Electrical computers and digital processing systems: memory – Storage accessing and control – Hierarchical memories
Reexamination Certificate
1999-07-19
2002-10-08
Peikari, B. James (Department: 2186)
Electrical computers and digital processing systems: memory
Storage accessing and control
Hierarchical memories
C711S130000, C711S159000
Reexamination Certificate
active
06463508
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to systems and methods for delivering video or audio streams over the World Wide Web and, particularly, to a system and method of caching media streams for delivering multimedia (video/audio) objects through proxy servers.
2. Discussion of the Prior Art
The recent increase in popularity of the World Wide Web (WWW or web) has led to a considerable increase in the amount of traffic over the Internet. As a result, the web has now become one of the primary bottlenecks of network performance. When web objects are requested by a user who is connected to a server via a slow network link there can be noticeable latency at the user end. Further, transferring web objects over the network leads to an increase in the level of traffic over the network. This reduces the bandwidth available for other requests, especially when large media objects are transferred. In order to facilitate the deliver of Web objects to users, it is desirable to install proxy servers which can cache web objects to facilitate the content distribution and reduce access delays.
Previous work on caching of web objects mainly relates to on-demand caching of the whole web object, i.e., a caching decision is made after the object is requested as to whether to cache the object. See, for example, the references entitled “Caching Proxies: Limitations and Potentials”, Proc. 4th International World Wide Web Conference, 1996, by Abrams et. el., and “Caching on the World Wide Web”, IEEE Trans. on Knowledge and Data Engineering, vol. 11, No. 1, January 1999, pp. 94-107, by C. Aggarwal, et al., for different web caching polices. In commonly owned, co-pending U.S. patent application Ser. No. 08/831,237, entitled “Collaborative Caching”, now U.S. Pat. No. 5,924,116, a collaborating caching method among the proxies is considered. In commonly owned U.S. patent application Ser. No. 08/854,226, entitled “Dynamic Push Filtering with Staging/Buffering in a Proxy Hierarchy”, now U.S. Pat. No. 6,065,058 a push filtering method is proposed in which meta information is used to facilitate the push filtering and staging or caching decision.
In commonly owned U.S. Pat. No. 5,878,223, entitled “System and Method for Predictive Caching of Information Pages”, the server determines and sends the pages predicted to be requested next to the requesting computer without a specific request by the user. This is basically a sender initiated approach. In commonly owned U.S. patent application Ser. No. 08/939,277, entitled “Method for Dynamically Prefetching Information via a Server”, now U.S. Pat. No. 6,085,193, a prefetching method is provided which can incorporate user specific information dynamically into the object selections. This provides an improved method for prefetching in a proxy hierarchy in order to reduce object access through the network (internet) by analyzing and identifying the common user reference patterns at the content server and proxy sites and providing Prefetch Hint Information (PHI) on related accesses via meta information piggybacked with the requested object. The PHI gets updated as the object passes through the proxy hierarchy to reflect prefetch operations performed and caching status at the higher levels of the hierarchy, and other considerations such as local reference patterns.
As requests for and delivery of streaming video and audio over the web becomes more popular, proper caching of media objects becomes increasingly important. In addition to the large bandwidth requirement, there is an isochronism requirement for the media stream. Caching can certainly help addressing these issues. However, for a large media file, such as a 2-hour video, treating the whole video as a single web object to be cached is impractical. Even a simple division of the media file into smaller web objects will not be efficient for caching purposes.
Recently, there have been developed several companies that provide multimedia content distribution services using a number of caching proxies distributed over different geographic locations. These companies include Akamai (www.akamia.com) and Sandpiper (www.sandpiper.net). Sandpiper uses the Inktomi Traffic Server (www.inktomi.com) in which the proxy server will get the video file from the content server upon the first client request and cache the whole video file so that subsequent requests to the same video can be served from the proxy cache.
It would be highly desirable to provide a system and method that enables more intelligent partial caching of media streams at proxy servers.
SUMMARY OF THE INVENTION
The present invention pertains to an improved method and apparatus for caching media streams at one or more proxy servers. In the current invention, blocks received for a media stream are grouped into segments by the proxy servers, where the cache admission and replacement policies attach a different caching value to different segments, taking into account the segment distance to the start of the media object. These policies may give preferential treatment to the beginning segments and may take into account other considerations such as the media object reference frequency, time since the last reference to the object, or its access time.
The present invention enables the proxy server to cache a subset of the segments, which will generally be the beginning segments.
In another aspect of the present invention, if only a portion of a media stream is cached in the proxy server when the stream is requested, a method is provided to prefetch the remaining segments. Upon a media request, the proxy can immediate serve the request using the segments cached, and compose and issue a prefetch request to obtain the remaining blocks for segments which are not currently cached.
To improve the manageability (or, more specifically, tracking overhead) of the segments and caching efficiency, a method is provided to divide the media object into unequal size segments, where a segment further away from the start will be assigned a larger number of blocks.
Advantageously, the approach of the invention is readily generalized to the case of a caching system implementing multiple proxies.
REFERENCES:
patent: 5467460 (1995-11-01), Patel
patent: 5878223 (1999-03-01), Becker et al.
patent: 5924116 (1999-07-01), Aggarwal et al.
patent: 6065058 (2000-05-01), Hailpern et al.
patent: 6085193 (2000-07-01), Malkin et al.
patent: 6243791 (2001-06-01), Vondran, Jr.
Wolf Joel L.
Yu Philip Shi-lung
International Business Machines - Corporation
Peikari B. James
Scully Scott Murphy & Presser
Zarick Gail H.
LandOfFree
Method and apparatus for caching a media stream does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for caching a media stream, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for caching a media stream will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2991165