform their goal. These subtasks can be configured to migrate form ....

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. Proceedings of the first ACM international conference on Autonomous agents, 1997, pp. 228 -- 236.

.... program that acts in behalf of its owner (agent owner) GHNCSE97] A mobile software agent, or a software agent for now on, is an object that migrates through many nodes of a heterogeneous network of computers, under its own control, in order to perform tasks using resources of these nodes [IH99; RGK97] It travels from node to node of a distributed system performing tasks in behalf of its owner. At the end of this process, an agent can return to its home site and report itself to the user who injected this object in the distributed system [KT98] Mobile agents are used in the development of ....

....Moreover, as the interaction between the agent and the resource (after moving) is performed in the same host, without the transmission of messages through the network, this paradigm is indicated for some kinds of real time distributed applications. 4. Applications The literature [CHK94; KT98, RGK97, GKNRC96, LO99] describes many applications that can benefit from the use of the mobile agent paradigm. These are mobile computing, fault tolerance, load balancing, workflow management and electronic commerce. Additionally, new applications as runtime software change and software deployment can ....

[Article contains additional citation context not shown here]

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. Proceedings of the First ACM International Conference on Autonomous Agents , 1997, pp. 228 -- 236.

....be disconnected from the network. In such cases, you can send off the agent onto the network first, then disconnect your computer and connect it again somewhere else. The agent will find you in your new location and dock with your computer again, providing you with information you have requested [33]. Agent systems can use one of two approaches; a federated approach or a fully autonomous [27] In a federated approach agents are not truly autonomous and relies on support from a facilitator, an agent host. All communication goes through this facilitator and it provides the agents with their ....

Rus, D., Gray, R., Kotz, D. "Transportable Information Agents", Proceedings of the International Conference on Autonomous Agents, pp. 228-236, 1997.

....and intrusion route tracing in local area networks have been completed. We created a development environment for IDA as described below. ffl Hardware Sun SS10, Ultra5, Ultra10, Ultra30 ffl Operating System Sun Solaris 2.5.1 ffl Languages GNU C C 2.7.2.2 perl 5.004.01 D Agent 2. 0[8] ffl Auditing Sun Solaris BSM[9] Basic Security Module) We adopted the Sun BSM as the audit mechanism in IDA, which is the Solaris standard module, and it can audit all users events at the system level. The auditing mechanism of the BSM is evaluated at the C2 level. In IDA, mobile agents are ....

R. Gray, D. Rus, and D. Kotz, "Transportable Information Agents," Technical Report TR96-278, Department of Computer Science, Dartmouth College, 1996. http://www.cs.dartmouth.edu/ agent/

....the focus of much attention in the AI and distributed systems communities. Mobile agents are program segments sent across a network which execute on host machines (very much like a friendly virus) Their aim is to perform some task on behalf of the user with a certain degree of autonomy (see [29] for an overview) Proposed uses include bartering, negotiating, entertainment, monitoring, data selection and filtration, searching, and distributed processing. Current suggestions for payment schemes are not well adapted to use with mobile agents: if an agent carries digital cash, for instance, ....

D. Rus, R. Gray and D. Kotz, "Transportable Information Agents", 1st Intl. Conf. Autonomous Agents, 1997.

....and intrusionroute tracing in local area networks has been completed. We created a developing environment of IDA as described below. ffl Hardware Sun SS10, Ultra5, Ultra10, Ultra30 ffl Operating System Sun Solaris2.5.1 ffl Languages GNU C C 2.7.2.2 perl 5.004.01 D Agent 2. 0[5] ffl Auditing Sun Solaris BSM[6] Basic Security Module) In IDA, mobile agents are written by D Agent, which offers PGP, and this is able to authenticate and encrypt the mobile agents[7] 4.2 Experiment and Evaluation 4.2.1 Local Attack Detection The goal of IDA is not to detect all ....

R. Gray, D. Rus, and D. Kotz, "Transportable Information Agents," Technical Report TR96-278, Department of Computer Science, Dartmouth College, 1996. http://www.cs.dartmouth.edu/ agent/

....of most IR systems. Is SodaBot necessary Well, it is certainly required for our interactions with the Intelligent Room s existing infrastructure of agents. However, there are several other agent systems which provide similar levels of network accessibility and easy user interaction. AgentTCL[14], and to some extent TeleScript[48] could be used to build a similar infrastructure. Given the functionality gained by building our application in the context of Intelligent Room, SodaBot is the logical choice. 3.1 SodaBot SodaBot[6] is a software agent environment developed at the MIT ....

R. S. Gray, D. Rus, and D. Kotz. Transportable information agents. Technical Report PCS-TR96-278, Dartmouth College, Computer Science, Hanover, NH, February 1996.

....machine, improving network bandwidth. Secure languages and protocols are needed before information providers will welcome trusted autonomous agents to their servers; agent research is providing systems technology with the thrust that may enable such mobile agents developments feasible very soon [19]. Many different models of interaction among agents are also worth exploration. For example, research is under way into the issues of agents learning from other agents, agent collaboration, and agent communication languages. The first form of direct agent interaction that we are going to address ....

D. Rus, R. Gray, and D. Kotz. Transportable information agents. In 1st Intl. Conf. Autonomous Agents, 1997.

....for itself based on its environment, the agent is capable of continuing with its tasks autonomously, even when the client at which it was generated is unavailable. A body of research exists that has focused on disconnected autonomy as a prime lever for information retrieval and dissemination[14][19], but this capability is equally applicable to shop floor traceability, 6 order tracking, and sales order processing. This feature is exploited within the work described in this paper to enable sales staff to participate in the sales order processing system using mobile clients on laptop PCs. ....

Rus, D., Gray, R., Kotz, D., "Transportable information agents", Journal of Intelligent Information Systems, 9:215-238, 1997

....files and Panda s optimization engine is able to automatically select an efficient I O policy, based on a description of the application I O requests generated by VFC (see Fig. 1) and the system configuration. ffl MAS a library which provides interface to the mobile agent system D Agents [13] developed at the Dartmouth College Hanover. By means of agents, the application program can receive data selected from remote files. 1 In the SPMD (Single Program Multiple Data) model, each processor executes the same program on different parts of the data space. These libraries interact with ....

....DO Fig. 6. Specifying Checkpoint Restart Operations 3.3 Remote File Access Using Mobile Agents Traditional approaches to distributed information access co locate the data and the computation needed to process it by bringing the data to the computation. The developers of the mobile agent technology [13] advocate a novel approach that brings the computation to the data in the form of mobile agents. A mobile agent is a program that can migrate from machine to machine in a heterogeneous network. This approach requires an agent to have substantial intelligence in accessing remote files, making ....

D. Rus, R. Gray, D. Kotz, Transportable Information Agents, Journal of Intelligent Information Systems 9 (Kluwer Acad. Publ., Boston, 1997) 215-238.

....concentration on fundamental issues and partly because the ACL alludes to a declarative approach that contradicts the procedural approach implicit in the mobile agent paradigm. There are some exceptions to this situation, of course, such as some experimental work in integrating Agent Tcl and KQML [32]. 5. Current ACLs The Knowledge Sharing Effort [26, 29] KSE) was initiated circa 1990 by DARPA and it enjoyed the participation of dozens of researchers from both academia and industry. Its goal was to develop techniques, methodologies and software tools for knowledge sharing and knowledge ....

D. Rus, R. Gray, and D. Kotz. Transportable information agents. In 1997 International Conference on Autonomous Agents, Marina del Ray, CA, 1997.

....in trading off CPU cycles for improved network bandwidth. Secure languages and protocols are surely needed before trusted autonomous agents will become a welcome reality; agent research is providing systems 42 technology with the thrust that may very soon make such mobile agents possible [42]. Finally, the feasibility of integrating agent based on line search with index based search engines must be put to the test. Hybrid systems can be constructed in which search engines provide agents with good staring points, based on the statistical (word based) topology of the search space. This ....

D Rus, R Gray, and D Kotz. Transportable information agents. In Proc. 1st International Conference on Autonomous Agents, 1997.

.... (Tambe et al. 1995) and entertainment (Mock, Lawton, Hoyle 1996) In the context of increasing interconnectivity, mobile computing is of increasing interest (Harrison 1995) Mobile agents are capable of roaming entire networks and traveling from one machine to another (Coen 1994; Gray 1995; Gray, Rus, Kotz 1996; General Magic, Inc. 1995; Sun Microsystems, Inc. 1996) Other recent approaches to software agents are strongly influenced by Artificial Life, where biological tenets of evolution and genetics are applied to the development of agents. The application of artificial life techniques to an ....

Gray, R. S.; Rus, D.; and Kotz, D. 1996. Transportable information agents. submitted to AAAI96.

....business rules for the successful execution of each step of the task. In short, transportable agent based workflow management systems can be more efficient, flexible and fault tolerant. 4. The Agent Tcl System For our research we are using a transportable agent system called Agent Tcl. Agent Tcl [Gra95, Gra96] is a system for transportable agents currently being developed at Dartmouth College. Agent Tcl extends the standard Tool Command Language (Tcl) Ous90] a high level scripting language that was developed in 1987 and has enjoyed enormous popularity. Agent Tcl addresses four main goals: Reduce ....

R. Gray, D. Rus, D. Kotz. Transportable Information Agents. Technical Report: PCSTR96 -278, Department of Computer Science, Dartmouth College, 1996.

No context found.

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9:215-238, 1997.

....with other mobile agent systems, touches on only some of the more important features of a mobile agent system. Again, Tables 1 and 2 summarize our comparison effort, and include citations to appropriate literature. The D Agents system is explained in more detail in several earlier papers [RGK97, KGN 97, GKCR97, Gra97, CCMG98, GKCR98, BGM 99, GCKR01, KCG 02] Despite the differences described above, and other small differences, all of the systems discussed above (with the exception of Messengers and CODE, which are lighter weight mobile code systems) are intended for the same ....

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9:215--238, 1997.

No context found.

D. Rus, R. Gray, and D. Kotz, "Transportable Information Agents", Intelligent Information Systems, vol 9. pp 215-238, 1997.

....documents into (automatically generated) categories, and (4) summarize each category with a few distinctive keywords. Figure 1 shows the structure of the technical report application. The stationary interface agents, which appear at the bottom of the gure, register with the yellow pages [RGK97] when they begin execution. The yellow pages, which are not shown in the gure, are a simple distributed directory service. When the interface agent registers, it provides its location (i.e. its identi er within the D Agents namespace) and a keyword description of its service (i.e. ....

....a small mobile agent onto the local machine. The agent s task is to nd the available interface agents and retrieve the relevant technical reports. The agent can go about the retrieval task in several ways. It makes two main decisions after consulting with a simple set of network sensors [RGK97] First, if the network connection between the user s machine and the network is reliable and has high bandwidth, the agent stays on the user s machine. If the connection is unreliable or has low bandwidth, the agent jumps to a proxy site within the permanent network. This proxy site is shown in ....

[Article contains additional citation context not shown here]

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9:215-238, 1997.

No context found.

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. Journal of Intelligent Information Systems, vol 9. pp 215-238, 1997.

....are currently implemented in D Agents. Since a later section is devoted to describing these techniques, we will present related work in that section. 3 D Agents D Agents is a mobile agent system whose agents can be written in Tcl Java and Scheme 1 D Agents has extensive navigation services [23], security mechanisms [9] and debugging and tracking tools [11] In addition, it is active use at numerous academic and industrial research labs, including labs at Lockheed Martin, Siemens, Cornell, and the University of Bordeaux, and is starting to find its way into production quality ....

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, May 1997. To appear.

....(3) obtain pair wise similarity scores for every pair of documents in a list of documents. The pair wise similarity scores are used to construct different graphical representations of the query results. When each stationary Smart agent starts execution, it registers with a virtual yellow pages [RGK97] The yellow pages are a simple, distributed, hierarchical directory service. When the Smart agent registers with the yellow pages, it provides its location (i.e. its identifier within the D Agents namespace) and a set of keywords that describe its service (i.e. smart, technical reports, text) ....

....The GUI first lets the user enter a free text query and optionally select specific document collections from a list of known document collections. Once the GUI has the query, it spawns a mobile agent onto the local machine. This mobile agent first consults one or more local network sensing agents [RGK97, Moi98] which keep track of the network connection between the user s machine and the rest of the network. These network sensing agents know what type of network hardware is in the machine, the maximum bandwidth of that hardware, an uptime downtime history of the network link, and the current ....

[Article contains additional citation context not shown here]

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9:215--238, 1997.

....Thus, mobile agents co locate data and computation by bringing the computation to the data, rather than by bringing the data to the computation. Agents have the necessary autonomy to make decisions, and to interact with other agents and services to accomplish their goals. Our previous research [RGK97, KGN 97] shows that mobile agents have tremendous promise as a uniform paradigm for developing distributed applications, primarily because agents are easier to write than message or RPC based applications, their autonomy makes them well suited to poorly connected network environments, and ....

....of a single transaction and the e#ect of defaulting on one s reputation. By adding a penalty for decommitting [SL96] agents can be persuaded to act in responsible fashion, while still having the flexibility to back out of a transaction in an emergency. 3 D Agents: a Mobile Agent System D Agents [RGK97, Gra97, KGN 97] was developed to support mobile agents written in Agent Tcl, Agent Java, and Agent Scheme (extensions of Tcl, Java, and Scheme, respectively) over the past two years. The primary goal of the project has been to implement a computational paradigm that co locates data and ....

[Article contains additional citation context not shown here]

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems, 9:215--238, 1997.

....between users and agents and propose statistical and machine learning methods for building user models to control the agent actions. Rus and Subramanian [24, 23] propose a modular, open, and customizable agent architecture organized around a notion of structure recognition. In our previous work [7, 8, 20, 10, 16, 25] we describe other aspects of transportable agents in Agent Tcl. 3. Agent Tcl: a Transportable Agent System Like all mobile agent systems, the main component of Agent Tcl 4 is a server that runs one each machine. When an agent wants to migrate to a new machine, it calls a single function, ....

D. Rus, R. Gray, and D. Kotz, Transportable Information Agents, in Proceedings of the 1997 International Conference on Autonomous Agents, Marina del Ray, California, 1997.

....possible to generate a centralized, multi level index of the corpus. When the index does not exist or is distributed, the problem becomes more difficult as resource discovery becomes a sub task. GGT93] We are currently pursuing an approach to information gathering that uses transportable agents. [Gra94, GRK96] Transportable agents are programs that can suspend execution in the middle, migrate to a different machine, and restart from where they left off. Our goal is to use the structural indices and the organization structures computed here as navigation tools to guide the motion of information ....

R. Gray, D. Rus, and D. Kotz. Transportable Information Agents. Technical Report PCS-TR96-278, Department of Computer Science, Dartmouth College, 1996.

....Thus, mobile agents co locate data and computation by bringing the computation to the data, rather than by bringing the data to the computation. Agents have the necessary autonomy to make decisions, and to interact with other agents and services to accomplish their goals. Our previous research [RGK, KGN 97] shows that mobile agents have tremendous promise as a uniform paradigm for developing distributed applications, primarily because agents are easier to write than message or RPC based applications, their autonomy makes them well suited to poorly connected network environments, and ....

....a single transaction and the effect of defaulting on one s reputation. By adding a penalty for decommitting [SL96] agents can be persuaded to act in responsible fashion, while still having the flexibility to back out of a transaction in an emergency. 3 D Agents: a Mobile Agent System D Agents [RGK, Gra97, KGN 97] was developed to support mobile agents written in Agent Tcl, Agent Java, and Agent Scheme (extensions of Tcl, Java, and Scheme, respectively) over the past two years. The primary goal of the project has been to implement a computational paradigm that co locates data and ....

[Article contains additional citation context not shown here]

Daniela Rus, Robert Gray, and David Kotz. Transportable information agents. Journal of Intelligent Information Systems. To appear.

No context found.

R. Gray, D. Rus, and D. Kotz, Transportable Information Agents, Technical Report PCS-TR96-278, Department of Computer-Science, Dartmouth College, 1996.

No context found.

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. In Huhns M.N. and M. Singh, editors, Readings in Agents, pages 283-291, San Francisco, California, 1997. Morgan Kaufmann Publishers.

No context found.

D. Rus, R. Gray, and D. Kotz, Transportable Information Agents, Journal of Intelligent Information Systems, vol. 9, pp. 215-238, 1997.

No context found.

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. In M. Huhns and M. Singh, editors, Readings in Agents, pages 283-291. Morgan Kaufmann Press, 1997.

No context found.

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. In M. Huhns and M. Singh, editors, Readings in Agents, pages 283--291. Morgan Kaufmann Press, 1997.

No context found.

D. Rus, R. Gray, and D. Kotz. Transportable Information Agents. In M. Huhns and M. Singh, editors, Readings in Agents, pages 283--291. Morgan Kaufmann Press, 1997.

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