Russell J. Clark and Mostafa H. Ammar. Providing Scalable Web Services Using Multicast Communication. Computer Networks and ISDN Systems, 29(7):841--858, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the famous Starr report case) There are several methods to try to overcome the problem. One is to use caches [13, 7, 1] However, caches are not e ective for frequently changing content or for long les (e. g video, audio) A di erent possibility that we consider in this paper is to use multicast [4, 8, 6], i.e. to deliver the content simultaneously to many (all) users via multicast dynamic tree. Obviously, one may also combine both caching and multicasting to further improve the solution. At rst, it may seem that multicast could be e ective only if many users requests exactly the same content ....

R. J. Clark and M. H. Ammar. Providing scalable Web services using multicast communication. Computer Networks and ISDN Systems, 29(7):841-858, 1997.

....top of a multicomputer machine designed and implemented in our laboratory.Themachine uses a multistage switching network to connect multiple clusters of multiprocessors. Each processor This work was partially supported by the National Science Council under grants NSC84 2221 E 002 004, and NSC85 2221E 002 029. can either be a simple CPU module, or with individual storage devices and or network adapters attached depending on the needs of applications. As a result, multiple I O devices can be accessed concurrently to provide higher I O bandwidth than single bus machines. An ATM like ....

....and the availability issues first. Some related works are also discussed. Then, we present our multicomputer platform in Section 3. In Section 4, a general software structure of a web server proposed for multicomputers like ours is depicted. An implementation of the server is discussed in Section 5. Section6 draws conclusions. 2 Issues of WWW Servers In World Wide Web, user agents use the application level stateless Hypertext Transfer Protocol (HTTP) 3] to request documents or other kinds of objects from web servers (or from proxies gateways to other Internet servers) In practice, a ....

[Article contains additional citation context not shown here]

R.J. Clark, and M.H. Ammar, "Providing Scalable Web Service Using Multicast Delivery," Second International Workshop on Services in Distributed and NetworkedEnvironments,1995.

....an application to maximize the impact of an action by aggregating dependencies. In TCP, the Nagel algorithm [Nag84] delays transmission of small packets awaiting more data. In the web, there have been several attempts to collect requests for popular pages and multicast the response to all [CA95, Ham95] This minimizes the number of times that the server must transmit the page and can reduce the number of copies sent into the network, but it may cause users to wait longer for pages. All prefetching works discussed above, whether it occurs at the client, the server, or an ....

Clark, R. J. and Ammar, M. H., "Providing Scalable Web Service Using Multicast Delivery", Technical Report GIT-CC-95-03, College of Computing, Georgia Institute of Technology, 1995.

....throughput performance. After the rst installation of the MBONE [Dee88] a few years ago, a growing part of the internet routers supports IP multicast. With the availability of the MBONE reliable multicast protocols have been implemented for the WAN environment and results have been reported in [RA95] FJM 95] YGS95] The WAN environment provides a new challenge: reliable multicast communication had to face the scalability problem that arises for a large number of receivers. The research community proposed two directions to deal with the scalability problem: ffl timer approaches ....

Clark Russell and Mostafa Ammar. Providing scalable web service using multicast delivery. In Proceedings of the IEEE Workshop on Services in Distributed and Networked Environments, Whistler, Canada, June 1995.

....deals with the efficient delivery of information to multiple users in overly crowded internet environment. Optimization of data flow between users and servers, can be achieved by intelligent caching, and multicasting schemes. The most recent work in this field has been done by Ammar et al. [6], 7] On top of a reliable multicast data transfer protocol, they developed an application level data delivery policy, which partitions the web pages at the server side into hot, warm, and cold pages according to their popularity. In their proposed architecture, user requests reach the server ....

....architecture, user requests reach the server via an HTTP connection. If the request is for a cold page, then it is delivered via the existing HTTP connection. If it is for a warm page, then all of the requests for this page are grouped together, and serviced by a single multicast connection [6]. The hot pages are multicast at a known multicast address periodically, and continuously, as soon as a request for them is received by the server [7] The work described above, only dealt with the effective and efficient delivery of the information. Caching at both intermediate nodes and ....

Russell Clark and Ammar, Mostafa. " Providing Scalable Web Service Using Multicast Delivery ", Proceedings of the IEEE Workshop on Services in Distributed and Networked Environments, June 1995.

....Our approach is scalable and enables sites to cut down costs, and still be able to handle increased demand. Users benefit from better response time due to lower load on the server, and less congested core links. An approach to use multicast in the delivery of web resources was first introduced in [10]. In their architecture, clients use http requests. Several simultaneous requests are assembled, and grouped into a multicast group. Then, the clients join the dynamically generated multicast group and receive the page using a reliable multicast protocol. A later work [11] suggested the use of ....

R. Clark and M. Ammar, "Providing scalable web services using multicast communication," Proceedings of the IEEE Workshopon Services in Distributed and Networked Environments, Whistler, Canada, Feb. 1995.

....of redundant contents in their corresponding update logs[20] ffl Reduced transmission cost Individual datagroups can be designed so that increasing numbers of clients become interested in them. Their resultant update logs become good candidates for broadcast or multicast transmission[1, 3, 5, 18]. For brevity, we only consider unicast transmission in this paper. Our goal is to increase scalability in server update processing and thereby reduce the overall cost (e.g. time) to synchronize disconnected clients. In this paper, we model update processing costs for ISDBs, propose a method ....

....these patterns to reduce update log dissemination costs. Effective use of bandwidth becomes an issue as the volume of update data increases. Grouping data can have the side effect that many clients share interests in common update objects, and multicasting or broadcasting them becomes more feasible[2, 5]. 5 Conclusions ISDB systems on the market today handle update dissemination in a client centric manner, by generating update logs on a client by client basis. Unfortunately, the computational complexity of the system increases directly with the number of clients. Previous research proposes an ....

R. J. Clark and M. H. Ammar. Providing scalable web service using multicast communication. Computer Networks and ISDN Systems, 29:841--858, 1997.

....to other materialized views. This work does not 77 address the issue of multiple differential update files, as required by our approach. Related work in aggregation of data for broadcast has been performed in many areas. This approach has been used to reduce bandwidth consumption of web servers[13, 42]. In [165] the authors analyzed data aggregation techniques for broadcasting data in Videotex systems. Our approach of dissemination of updates from server to the clients is related to work in broadcast disks[1, 2, 3, 4, 5, 6, 167] The motivation for this related work is to operate in an ....

....with multicast. One example is in the area of streaming video across the Internet. Researchers have adapted traditional unicast video server technology to multicast by developing mechanisms to maintain bandwidth fairness[38, 102, 101, 112] and even provide VCR style functionality[9] Web servers[42] have also been converted to multicast their high volume pages. Work has also been done in multicast transport layer protocols[15, 37] to allow better flow control for all receivers. 78 3.3 Data Centric Aggregation Approaches We must determine the best method of deriving the data groups (called ....

Russell J. Clark and Mostafa H. Ammar. Providing scalable Web services using multicast. In Proceedings of the 2nd International Workshop on Services in Distributed and Networked Environments (SDNE 95), June 1995. 257

....to transmit the reply using a multicast connection to the clients. The multicast transmission allows the server to satisfy many requests with a single transmission, and hence decreases the server load. Furthermore, it reduces the overall bandwidth required to transmit the document to the clients [7]. The other way to control the amount of Web data in the network is to minimize the number of client requests sent to distant servers. This is typically done with caching which we examine in next section. 2. WEB CACHING Caching and related techniques attempt to prevent network congestion by ....

R. J. Clark, M. Ammar, Providing scalable Web services using multicast communication, to appear in Proc. IEEE Infocom'97, Kobe, Japan, April 1997

....file from the origin server. In fact our scheme is not just limited to the origin server but could also be implemented at cache proxy servers situated between the origin server and the clients. The use of multicast for distribution of web documents has been proposed elsewhere in the literature. [4] presents a technique whereby outstanding requests at the server are queued in a central queue prior to commencement of service. Associated with each such request is a list of clients awaiting delivery of the requested file. Files at the server are labelled as either hot(popular) or cold(not ....

....the new requesting client and no new entry in the central queue results. When each request reaches the head of the central queue the request is scheduled for multicast delivery if it has more than one client in its associated client list. Otherwise the request is scheduled for unicast delivery. [4] suggests that a portion of the multicast address space could be reserved for WWW multicast and each server could be allocated a contiguous block of addresses within this WWW multicast address space. 4] presents a many to one mapping of a server s files to the individual addresses within the ....

[Article contains additional citation context not shown here]

R. Clark, M. Ammar. Providing Scalable Web Service Using Multicast Delivery, Proceedings of 2nd International Workshop on Services in Distributed and Networked Environments (SDNE 95), June, 1995.

....we view CMP as one of the three complementary delivery options integrated on the World Wide Web: Caching, AMP, and CMP: ffl Caching: Popular documents that do not rapidly change and documents that are rarely requested are distributed via a caching hierarchy. ffl Asynchronous Multicast Push (AMP) [6] [16] This method can also be used for popular documents that do not frequently change. Requests for the same document are accumulated over a time interval, and answered together via multicast. The drawback of this method is that the grouping time increases the latency to the receivers. However, ....

R. Clark and M. Ammar, "Providing Scalable Web Service Using Multicast Delivery", In Proceedings of the IEEE Workshop on Services in Distributed and Networked Environments, Whistler, Canada, June 1995.

....realize how much of a traffic load is generated even by one audio and one visual feed. 6.3 Scalability There will also be issues as the MBONE grows. There has been an effort at the Georgia Institute of Technology to use multicast to relieve the load on World Wide Web servers for popular sites [4], and this raises the administrative issue of assigning multicast group addresses. While these are currently assigned by collaborative agreement, a more structured approach will be necessary if the popularity of multicast grows. It is also not clear whether the protocols presented in this paper ....

Russell Clark and Mostafa Ammar. Providing Scalable Web Service Using Multicast Delivery. Technical Report GIT-CC-95-03, College of Computing, Georgia Institute of Technology, Atlanta, GA, January 1995.

....of a caching scheme. If the number of requests in a clientinitiated system for a certain document is very high, the final server could try to group several requests in a certain time interval and answer the clients concerned to that period of time by multicasting them the same information [7]. Clients have to wait until a group is formed before they get the data. Only if the number of receivers is high, the bandwidth gain will be appreciable. This requires to wait for long intervals to have enough receivers. In the case of a server initiated scheme, the multicast operation could be ....

C. Russell and M. Ammar, "Providing scalable web service using multicast delivery," in Proceedings of the IEEE Workshop on Services in Distributed and Networked Environments, (Whistler, Canada), June 1995.

....request due to server overload, and poor performance observed by other network connections due to hot spots. To help maintain performance under higher demands, there have been several proposals for multicast service in which a server delivers information to multiple clients simultaneously (e.g. [4], 5] 9] Through the aggregation of request processing at the server and the aggrega This work is supported in part by research grants from Sprint, and NSF under contract number NCR 9628379 and ANI 9973115. tion of bandwidth requirements in the network, a multicast service can reduce the ....

Russell Clark and Mostafa Ammar. Providing scalable web service using multicast delivery. Computer Networks and ISDN Systems Journal, 29:841--858, 1997.

....(e.g. response time, throughput, reliability) Several approaches have been proposed for improving the scalability of a networked service. These include server replication [15] caching [16, 35] batching of requests at the server [14] and multicasting of server responses over the network [19, 10]. In this paper, we consider the server replication technique and focus on the design and evaluation of a system that supports client access to the best server for a broad range of definitions of best . 1.1 Server Selection Techniques When replication is used, one primary concern is how a ....

R. Clark and M. Ammar. Providing scalable web service using multicast delivery. In Proceedings of 2nd IEEE Workshop on Services in Distributed and Networked Environments, pages 19--26, 1995.

....servers is done with some goal in mind, e.g. minimizing client response time or balancing server load. Recently, there have been several proposals for multicast services in which a server delivers information to multiple clients simultaneously. Such proposals include multicasting of web content [7 11], multicast based video services (on demand and pay per view style services) 1216 ] multicasting of database content [17] and broadcast disks [18, 19] The goal of many of these proposals is to use multicast to enhance the ability of the service to handle a large number of clients economically. ....

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," Computer Networks and ISDN Systems Journal, vol. 29, pp. 841--858, 1997.

....with layered encoding or multi stream schemes[10, 11, 12, 13] Recent work has also looked at integrating the services of the MBone and the Real Time Protocol (RTP) into the the WWW. Several researchers are looking at various ways of using the MBone and multicast protocols to deliver WWW pages[14, 15, 16, 17]. Other issues are based on the integration of WWW and MBone style conferencing[18, 19] This paper is organized as follows. Section 2 describes our proposed jukebox paradigm and related paradigms. Section 3 details our implementation of a jukebox prototype called the Interactive Multimedia ....

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," in IEEE Workshop on Services in Distributed and Networked Environments, (Whistler, CANADA), June 1995. (to appear in Computer Networks and ISDN Systems).

....interactive conferencing and program broadcasts[6, 7, 8, 9] Finally, recent work has looked at integrating the services of the MBone and the Real Time Protocol (RTP) into the the WWW. Several researchers are looking at various ways of using the MBone and multicast protocols to deliver WWW pages[10, 11, 12, 13]. Other issues are based on the integration of WWW and MBone style conferencing[14, 15] This paper is organized as follows. Section 2 describes our proposed jukebox paradigm and other related paradigms. Section 3 details our implementation of a jukebox prototype called the Interactive Multimedia ....

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," in IEEE Workshop on Services in Distributed and Networked Environments, (Whistler, CANADA), June 1995. (to appear in Computer Networks and ISDN Systems). 17

....like conferencing and conference broadcasts[6, 7, 8] Finally, recent work has looked at integrating the services of the MBone and the Real Time Protocol (RTP) into the the WWW. Several researchers are looking at various ways of using the MBone and multicast protocols to deliver WWW pages[9, 10, 11, 12]. Other issues are based on the integration of WWW and MBone style conferencing[13, 14] This paper is organized as follows. Section 2 describes our proposed jukebox paradigm and other related paradigms. Section 3 details our implementation of a jukebox prototype called the Interactive Multimedia ....

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," in IEEE Workshop on Services in Distributed and Networked Environments, (Whistler, Canada), June 1995. (to appear in Computer Networks and ISDN Systems).

....often referred to as the scalability of the service. There have been several approaches proposed for improving the scalability of a networked service. These include server replication [1] caching [2, 3] batching of requests at the server [4] and multicasting of server responses over the network [5, 6]. We investigate the anycasting communication paradigm which has been proposed [7] to support server replication. As originally defined [7] anycasting provides: a stateless best effort delivery of an anycast datagram to at least one host, and prefer 1 ftp: ftp.cc.gatech.edu pub coc tech ....

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," in Proceedings of 2nd IEEE Workshop on Services in Distributed and Networked Environments, pp. 19--26, 1995.

....proposals while improving on their performance when other measures of interest are considered. Keywords Multicast, Receiver Heterogeneity, Scheduling I. INTRODUCTION A key technique for allowing the server to handle a large volume of requests is to multicast the data to the set of clients [20] [7], 1] 11] 2] Multicast in this context can be difficult. First, the clients will typically make requests over time, without synchronization. While the clients may be able to tolerate a brief delay in response allowing the server to batch more requests together this technique will be ....

Russell J. Clark and Mostafa H. Ammar. Providing scalable Web services using multicast. In Proceedings of the 2nd International Workshop on Services in Distributed and Networked Environments (SDNE 95), June 1995.

No context found.

Russell J. Clark and Mostafa H. Ammar. Providing Scalable Web Services Using Multicast Communication. Computer Networks and ISDN Systems, 29(7):841--858, 1997.

No context found.

M.H. Ammar R.J. Clark. Providing scalable web service using multicast delivery. College of computing technical reports, Georgia Tech, 1995.

No context found.

R. Clark and M. Ammar. Providing scalable web services using multicast communication. Proceedings of the IEEE Workshop on Services in Distributed and Networked Environments, Whistler, Canada, Feb. 1995.

No context found.

R. Clark and M. Ammar, "Providing scalable web service using multicast delivery," in IEEE Workshop on Services in Distributed and Networked Environments, (Whistler, Canada), June 1995. (to appear in Computer Networks and ISDN Systems).

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