A. Bestavros. Demand-based document dissemination for the world-wide web. Technical Report 95-003, Boston University, Computer Science Department, Feb. 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... 2 Citations [AFA97] AFAW97] ANF97] ALF98] Abd98] ASA 95] AW96a] ASF98] ABCdO96] dAMdO 96] AAY97] AC98a] ACF 98a] ACF 98b] AW96b] AW97] AC98b] BMS96] BBM 97] BDR97] BC98] BO00] BS96] BLFF96] FGM 99] Bes95b] BCC 95a] Bes95a] BCC 95b] BC95] Bes95c] Bes96b] BC96] Bes96a] BVG96] BH96] BDH 94] Bra96] wBC95] BCFP98] BCF 99] BGM98] BD96] CDF 98] CFKL95] CI97] CZB98] CGL 98] CC96] Cat92] CP95] CDN 96] CY97] CMPS95] CLM99] CLLM00] Cor96] Cor98] CB95] CB96] CB97] ....

Azer Bestavros. Demand-based Document Dissemination for the World Wide Web. Technical Report TR-95-003, Boston University, CS Dept, Boston, MA, February 1995.

....of servers according to some local selection scheme (usually round robin) 1, 9] While the latter approach due to the single common entry point is neither scalable nor faulttolerant the former suffers from the need to regularly update aging statistics. We are aware of three methods [7, 8, 15] that like our method use client server cooperation to level loads on replicated servers. They all address two separate concerns [8] 1) the problem of disseminating data closer to clients, and (2) finding the nearest replica. These two concerns are tackled by a cooperative replication ....

....point is neither scalable nor faulttolerant the former suffers from the need to regularly update aging statistics. We are aware of three methods [7, 8, 15] that like our method use client server cooperation to level loads on replicated servers. They all address two separate concerns [8]: 1) the problem of disseminating data closer to clients, and (2) finding the nearest replica. These two concerns are tackled by a cooperative replication strategy based on client access statistics collected by the servers. When requesting a document, a client queries the primary server for this ....

[Article contains additional citation context not shown here]

A. Bestavros. Demand-based document dissemination for the world-wide web. Technical Report 95-003, Boston University, Computer Science Department, Feb. 1995.

....organizational boundaries. Their choice was motivated by their finding that cache performance could be improved by biasing the cache replacement policy in favor of more heavily shared local documents. Bestavros also explored a mechanism for distributing popular documents based on server knowledge [3]. There have also been a number of simulation studies of caching in large environments. Using trace driven simulations Alonso and Blaze showed that server load could be reduced by 60 90 [1, 2] Muntz and Honeyman showed that a caching hierarchy does not help for typical UNIX workloads [17] A few ....

Azer Bestavros. Demand-Based Document Dissemination for the World-Wide Web. Computer Science Department, Boston University, February 1995. Available from ftp://cs-ftp.bu.edu/techreports/95003 -web-server-dissemination.ps.Z.

....push caching as a method for reducing network traffic. This method involves replicating files. The degree to which a file is replicated is proportional to the file s global popularity, which can be determined from the global knowledge of the server. A similar approach is suggested by Bestavros [9]. One problem that arises with the use of network file caching and replication of files is the locating of the cached copy by the client, without communicating with 3 See http: java.sun.com for more information. 4 See http: vrml.wired.com for more information. 18 the server. Current ....

A. Bestavros, "Demand-Based Document Dissemination for the World Wide Web", Technical Report BU-CS-95-003, Boston University Computer Science Department, 1995. Available at URL: ftp://cs-ftp.bu.edu/techreports/ 95-003-web-server-dissemination.ps.Z

....systems community turned toward globally distributed systems such as the World Wide Web and FTP. Several groups are working on similar problems, but most are focused on client side caching. Only one group that we know of, at Boston University, is also working on server initiated caching. [2] One such client side system is included in The Harvest system [4] Their object caching subsystem provides a hierarchically organized means for efficiently retrieving Internet objects such as FTP and HTML files. The group at the University of Colorado at Boulder has also examined the resource ....

Azer Bestavros. Demand-based document dissemination for the world-wide web. Technical Report 95-003, Boston University, 1995.

....that a small percentage of the files available on a given server are responsible for a disproportionate share of the requests from that server. For example, the top 5 of the files on the NCSA server were responsible for 90 of the total requests from that server. Bestavros confirms these results [2], adding that the more globally popular a server, the smaller the fraction of pages that account for most of its accesses. Our results agree with this observation: the two most popular servers, NCSA and FAS, are also the two with the smallest percentages of files responsible for the most requests. ....

....time period were not included in these statistics. Notice that the most popular server, the FAS server, is also the one with the fewest mutable files. Bestavros provides further reasons why this random lifetime generation is flawed; he found that on a given server only a few files change rapidly [2]. Furthermore he notes that globally popular files are the least likely to change, evidence that Worrell s random file selection is flawed as well. If file request distribution is skewed toward the most popular files then the overall number of stale hits would go down even further. 6.3.1 Trace ....

Azer Bestavros. Demand-based document dissemination for the worldwide web. Technical Report 95-003, Boston University, 1995.

....Braun and Kim Claffy at the San Diego Supercomputer Center, to put Harvest caches into place at high level parts of the U.S. research and educational networking infrastructure, for the purpose of reducing backbone network traffic [8] Server based replication of objects is argued for in [2] and [6] Both papers argue that network traffic and server load could be reduced considerably by having servers disseminate their most popular documents on servers closer to clients. Both also point out the need for integration of producer based dissemination and consumer based caching of ....

A. Bestavros. Demand-based document dissemination for the World-Wide Web. Technical Report TR-95-003, Computer Science Department, Boston University, Feb. 1995. Available as ftp://csftp. bu.edu/techreports/95-003-web-server-dissemination.ps.Z.

....organizational boundaries. Their choice was motivated by their finding that cache performance could be improved by biasing the cache replacement policy in favor of more heavily shared local documents. Bestavros also explored a mechanism for distributing popular documents based on server knowledge [3]. There have also been a number of simulation studies of caching in large environments. Using trace driven simulations Alonso and Blaze showed that server load could be reduced by 60 90 [1, 2] Muntz and Honeyman showed that a caching hierarchy does not help for typical UNIX workloads [18] A few ....

Azer Bestavros. Demand-Based Document Dissemination for the World-Wide Web. Computer Science Department, Boston University, February 1995. Available from ftp://cs-ftp.bu.edu/techreports/95-003-webserver -dissemination.ps.Z.

....unicast connections. In order to calculate the achieved bandwidth reduction, we need to characterise the distribution of the popularity between the documents inside a Web server. The high popularity of a small number of Web documents is a well known phenomenon reported in several studies [24] [11], which have shown that Popular documents are very popular in the sense that some popular documents account for most of the requests on a Web server. Cunha, Bestavros, and Crovella [19] characterised the popularity of Web documents and confirmed the strong applicability of Zipf s law to Web ....

A. Bestavros, "Demand-based Document Dissemination for the World-Wide Web", 95-003, Boston University, Computer Science Department, February 1995.

....Another challenge in replication schemes is to locate the nearest server having the wanted object. If the server is chosen randomly, network traOEc is certainly not optimized. 6.2. 1 Demand based replication Bestavros has considered disseminating the most popular documents to dioeerent servers [Bestavros 95a] In his proposal documents would be automatically disseminated according to their popularity. In order to do so, each server must collect statistics on the popularity of each document it maintains. To make the dissemination more eoeective, servers would maintain a list, mapping local URLs to ....

Bestavros, A. 1995. Demand-based Document Dissemination for the World-Wide Web. Technical Report TR-95-003, Computer Science Department, Boston University, February 15, 1995. 22 pages.

....to these hints. Work along this line has also been developed elsewhere [17] This dissertation will follow a third approach which will disseminate file replicas to strategic points in the network in order to reduce server and network load. 8 Identification of which files to replicate is done in [14]. In this dissertation, position identification techniques are studied. As it will be shown, little gain is achieved depending on where the replicas are placed. This means that replica placement is a forecasting process. An aggressive guessing policy, i.e. putting copies too close to the most ....

....two from an academic and one from a commercial environment, searching for common characteristics that servers share. They came up with a list of 10 items that all servers analyzed presented. Among these items, it is interesting to note the confirmation of certain points claimed in other studies [32, 14, 30], for example, that small files are responsible for most of the transfers, and that few files are responsible for most of the bytes transfer and requests received. 2.6 Summary In this Chapter, the trace collection procedure at clients was presented. The modifications made to Mosaic and the types ....

[Article contains additional citation context not shown here]

Azer Bestavros. Demand-based Document Dissemination for the World-Wide Web. Technical Report TR-95-003, Computer Science Dept., Boston University, 111 Cummington St, Boston, MA 02215, February 1995.

....based on the logs of our departmental HTTP server http: cs www.bu.edu, we show that both server load and service time could be reduced considerably, if speculative service is used. This is above and beyond what is currently achievable using client side caching [3] and server side dissemination [2]. We identify a number of parameters that could be used to fine tune the level of speculation performed by the server. This work has been partially supported by NSF (grant CCR 9308344) 1 Introduction Current protocols for accessing distributed information systems do not scale, partly due to ....

....in on older traces, in favor of dependencies exhibited in more recent traces. This aging mechanism depends highly (among other things) on the frequency and pattern of document updates on the server. The relative stability of P and P observed in the above experiments reinforces our findings in [2] and the findings of Gwertzman in [9] that for WWW documents the popularity profile tends to be stable and updates tend to be infrequent. Ratio Level of Speculation more speculative less speculative T p 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 0.20 0.40 0.60 0.80 1.00 Server load ....

[Article contains additional citation context not shown here]

Azer Bestavros. Demand-based document dissemination for the world wide web. Technical Report TR-95-003, Boston University, CS Dept, Boston, MA 02215, February 1995. (submitted for publication).

....information systems, such at the WWW, is based on the popularity based dissemination of information from servers to proxies, which are closer to clients. There are three problems to be tackled for such an approach, namely what, how far, and in which direction(s) to disseminate. The work in [2] addresses the first two aspects, whereas the work in [8] investigates the third. Supply based dissemination of information is complementary to demand driven caching; the former aims primarily at reducing traffic and balancing load (through replication) whereas the latter aims primarily at ....

Azer Bestavros. Demand-based document dissemination for the World Wide Web. Technical Report TR-95-003, Boston U., CS Dept, Boston, MA 02215, Feb. 1995.

No context found.

A. Bestavros. Demand-based document dissemination for the world-wide web. Technical Report 95-003, Boston University, Computer Science Department, Feb. 1995.

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