Patent No. US7246173 (titled "Method and apparatus for classifying IP data") on Apr 16, 2001. The application was issued on Jul 17, 2007.
’173 is related to the field of packet-switched networking and specifically addresses the challenges of maintaining Quality of Service (QoS) in environments utilizing source routing. In standard IP networking, routers typically prioritize or manage traffic based on the destination address found in the primary IP header. However, when source routing is employed—allowing a sender to specify a path through specific intermediate nodes—the destination field in the IP header is overwritten with the address of the next hop rather than the ultimate recipient. This creates a visibility gap for resource reservation protocols like RSVP, which may fail to apply the correct traffic policies because they cannot identify the true end-to-end session.
The underlying idea behind ’173 is to shift the classification logic from the transient destination address in the IP header to the permanent final destination address embedded within the source routing extension headers. By forcing the router to look deeper into the packet structure—specifically targeting the tail end of the routing list—the system ensures that the packet is mapped to its reserved resources regardless of how many intermediate hops it must traverse. This insight allows the network to maintain a consistent traffic control state for a specific flow, even as the visible destination address in the outer header changes at every hop along the source-defined path.
The claims of ’173 focus on a method and apparatus for classifying IP data by extracting routing information from specific sub-fields within the packet header. The independent claims describe receiving data that includes a list of intermediate nodes and performing classification based on an entry within that header rather than the standard destination field. This mechanism is specifically applied to both IPv4 and IPv6 architectures, targeting the Loose/Strict Source and Record Route (LSRR/SSRR) options in the former and the Type 0 routing header in the latter. The scope covers the logic required to identify when source routing is active and redirect the classifier to the final address in the list.
In practice, the invention functions as a diagnostic check performed by the router's processor during the initial packet inspection phase. If the router detects that the source routing has not yet been fully consumed—indicated by a pointer in IPv4 or a non-zero segments left field in IPv6—it bypasses the standard destination field. Instead, it calculates the offset for the very last address in the routing data field. This final address is then used as the key for the RSVP session lookup, ensuring the packet receives the specific bandwidth or latency guarantees previously negotiated between the source and the ultimate destination.
This approach differs from prior solutions that often failed to provide QoS for source-routed packets until they reached the final leg of their journey. By integrating the final destination address into the classification flow at every intermediate node, the invention prevents the 'misclassification' that occurs when a router mistakes an intermediate hop for the session endpoint. Furthermore, the system extends this logic to PATH refresh messages, allowing intermediate routers to store and propagate source routing states, thereby maintaining stable resource reservations across the entire network fabric.
In the early 2000s when ’173 was filed, network architectures were transitioning between established IPv4 frameworks and the emerging IPv6 standard at a time when Quality of Service (QoS) was typically implemented using Resource Reservation Protocol (RSVP) based on static destination addresses. During this era, systems commonly relied on standard packet headers for traffic classification rather than inspecting the internal components of routing headers, and hardware constraints made the deep inspection of source-routed paths non-trivial for real-time resource allocation. Consequently, network nodes often lacked the mechanism to reconcile explicit source-defined paths with the flow-based classification required for guaranteed service levels.
The disclosed invention represents a technical advancement by integrating source routing information directly into the packet classification process for QoS resource reservation. By configuring network nodes to classify data based on specific fields within an IPv6 routing header—such as the last destination address—or IPv4 source routing options like LSRR/SSRR, the architecture enables precise resource allocation along an explicitly defined path. This shift overcomes the technical constraint where intermediate nodes were previously unable to associate source-routed packets with their ultimate destination for the purpose of bandwidth or priority reservation, thereby ensuring consistent service quality across complex, multi-node network topologies.
This patent contains 37 claims, with claims 1, 5, 13, 17, 25, and 30 serving as the independent claims. These independent claims focus on methods and router configurations for classifying Internet Protocol data at a network node by utilizing routing information, such as a list of intermediate nodes or specific source routing entries like LSRR and SSRR, contained within the data header. The dependent claims provide additional details regarding specific protocol versions like IPv4 and IPv6, the use of destination address fields for classification, resource reservation procedures, and the storage or forwarding of source routing information.
Definitions of key terms used in the patent claims.
US Latest litigation cases involving this patent.

The dossier documents provide a comprehensive record of the patent's prosecution history - including filings, correspondence, and decisions made by patent offices - and are crucial for understanding the patent's legal journey and any challenges it may have faced during examination.
Get instant alerts for new documents