V. Gupta and R. H. Campbell. Internet search engine freshness by web server help. In Proceedings of the Symposium on Internet Applications (SAINT), pages 113--119, 2001.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....fingerprints to compress and send only the modified parts, and CTM [13] that is a method for sending differences via e mail, used to keep up to date copies of the source code of the Open BSD operating system. A specific proposal for pushing last modification data to web crawlers is presented in [12], including a cost model in which the meta data is sent only if the web server is expected to be misrepresented above a certain threshold in the search engine. A more general Internet notification system was presented in [7] The Distribution and Replication Protocol (DRP) 28] provides a ....

....page) but the web server can wait and send the updates in batches. When receiving the meta data about the update, the search engine may download, enqueue or ignore the event. Examples: the Keryx notification service, developed in the apogee of push based content delivery in 1997 [7] Fresh Flow [12]. Send differences of content: Whenever an event happens, the web server sends a file con taining the difference between the last version and the current one (if the change is big, then the web server may send the entire file) The search engine may accept or ignore the file received, but if it ....

V. Gupta and R. H. Campbell. Internet search engine freshness by web server help. In Proceedings of the Symposium on Internet Applications (SAINT), pages 113-119, 2001.

....Crawlers receive web object meta data (URL, last modified date, file size, file checksum, etc. which can be used to decide if their local copy is stale. This idea is carried still further by Gupta and Campbell, by allowing the local web server to measure the popularity of its own object repository[20]. Web crawlers can then use the popularity and frequency of change data measured at the web server to schedule e#cient crawls. Furthermore, the data is pushed to subscribed search engines rather than remaining passive on the local web server. Ho s own protocol for web event dissemination among a ....

....which are beyond the scope of this work. Providing web event dissemination services with other sources (web servers, proxy servers, etc. could be realized through development of an Apache server module. This could be used to provide the services described by Brandman et al. [5]and Gupta et al. [20]. http: www.apache.org ....

Vijay Gupta and Roy H. Campbell. Internet Search Engine Freshness by Web Server Help. In Symposium on Applications and the Internet, pages 113--119, 2001.

....answer queries is not usually directly visible to the outside world. Although the aggregate amount of new information created is not very large, there are a large number of smallsized updates. A separate multicast of every small update would pose a significant control overhead on the network. In [13], we discuss an algorithm to propagate web updates such that the freshness of data is preserved and at the same time, the number of transfers needed to propagate the updates is reduced. There the particular application being considered is the update of search engines so that the results of queries ....

....To summarize the motivation, we are targeting the use of reliable multicast for improved QoS for the following applications: 1. Search engines: By considering freshness to be a QoS metric, we show how multicasting the web updates can be useful for improving the freshness of search engines [13]. 2. Audio and video: If audio and video are distributed to locations which are closer to the consumers, then facilitating improved QoS becomes much easier. As such, we consider audio and video to be loss intolerant data which ought to be transmitted later in a real time fashion. C. High level ....

V. Gupta and R. Campbell. Internet search engine freshness by web server help. Technical Report UIUCDCS-R-2000-2153, Digital Computer Lab., U. of Illinois, January 2000.

....an evolving area and the focus of these early studies has been on the pull of updates. There is an alternative way of quickly propagating the updates, which is known as push approach [GS96] Our prior work has developed algorithms to support push of updates for facilitating web cache consistency [GC00] While that report is more concerned with the architectural details, this paper presents the analysis. Since disconnected operation is frequent in mobile computing, updates which occur on the laptop computer cannot be easily polled by any other computer. As such, push of updates from the laptop ....

....on the replica at the backbone server when modi cations are made to the laptop replica. The reason for introducing OC are: If a user is modifying les in her public html directory, then it is desirable to propagate the updates to the backbone server for freshness considerations [CGM00, GC00] 3 In mobile computing, there is a notion called queued RPC [JdT 95] Because of queued RPC, operations performed on replicated databases are not immediately transmitted. Similarly, if a user is sending Emails, the Emails won t be immediately sent. It is not inconceivable that excessive ....

[Article contains additional citation context not shown here]

V. Gupta and R. Campbell. Internet search engine freshness by web server help. Technical Report UIUCDCS-R-2000-2153, Digital Computer Laboratory, University of Illinois at UrbanaChampaign, January 2000.

No context found.

V. Gupta and R. H. Campbell. Internet search engine freshness by web server help. In Proceedings of the Symposium on Internet Applications (SAINT), pages 113--119, 2001.

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