G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proceedings of High Performance Distributed Computing (HPDC), 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....not lost if they cannot be delivered immediately. Java based and C based APIs are available for XMessages [23] XMessages uses SOAP and XML as the wire format. A JMS Bridge is also available that allows for messages to be exchanged between JMS peers and the XMessage framework. 3. 7 Echo Echo [8] is also a hybrid of subject based and content based event communication system. Its architecture we categorize as peer to peer graph topology. ECho provides an efficient lightweight implementation of CORBA style event channels. Event channels 12 are created by a provider and subscribed by ....

Greg Eisenhauer, Fabian Bustamente, and Karsten Schwan. Event services for high performance computing. In Proc. 9th IEEE Intl. High Performance Distributed Computing (HPDC), Los Alamitos, CA, August 2000. IEEE Computer Society.

....such as OpenORB [3] and Bossa Nova [14] o#er a flexible infrastructure that supports QoSaware composition and reflection. While these frameworks do not provide specific streaming support, they can serve as a basis for building information flow middleware. Event based middleware such as Echo [7, 11] provides a type safe and e# cient way of communicating data and control information in a distributed and heterogeneous environment. A higher level Infopipe layer can also be built on top of these platforms. Ensemble [34] and DaCaPo [27] are protocol frameworks that support the composition and ....

G. Eisenhauer, F. Bustamante, and K. Schwan. Event services for high performance computing. In International Conference on High Performance Distributed Computing (HPDC), August 2000.

No context found.

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proceedings of High Performance Distributed Computing (HPDC), 2000.

No context found.

G. Eisenhauer, F. Bustamante, and K. Schwan, \Event Services for High Performance Computing," in Proceedings of High Performance Distributed Computing (HPDC), 2000.

No context found.

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proc. of HPDC9, Pittsburg, PA, Aug. 2000.

No context found.

G. Eisenhauer, F. Bustamente, and K. Schwan. Event services for high performance computing. In Proceedings of High Performance Distributed Computing-9(HPDC9) , August 2000.

....CPU load on maui, and network, CPU load and disk usage on etna. Dproc makes this information available at each of these three nodes in the same hierarchy as shown in Figure 1. The communications infrastructure is implemented using a kernel implementation of the ECho event channel infrastructure [5], called KECho [16] KECho provides a publish subscribe mechanism for direct kernel kernel communications. Each kernel connects to both a data communication channel, or monitoring channel, and a control channel. To improve scalability, the exchange of data is triggered only when an application ....

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proceedings of High Performance Distributed Computing (HPDC), 2000.

.... information captured by remote instruments[7] In addition, real time collaboration tools[4] permit evaluation and discussion of results and insights with remote colleagues [10] Finally, modern portal technologies[1, 11] and their underlying publish subscribe communication infrastructures[5, 18] not only enable the real time viewing and inspection of the results of remote simulations and or instruments, but also their online steering and control[14, 17] Portals and Virtual Workbenches for Scientists and Engineers. The purpose of portals is to give users remote access to resources ....

....JNI Java Mirror Object MOSS Mirror Java Object User Interface ECho JECho JECho JECho Event Figure 1. Mirror Object Architecture and Experiment Configuration data, each in ways customized to their individual needs. Implementation. The ECho and JECho publish subscribe event infrastructures[5, 18] implement the communications within mirror object structures, as depicted in Figure 1. Specifically, each mirror object subscribes to the event channels to which the application level monitoring events they desire are sent. Thus, any one mirror object can receive events emanating from any number ....

[Article contains additional citation context not shown here]

G. Eisenhauer, F. Bustamente, and K. Schwan. Event services for high performance computing. In Proceedings of High Performance Distributed Computing-9(HPDC-9), August 2000.

.... on runtime code specialization [14] and on code generation of embedded systems [11, 21, 32] Our research is based on our previous work with distributed, real time, adaptive, and multimedia systems [24, 27, 30] with sensor based embedded applications [23] and with high performance middleware [4, 9] and applications [31] We are also leveraging prior large scale funding from the National Science Foundation and the Department of Energy that has created the basic publish subscribe middleware used in this work [4, 33] and is currently creating the new network measurement techniques [10, 22] and ....

.... sensor based embedded applications [23] and with high performance middleware [4, 9] and applications [31] We are also leveraging prior large scale funding from the National Science Foundation and the Department of Energy that has created the basic publish subscribe middleware used in this work [4, 33] and is currently creating the new network measurement techniques [10, 22] and middleware mechanisms [7] needed for making middleware platform aware. Compiler techniques are focused on embedded devices, including static analysis methods like on chip memory data allocation [28] restructuring for ....

[Article contains additional citation context not shown here]

G. Eisenhauer, F. Bustamente, and K. Schwan. Event services for high performance computing. In Proceedings of High Performance Distributed Computing (HPDC-2000.

....wasting a significant amount of bandwidth. A more complete discussion analyzing the causes of the performance differences in more detail and also presenting an analysis of the behavior of these communications systems in the context of application evolution and type extension can be found in [4]. 3.2 Effects of Stream Specialization The primary benefit of stream specialization is the elimination of wasteful message traffic and the subsequent reduction of bandwidth requirements. As such, many of the benefits that are achievable through such specialization vary significantly with the ....

Greg Eisenhauer, Fabian E. Bustamante, and Karsten Schwan. Event services for high performance computing. To appear at High Performance Distributed Computing - HPDC'2000.

....load conditions, using the example of a web server s responses to its current vs. newly arriving requests. Previous work. Our research is based on our previous work with distributed, real time and adaptive systems, with sensor based embedded applications [10] and with high performance middleware [2]. Previous and ongoing related work performed by project members includes funding from DARPA s Quorum, PCES, and embedded systems programs. We are also leveraging prior large scale funding from DARPA and from the National Science Foundation that has created the basic publish subscribe middleware ....

....related work performed by project members includes funding from DARPA s Quorum, PCES, and embedded systems programs. We are also leveraging prior large scale funding from DARPA and from the National Science Foundation that has created the basic publish subscribe middleware used in this work [2, 13]. Concerning compiler technologies for embedded devices, we have investigated new static analysis techniques like on chip memory data allocation [11] restructuring for code compaction [5] and efficient data layouts for indirect addressing modes [9] The InfoFabric project goes beyond such work ....

G. Eisenhauer, F. Bustamente, and K. Schwan. Event services for high performance computing. In Proceedings of High Performance Distributed Computing (HPDC-2000.

....and its synchronization mechanisms appears in [14] 4 Evaluation Experiments are performed with the replicated server (REDE and mirror sites) running on up to eight nodes of a cluster of 300MHz Pentium III dual processor servers running Solaris 5.5.1. The ECho event communication infrastructure [11] is used to efficiently move data events across nodes. The flight positions data stream used in these experiments originates from a demo replay of original FAA streams, and it contains flight position entries for 50 different flights. The evaluation metric is the total execution time of the ....

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proc. of Ninth High Performance Distributed Computing (HPDC-9), Pittsburg, PA, Aug. 2000.

....implementation of its dynamic code generation framework. 3. A Proactive Directory Service Prototype To validate our approach we have designed and implemented PDS, a prototype of a proactive directory service. The PDS prototype is implemented in C C on top of an event communication mechanism [10] and makes use of an efficient and extensible binary transport for communication in heterogeneous environments [2] The PDS architecture includes three main components: PDS clients, servers and objects owners (Figure 1) PDS clients want to discover available objects in the environment and become ....

....constructs and a richer specification language that allows compound matching events and QoS directives. Elvin [30] is a centralized event dispatcher with an expressive event filtering language. Siena [6] and Gryphon [34] are both content based message brokering systems while our group s ECho [10] and the Java Distributed Event Specification [35] adopt a channel based addressing scheme. Yu et. al [44] proposed an event notification service with a novel peer to peer architecture of proxy servers. The CORBA Event Service [25] and the Notification Service [24] specification, as well as the ....

G. Eisenhauer, F. E. Bustamante, and K. Schwan. Event services for high performance computing. In Proc. of the 9th High Performance Distributed Computing (HPDC-9), Pittsburgh, PA, August 2000. IEEE.

....down processing on a mobile device. You can also observe that at some frequencies the energy consumption rises. We disregard this behavior in this paper, but point out that others (e.g. in [9] have investigated this issue. Solution Approach. Event services or publish subscribe mechanisms [2, 4, 7, 16] are increasingly being deployed in applications that range from remote sensing, to multimedia and video streaming [1, 10] to transactional systems [3] We adopt this approach to enable per frame power management for multimedia applications. That is, video or audio frames are distributed as data ....

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proceedings of High Performance Distributed Computing (HPDC), 2000.

....parties must be flexible. This need for flexibility has led some designers to adopt techniques like Java s RMI or meta data representations such as XML SOAP. These methods have high costs which interfere with total performance, because data marshalling becomes a key issue. The ECho middleware[ECho,EBS00] addresses these concerns in several different ways. In particular, it provides the following: Information flows are represented as event streams, using the publish subscribe model. There is no centralization; channels are represented by distributed data structures. Connections are managed ....

G. Eisenhauer, F. Bustamente and K. Schwan. "Event Services for High Performance Computing". Proceedings of High Performance Distributed Computing (HPDC-2000).

....have been performed on a cluster of 8 Quad Pentium Pros with 200MHz each, 512 MB RAM, running Linux (version 2.4. 17) Figure 3 compares the cost associated with event submission in the kernel level event service, KECho, with the event submission of a similar user level implementation, called ECho [3]. Both ECho and KECho use TCP, although work 0 0.2 0.4 0.6 0.8 1 0 2 4 6 8 10 12 Update Frequency (1 s) Submission Overheads (100 sinks) Kernel, 100b Kernel, 2kB User, 100b User, 2kB Figure 3. Event submission overheads. 0 5 10 15 20 25 30 0 10 20 30 40 50 60 70 80 CPU Load ....

G. Eisenhauer, F. Bustamante, and K. Schwan. Event Services for High Performance Computing. In Proceedings of High Performance Distributed Computing (HPDC), 2000.

.... We have built a kernel based event service aimed at supporting the coordination among multiple kernel services in distributed systems [19] Event and publish subscribe services have become prevalent in distributed applications ranging from virtual reality, avionics, to support for mobile users [8, 21, 6]. Our current work investigates the utility of a kernel based event service for resource management, distributed monitoring, and load balancing mechanisms. 4 An Example in Cluster based Servers We have started to apply this work to the monitoring of cluster based Internet services. In ....

Greg Eisenhauer, Fabian E. Bustamante, and Karsten Schwan. Event services for high performance computing. In Proc. of the 9th High Performance Distributed Computing (HPDC-9), Pittsburgh, PA, August 2000.

....such cluster servers, which is their ability to continue to provide high levels of service even under increasing loads or when clients request behaviors vary dynamically. Operational Information Systems. In contrast to the interactive high performance applications considered in our previous work [5, 6], this paper addresses an emerging application domain for high performance computing and for the cluster servers on which these applications run, which is that of operational information systems [7] OIS) An OIS is a large scale, distributed system that provides continuous support for a company s ....

.... vary checkpointing frequency; and install a different mirroring function. With the API call set adapt( we identify which of these parameters are to be modified if the threshold value is surpassed, and by what extent. 3. 3 Implementation We use the ECho event communication infrastructure [6] to efficiently move data across machines. Thus, communication is carried out via multiple logical event channels, both between mirrors and among sources, mirrors, and clients. Two separate event channels exist between the auxiliary and the main unit on each site, as well as among the auxiliary ....

G. Eisenhauer, F. Bustamante, and K. Schwan, "Event Services for High Performance Computing", In Proc. of Ninth High Performance Distributed Computing (HPDC-9), Aug. 2000.

....event has no size restriction and contains timestamped data about the behavior or state of a computational entity, physical instrument, or user. Event streaming can be initiated by a data provider (push model) or by a consumer (pull model) Publish subscribe event communication packages like ECho [8] implement event channels similar to those provided by CORBA [14] but focused on large scale event flows. The publish subscribe semantics allow any number of consumers to subscribe to an event channel; the provider need not have knowledge of the location or number of users. Data streams have been ....

....by which one can create computational entities to manipulate data streams. A data stream is a logical group of one or more event channels, one or more suppliers, a single consumer, intermediate query driven computation called quoblets, and a relational schema. Data streams are built on top of ECho [8], an efficient publish subscribe event channel implementation. Data suppliers and consumers are linked via logical event channels that permit suppliers to publish events unbeknownst of the types and numbers of consumers. The relational data model consists of data organized into relations (or ....

[Article contains additional citation context not shown here]

Greg Eisenhauer, Fabian Bustamente, and Karsten Schwan. Event services for high performance computing. In IEEE International High Performance Distributed Computing (HPDC), 2000. 20

....Streams Framework. notifications can be dynamically tuned by the clients. These three components, together with the repository service previously mentioned, conform the core of the Active Stream framework (Figure 3) The whole framework relies on ECho, a high performance event infrastructure [5], for its communication requirements. The implementation of Active Streams is mostly complete, and we plan on making it available by December 2001. ....

G. Eisenhauer, F. E. Bustamante, and K. Schwan. Event services for high performance computing. In Proceedings of the 9th High Performance Distributed Computing (HPDC9) , Pittsburgh, PA, August 2000.

....computations serve to transform, aggregate, or filter the data. For instance, aggregation might be used to sum values over neighbor points in a 3D space to reduce downstream bandwidth needs. Additionally, our group has developed a CORBA style publish, subscribe event middleware layer, called ECho [2] which creates event channels as the transport vehicle. ECho relies upon a reader makes right data format, PBIO [3] for binary I O. Computational data streams are one of the underlying mechanisms of the Infosphere project [12] Their viability has been considered by others in [4] 1] and [8] ....

....the created channel as an input parameter to each of the processes at startup. The programmer must also explicitly code the data format descriptions in the form that the binary library demands. The need for using binary I O to stream data has been established by earlier measurements by our group [2] comparing the cost of binary I O to an XML encoded data stream transmitted as ASCII. We have approached the problem of explicitly having to create event channels and declaring binary formats by leveraging the meta information available to us by our relational view of data streams. Specifically, ....

Greg Eisenhauer, Fabian Bustamente, and Karsten Schwan. Event services for high performance computing. In IEEE International High Performance Distributed Computing (HPDC), 2000.

No context found.

G. Eisenhauer, F. Bustamante, and K. Schwan. Event services for high performance computing. In International Conferemce on High Performance Distributed Computing (HPDC), August 2000.

No context found.

Eisenhauer G, Bustamante F, Schwan K. Event services for high performance computing. In International Conference on High Performance Distributed Computing (HPDC). 2000; .

No context found.

G. Eisenhauer, F. Bustamente and K. Schwan, Event Services for High Performance Computing. Proceedings of High Performance Distributed Computing (HPDC-2000).

No context found.

EISENHAUER, G., BUSTAMENTE, F., AND SCHWAN, K. Event services for high performance computing. In IEEE International High Performance Distributed Computing (HPDC) (August 2000.

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