Encapsulation Header Packet Tracking in Network Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication networks face challenges in efficiently identifying and characterizing network issues such as packet loss, delay, and congestion, particularly in complex network topologies like leaf-spine or fat-tree networks, where intermittent issues and inconsistent sampling of user network traffic lead to unhelpful information and interference with normal traffic forwarding.
Innovation Solution
Implementing a consistent sampling policy across network nodes using low and high entropy data field matching, where low entropy fields define the network flow and high entropy fields provide a pseudo-random representation, allowing for detailed spatial and temporal analysis of packets, and marking packets for selective sampling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network probes are used to diagnose packet loss, then network issues can be identified, but the probes may traverse different paths and overlook intermittent packet loss issues
Solution Approach 1:
The patent creates a copy of the original packet's identification information and embeds it in the encapsulation header of sampled packets. This copy allows downstream network devices to track the same packet through multiple hops without needing to re-generate probe packets, ensuring consistent path traversal and reliable detection of intermittent packet loss issues.
Solution Approach 2:
The patent performs preliminary sampling and marking of packets at the ingress network device before they traverse the network. By pre-embedding identification information and sampling indicators in the packet header, the system ensures that the same packets are consistently sampled across multiple network devices, enabling reliable detection of intermittent issues that would otherwise be missed.
2Measurement precision
If user network traffic is sampled to identify network issues, then real-world packet characteristics are captured, but inconsistent sampling leads to unhelpful information and interferes with normal traffic forwarding
Solution Approach 1:
The patent applies different processing qualities to different packets: sampled packets receive additional encapsulation and tracking information, while non-sampled packets continue normal forwarding without modification. This local differentiation ensures that only selected packets are subjected to additional processing, minimizing interference with normal traffic while maintaining accurate packet information for analysis.
Solution Approach 2:
The patent extracts only the necessary identification and sampling information from the original packet and places it in the encapsulation header, rather than copying the entire packet. This extraction approach reduces the overhead on normal traffic forwarding while preserving the essential characteristics needed for network issue identification.
3Measurement precision
If detailed packet tracking information is collected across multiple network devices, then network issues can be precisely diagnosed, but the burden on collector circuitry and network paths increases
Solution Approach 1:
Instead of forwarding complete packet copies to collector circuitry at every network device, the patent embeds a compact identification copy in the encapsulation header that can be efficiently processed. This reduced-copy approach maintains the ability to perform detailed spatial and temporal analysis while significantly reducing the processing burden on collector circuitry and network paths.
Data Source
AI summary
In general, one aspect, the disclosure relates to a method for sampling packets in a network. The method includes receiving, by a first network device, a packet, making a first determination, by the first network device, that the packet is to be sampled, in response to the first determination: sampling the packet to obtain sampling data, storing sampling metadata associated with the packet, encapsulating, after the sampling, the packet to obtain an encapsulated packet, where the encapsulated packet comprises a bit that is set in an encapsulation header, wherein the bit is set based on the presence of the sampling metadata, and transmitting the encapsulated packet to a second network device.