A. Fox, S.D. Gribble, Y. Chawathe, and E.A. Brewer, "Adapting to network and client variation via On-Demand Distillation", in Proceedings of the 16th Symposium on Operating Systems Principles (SOSP-16), Oct. 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....may leave an application unable to function. Adaptive applications are able to change their degree of communication or mode of operation to suit the currently available bandwidth. Examples of classic adaptive applications include streaming video playback and image retrieval during web browsing [2], 3] In response to low bandwidth, a video player might reduce the quality of the frames it displays, or a web browser might retrieve degraded versions of images appearing in web pages, instead of the original high quality versions. We call this modal adaptation, since the application has a set ....

....VI. RELATED WORK While there has been a great deal of research in adapting applications and protocols to variations in network characteristics, and in bandwidth division algorithms, for the sake of brevity we will mention only a few related projects. We have already proposed web browsing [2] and remote file systems [4] as specific applications which can adapt to bandwidth availability; several systems provide more general adaptation mechanisms to applications. Odyssey [3] allows applications on a mobile host to adapt to changes in availability of many kinds of resources; the ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir, "Adapting to network and client variation via on-demand dynamic distillation," in Proceedings of the Seventh International Conference on Architectural Support for Programming Languages and Operating Systems, Cambridge, Massachusetts, Oct. 1996, pp. 160--170.

....Units comprise the myriad of devices that may be connected to the infrastructure. The active proxies provide adaptation 12 of content (similar to our containers) and are the result of previous research in data distillation using the TACC [Fox] model to perform on the y data transformations [FGBA96, FGG 98] The last component, paths, constructs ows of data that may be transformed while passing through di erent components, using their active proxies [CMI] The Ninja architecture is investigating scalable and robust ways to build scalable Internet services. We are focusing on physical ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOSVII, Boston, MA, October 1996.

....the context of VoD, scalability also applies to the server s ability to support the data requirements of multiple terminal types. One way to solve this problem is proxy based transcoding, where data streams are individually transformed according to the speci cation of each requesting receiver [31]. However, it typically imposes an administrative burden because it is not transparent to end users. Proxies are also dicult to deploy because a user behind a constrained network link might not have access to the optimal location of a proxy. There was a proposal to use active networks to solve ....

....this problem by o ering a common platform for such services as part of the basic network service model [79] but there remain many issues to be addressed before such an infrastructure can be deployed. Furthermore, transcoding proxies must be highly reliable and scalable which can be very costly [31]. Another ecient solution to heterogeneity is the use of layered media formats. This scheme encodes source data as a series of layers, the lowest layer being called the base layer and higher layers being called the enhancement layers. Layered encoding can be e ectively combined with multicast ....

A. Fox et al., \Adapting to network and client variation via on-demand, dynamic distillation," Proc. ACM ASPLOS-VII, pp.160-170, Oct. 1996.

....of contexts. Adaptability has been proposed as a way to conserve power in PDAs [16] as well as to deal with changes in the quality of wireless network connectivity caused by mobility [2] It can be used in multimedia applications to adjust the amount of data transmitted to the available bandwidth [7], and in highly dependable systems to react to failures or security threats [3] Adaptive solutions have also been used extensively in networking protocols [13] to change the functionality of middleware layers dynamically [15] as the basis for backend storage server architectures [4] and for ....

A. Fox, S. Gribble, E. Brewer, and E. Amir. Adapting to network and client variation via on-demand, dynamic distillation. In Proceedings of the 7th ASPLOS Conference, Oct 1996.

.... different Quality of Media can be achieved by ffl media scaling [19] the dynamic manipulation of media contentasitisdelivered to the client# for instance, the base layer (BL) and enhancement layers (E1 and E2) in a MPEG 2 video stream require different bit rates, ffl dynamic distillation [4] the controlled degradation of quality to adapt to the client platform or network connectivity# for instance, to deliver a video stream to a hand held device connected via a wireless link, wemay reduce the color to grey scale images with low resolution. ffl softwarefeedback the ....

Fox et al. Adapting to Network and ClientVariation via On-Demand, Dynamic Distillation. In ASPLOSVII, Boston, MA, 1996.

....Units comprise the myriad of devices that may be connected to the infrastructure. The active proxies provide adaptation of content (similar to our containers) and are the result of previous research in data distillation using the TACC [6] model to perform on the y data transformations [8]. Transcoding data formats was found to greatly increase the performance of certain applications [7] The last component, paths, constructs ows of data that may be transformed while passing through di erent components, using their active proxies. These are similar to our channels. Our methodology ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOS-VII, Boston, MA, October 1996.

....The possibility of compressing HTTP messages seems to have a longer history, going back at least to the early drafts of the HTTP 1.0 specification. However, until recently, it appears that nobody had attempted to quantify the potential benefits of loss free compression, although the GloMop project [7] did explore the use of lossy compression. A study done at the World Wide Web Consortium reports on the benefits of compression in HTTP, but for only one example document [15] Also, our traces suggest that few existing client implementations offer to accept compressed encodings of arbitrary ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Transcoding. In Proc. ASPLOS VII, pages 160-170. Cambridge, MA, October, 1996.

....data compression, and caching. Content simplification works, to a point. For example, Web designers can use common sense to reduce page complexity, or special software tools to optimize image coding (e.g. 12] Transcoding proxies can reduce image complexity for transmission over slow links [17]. But some content, such as medical images, broadcast quality video, and executable software, cannot be simplified without loss of meaning. Data compression directly targets the transmission of redundant bits within a single transfer. Existing general purpose compression algorithms provide ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Transcoding. In Proc. ASPLOS VII, pages 160-170. Cambridge, MA, October, 1996.

....A replacement for standard le system organization has been proposed that logically treats les as nested boxes [BA99] Remote copy operations and converters are incorporated into the design. Data distillation has been used in the TACC [Fox] model to perform on the y data transformations [FGBA96] Transcoding data formats was found to greatly increase the performance of certain applications. The e ects of mobile code were evaluated on a le service [SHG98] It was found that moving operations to the le server (i.e. agrep) is typically advantageous when client CPU power is below that of ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOS-VII, Boston, MA, October 1996.

....A replacement for standard le system organization has been proposed that logically treats les as nested boxes [BA99] Remote copy operations and converters are incorporated into the design. Data distillation has been used in the TACC [Fox] model to perform on the y data transformations [FGBA96] Transcoding data formats was found to greatly increase the performance of certain applications. Our system performs similar transformations, but includes the management of data on disk, creating chains of adaptors, and provides a uniform interface for data access. The e ects of mobile code were ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOS-VII, Boston, MA, October 1996.

....file system. The three tier model with an intermediate node for data filtering is a com 2 The home computer is basically a machine which is designated as the mobile user s PFS server (e.g. the mobile users desktop workstation) 2 monly used technique to reduce bandwidth utilization [7][3]. Based on the above model, PFS provides the following services to applications: ffl Support for a wide range of connectivity. Hoarding for disconnected operation, and variable granularity caching and consistency for low bandwidth operation. ffl Intelligent use of wireless bandwidth. Only the ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to network and client variation via on-demand, dynamic distillation. In Seventh Symposium on Architectural Support for Programming Languages and Operating Systems, October 1996.

....(e.g. contained in a collection) have not been investigated. 6. Conclusions Document customization has been an emerging trend in the WWW. However, in the WWW most of the customization occurs at the original servers, like my.yahoo.com, or by designated servers for network or client adaptation [3]. The Placeless Documents system allows individual users to customize their documents through active properties. This generalization introduces interesting new issues to the problem of document caching: 1) per user versions of the same document need to be cached, 2) cache consistency depends ....

A. Fox, S. D. Gribble, E. Brewer, and E. Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In Proceedings of the ACM 7 th International Conference on Architectural Support for Programming Languages and Operating Systems, Boston, Oct. 1996

....about the web) several systems are now making use of computation at proxy servers. Instead of just using these for caching visited pages, proxies or other intermediaries are now being used for a variety of purposes, for example to filter and transform pages for display on small portable devices [11] or to counter transmission protocol problems [12] The way an application is split over these sites affects resources and efficiency, but most important it affects the rate of feedback for a user. There are neurologically fixed targets for this for some kinds of hand eye coordination ....

....purchasing a home PC is now to use the Internet. However, a PC is not the only entrance to the Internet. Many libraries have public Internet access, recasting their role as information (not simply book) providers. Mobile Internet access is growing using proxies to reformat pages for small screens [11]. Perhaps most significant is the fact that games consoles and TV set top boxes are beginning to have Internet connectivity. This has the potential to really break the social barriers to connectivity within developed countries (in the UK more households have television sets than bathrooms ) ....

A. Fox, S.D. Gribble, E.A. Brewer, and E. Amir. "Adapting to Network and Client Variation via On-Demand, Dynamic Distillation", Proceedings of ASPLOS-VII, Boston, Masachusetts, U.S., 1996

....an environment in which a wide range of devices may interact. However, some computers, such as palmtops, may not posses the resources needed to accept full blown services. In these cases, proxies may be introduced to transform the services such that they are appropriate for the connected device [2]. Proxies may be used to o load computations or lter continuous media to accommodate resource impoverished devices. Use of such proxies may leave the server and client una ected, placing the responsibility of determining the peculiarities of the di erent devices on a decoupled third party. ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOS-VII, Boston, MA, October 1996.

....the services a powerful desktop system may receive at an appropriately degraded level. One approach for providing adaption is the introduction of proxies [21] into the system. Proxies may be used to o load computations or to lter transform streams to accommodate resource impoverished devices [8]. Use of proxies may leave the server and desktop clients una ected, focusing the intelligence into a decoupled third party. Although the infrastructure may need some knowledge of the proxies for management purposes, the servers may remain generally una ected, thereby allowing them to ....

....algorithms. In such cases, transcodings may be used to service these devices. An application level video gateway [2] has been developed to connect sessions with di erent bandwidth requirements. Data distillation has been used in the TACC [6] model to perform on the y data transformations [8]. The system has been utilized to create the Top Gun Wingman [7] web browser for PalmPilots. HTML pages and embedded images are transformed for the limited resources of PalmPilots and was found to signi cantly improve download speed and page layout. Compressed domain transcodings of MPEG video ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation. In ASPLOS-VII, Boston, MA, October 1996.

....which may have different characteristics, the sender alone cannot simultaneously meet the conflicting demands of different receivers. One solution to this problem is proxy based transcoding, where data streams are individually transformed according to the specification of each requesting receiver [7]. Such a solution, however, typically imposes an administrative burden because it is not transparent to end users. Proxies are also difficult to deploy a user behind a constrained network link might not have access to the optimal point of placement for a proxy. Although the active networks ....

....for such services as part of the basic network service model [27] numerous research problems must be addressed before such an infrastructure is widely deployed. Furthermore, Fox et al. argue that transcoding proxies must be highly reliable and scalable, which comes at considerable effort and cost [7]. Alternatively, we can avoid proxies with an entirely endto end solution that completely avoids computation within the network through the use of layered media formats. A layered encoding algorithm encodes source data as a series of layers l 1 ; l 2 ; l n . The layers are structured ....

Fox, A., Gribble, S. D., Brewer, E. A., and Amir, E. Adapting to network and client variation via on-demand, dynamic distillation. In Proceedings of ASPLOS-VII (October 1996).

....Furthermore, because there are multiple paths, which may have different characteristics, the sender alone cannot simultaneously meet the conflicting demands of different receivers. One solution to this problem is the deployment of a proxy to transcode the data, customizing it for each receiver[6]. Such a solution, however, is not transparent to end users. Proxies are also difficult to deploy a user behind a constrained network link may not have access to a site on the other side of the link where a proxy should be situated. Although the active networks and active services research ....

....for such services as part of the basic network service model, numerous research problems must be addressed before such an infrastructure is widely deployed. Furthermore, Fox et al. argue that transcoding proxies must be highly reliable and scalable, which comes at considerable effort and cost[6]. Alternatively, we can avoid proxies with an entirely end to end solution that completely avoids computation within the network through the use of layered media formats. A layered encoding algorithm encodes source data as a series of layers l 1 ; l 2 ; l n . The layers are structured ....

Fox, A., Gribble, S. D., Brewer, E. A., and Amir, E. Adapting to network and client variation via on-demand, dynamic distillation. In Proceedings of ASPLOS-VII (October 1996).

.... based on document hyperlinks and user access patterns have been proposed in several studies [1, 8] Filtering in order to adapt to dynamic network quality of service has been proposed in related work, both in the context of WWW accesses and in the context of application adaptation in general [12, 25, 7]. Collaborative filtering has been proposed in the context of newsgroups [13] as well as WWW [11] Hoarding has been proposed in the context of file systems support during disconnected operation [18, 22] The contribution of this paper is the combination of several mechanisms and the use of ....

A. Fox, S. Gribble, E. Brewer, and E. Amir. Adapting to Network and Client Variation via On-Demand Dynamic Distillation, Proceedings of ASPLOS-VII, Boston, Massachussetts, October 1996.

....[8] These services are middleware because they run inside the network infrastructure, rather than at the endpoints of the network (either clients or servers) Middleware services thus can be regarded as an extra level of indirection between clients and servers. TranSend provides distillation [7,8] services for the U.C. Berkeley home IP [2] modem user population, representing roughly 10,000 active users of a bank of 600 700 modems. Distillation is data type specific, lossy compression for example, a distilled image may have reduced resolution or color depth, sacrificing image quality for ....

A. Fox, S. D. Gribble, E. Brewer and E. Amir. Adapting to Network and Client Variation Via On-Demand Dynamic Distillation. Proceedings of ASPLOS-VII, Boston, October 1996.

....[8] These services are middleware because they run inside the network infrastructure, rather than at the endpoints of the network (either clients or servers) Middleware services thus can be regarded as an extra level of indirection between clients and servers. TranSend provides distillation [7,8] services for the U.C. Berkeley home IP [2] modem user population, representing roughly 10,000 active users of a bank of 600 700 modems. Distillation is data type specific, lossy compression for example, a distilled image may have reduced resolution or color depth, sacrificing image quality for ....

A. Fox, S. D. Gribble, E. Brewer and E. Amir. Adapting to Network and Client Variation Via On-Demand Dynamic Distillation. Proceedings of ASPLOS-VII, Boston, October 1996.

No context found.

A. Fox, S. D. Gribble, E. Brewer and E. Amir. Adapting to Network and Client Variation Via On-Demand Dynamic Distillation. Proceedings of ASPLOS-VII, Boston, October 1996.

....that do not want to limit the effective audience of their content must therefore be able to deliver content that suits the needs of all of these diverse clients. Either the services themselves must adapt their content, or they must rely on the emergence of middleware services (such as in [13] [14], and [7] to adapt content on the fly to better suit the clients particular needs. 3.2 Client Activity As seen in figure 2, the amount of activity seen from the client population is strongly dependent on the time of day. The Berkeley web users were most active between 8:00pm and 2:00am, with ....

....nature of web service response times. In the next section, we discuss how these observations relate to a real Internet middleware service designed at Berkeley, the TranSend distillation proxy. 4 System Design Experience from TranSend The TranSend middleware service provides distillation ( 13] [14]) services for the Berkeley Home IP modem user population, representing roughly 8,000 time(ms) time(ms) a) b) c) 1 10 100 1000 10000 100000 0 20000 40000 60000 80000 100000 dur cs.hist 0 50000 100000 150000 200000 250000 0 5000 10000 15000 20000 dur ss.hist 1 10 100 1000 10000 100000 ....

Armando Fox, Steven D. Gribble, Eric A. Brewer, and Elan Amir. Adapting to network and client variation via on-demand dynamic distillation. In Fifth International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS-V), October 1997.

No context found.

A. Fox, S.D. Gribble, Y. Chawathe, and E.A. Brewer, "Adapting to network and client variation via On-Demand Distillation", in Proceedings of the 16th Symposium on Operating Systems Principles (SOSP-16), Oct. 1997.

No context found.

A. Fox, S. D. Gribble, E. A. Brewer, and E. Amir, \Adapting to Network and Client Variation via On-Demand Dynamic Distillation," in ASPLOS-VII, (Boston, MA), October 1996.

No context found.

Fox, A., S.D. Gribble, E.A. Brewer, and E. Amir. "Adapting to Network and Client Variation via On-Demand, Dynamic Distillation." Proc. ASPLOS-VII, Boston, Masachusetts, U.S.

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