US 6708213

US Patent 6,708,213: Method for Streaming Multimedia Information Over Public Networks

Title: Method for streaming multimedia information over public networks

Assignee: The patent's current assignees are Alcatel Lucent SAS and Sound View Innovations LLC. The original assignee was Lucent Technologies Inc.

Inventors: Ethendranath N. Bommaiah, Katherine H. Guo, Sanjoy Paul, and Markus A. Hofmann.

Filing Date: March 29, 2000

Issue Date: March 16, 2004 (This is the publication date, which is synonymous with the issue date for granted patents).

Abstract: The patent describes a method and apparatus designed to improve existing caching systems for streaming multimedia (SM) over public networks like the Internet. This is achieved through the use of "helper servers" (HSs) acting as caching and streaming agents within the network. These helpers employ proxy caching, client request aggregation (utilizing memory and disk resources at the helpers), and data transfer rate control to decrease start-up latency. The invention aims to reduce server and network loads and enhance the quality perceived by end-users by effectively managing variations in client request arrival times and requested content ranges for SM objects.

Plain-Language Overview of Independent Claims:

• Claim 1: This claim outlines a method to reduce delays (latency) when delivering streaming multimedia (SM) objects in a network. This network includes a central content server, multiple intermediary "helper servers" (HSs), and numerous client devices. The method involves:

  • Handling initial requests: When a client first requests an SM object and specifies a starting point, a dedicated temporary storage area called a "first ring buffer" is set up in the memory of one of the HSs. This buffer stores the part of the SM object starting from the client's requested position.
  • Continuous buffering: The "first ring buffer" continuously updates itself by replacing older stored data with newer, subsequent parts of the SM object, acting like a sliding window of data.
  • Handling subsequent requests: If another client requests the same SM object, the system first checks if the data needed for this new request is already available in the "first ring buffer." If it is, the request is served from there. If not, a "second ring buffer" is created to fulfill this new request.

• Claim 13: This claim also details a method for reducing latency when streaming SM objects through HSs to clients. It covers:

  • Receiving a request: A helper server receives a client's request for an SM object, including a desired starting position.
  • Buffer allocation: Immediately upon receiving this request, a "first ring buffer" is allocated in the HS's memory.
  • Data retrieval: The requested SM object, which consists of time-ordered packets, is fetched from either the specific HS itself (if available) or the central content server, or a combination of HSs.
  • Sequential storage: These time-ordered packets are then stored in the "first ring buffer," with newer packets continuously replacing older ones.
  • Servicing further requests (conditional): If a second request for the same SM object comes in, it can be served from the "first ring buffer" only if the requested starting position falls within the time range of the packets currently held in that buffer.
  • Allocating a new buffer (conditional): If the second request's starting position is outside the range of data in the "first ring buffer," a "second ring buffer" is allocated in memory to handle it.

• Claim 16: This claim describes a method to reduce latency in a network when streaming media (SM) objects from a content server, through helper servers (HSs), to clients. The method includes:

  • Request reception: A helper server receives a request for an SM object from a client.
  • Buffer allocation: A buffer is allocated at that HS specifically to cache (store temporarily) at least a portion of the requested SM object.
  • Simultaneous delivery and fetching: The cached portion of the SM object is downloaded to the requesting client at the same time (concurrently) that the remaining parts of the SM object are retrieved from either another HS or the original content server.
  • Rate adjustment: The rate at which data is transferred from the HS to the client is adjusted.

• Claim 17: This claim defines the components of an interconnected network of helper servers (HSs). Each HS in this network is equipped with:

  • RTSP request handling: Mechanisms to receive and process client requests using the Real-Time Streaming Protocol (RTSP).
  • Request forwarding: Mechanisms to send client requests onward to a content server.
  • RTP data streaming: Mechanisms to stream data to clients using the Real-Time Protocol (RTP).
  • Memory management: Mechanisms to manage available memory as a "buffer pool," where each buffer is linked to a specific SM object via a Uniform Resource Locator (URL).
  • URL to filename mapping: Mechanisms to translate URLs (identifying SM objects in client requests) into local filenames.
  • Disk space management: Mechanisms to manage allocated disk space for caching through a defined cache replacement policy (determining how SM object segments are added or deleted).
  • Cache data operations: Mechanisms for both recording data onto the cache and reading data from it.
  • Event scheduling: A scheduler that manages various events, including "data producer events" (which source parts of an SM object to be stored in the buffers), "data consumer events" (which are the client RTSP requests), and "garbage collector events" (which remove no-longer-needed portions of SM objects from the buffer pool to free up memory).

CAFC 2026 Dockets:

US Patent No. 6,708,213 was notably involved in a Federal Circuit case decided on January 29, 2026, titled Sound View Innovations, LLC v. Hulu, LLC (Appeal No. 2024-1092). The Federal Circuit affirmed a district court's summary judgment of non-infringement in favor of Hulu.

The core of the dispute revolved around the interpretation of method claim 16 of the '213 patent. The Federal Circuit ruled that, despite the absence of explicit sequential language, the first two steps of claim 16—"receiving a request for an SM object" and "allocating a buffer...to cache...said requested SM object"—were implicitly required to be performed in a specific order due to their logical and grammatical dependencies. The court found that the term "said requested SM object" in the second step logically presupposes that a request for the object (the first step) had already occurred.

Hulu's accused streaming system was found not to infringe because it allocated buffers before receiving a request for a specific streaming object, which contradicted the court's interpretation of the required order of steps in claim 16. While the Federal Circuit also found that the district court had erred in its construction of the term "buffer" (by requiring a specialized buffer), this error did not alter the final outcome, as the mandatory ordering of the method steps was dispositive to the non-infringement finding. The patent is reported as expired.