SIP Packet Processing via Dedicated Network Co-Processor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The processing of Session Initiation Protocol (SIP) request messages incurs significant overhead due to the need for packets to traverse multiple servers, as SIP addresses are processed through the IP stack to the application layer, which is inefficient, especially in scenarios like contact centers where call redirection is common.
Innovation Solution
A networked server with a dedicated packet handling circuit, utilizing a network processor with co-processor or microengine circuitry, parses and modifies SIP packets without referring them to the application layer, identifying and redirecting SIP destination URIs directly at the network and transport layers, thereby reducing processing overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SIP request messages are processed through the IP stack to the application layer, then SIP addresses can be correctly identified and routed, but processing overhead increases significantly
Solution Approach 1:
The patent segments the packet processing function by separating SIP header parsing from full IP stack processing. The network processor extracts and parses SIP headers (To:, From:, Request-URI fields) independently at the network layer, identifying destination addresses without requiring complete application layer processing. This segmentation allows accurate SIP address identification while bypassing the overhead of full stack processing for each packet.
Solution Approach 2:
The patent introduces an intermediary mechanism - a dedicated network processor that acts as a mediator between the IP stack and application layer. This intermediary parses SIP messages and extracts routing information (destination IP addresses) at the network layer, then forwards packets directly to the destination without requiring application layer intervention. This intermediary approach maintains routing accuracy while dramatically reducing processing overhead.
2Adaptability or versatility
If SIP requests traverse multiple proxy and redirect servers, then routing flexibility and contact center offloading are enabled, but processing time and network latency increase
Solution Approach 1:
The patent applies preliminary action by having the network processor pre-parse SIP message headers and extract destination routing information before full packet processing. By identifying the final destination address in advance through SIP header analysis (To:, Request-URI fields), the system can optimize the packet forwarding path and reduce the number of intermediate hops, thereby maintaining routing flexibility while reducing overall session establishment time.
Solution Approach 2:
The patent implements skipping by enabling the network processor to rush through the packet processing pipeline using dedicated hardware acceleration for SIP header parsing and IP address extraction. The network processor can quickly identify and forward packets to their destinations without requiring sequential processing through multiple software-based proxy and redirect servers, thereby reducing latency while preserving routing options through intelligent packet inspection.
3Adaptability or versatility
If contact centers transfer SIP-based contacts to other network locations, then resource utilization across multiple contact centers improves, but processing overhead at each server increases
Solution Approach 1:
The patent replaces the mechanical software-based SIP processing system with a dedicated network processor that uses hardware-accelerated packet parsing. This substitution enables contact centers to efficiently transfer SIP-based contacts to other network locations by rapidly inspecting and forwarding packets based on SIP header information, without requiring complex software processing at each intermediate server. This reduces server processing complexity while maintaining contact center interoperability.
Data Source
AI summary
SIP (session initiation protocol) requests are processed by a networked server having a processor which has a dedicated packet handling circuit (or network-aware co-processor), without passing said requests through an IP stack to an application layer. The co-processor receives SIP requests in the form of IP packets. The co-processor then parses the packets, and thereby identifies a SIP destination address in the datagram of the packet. A memory area accessible by the co-processor has a stored list of SIP addresses and of associated forwarding addresses to which packets containing the listed SIP destination addresses are to be directed. The co-processor directly modifies the packets to include the stored further address, which may be an IP address or a SIP address, associated with the original destination SIP address and then forwards the packets via the network.