Probabilistic SLA Token Bucket Rate Regulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication networks face challenges in managing bursty traffic, particularly in backhaul networks, where conventional techniques like token buckets lack a formal approach for setting token rates, leading to inefficiencies and bottlenecks, especially when handling peak capacity demands and achieving statistical multiplexing gains.
Innovation Solution
The method involves determining specific token bucket rate parameters based on peak rate requirements, data link capacity, nominal speed, and file sizes to regulate data packet transmission across networks, using a matrix-based approach that includes multiple drop precedence levels and timescales to ensure efficient data handling and prevent bottlenecks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If link capacity is dimensioned by a CIR according to an SLA, then all packets will be transmitted with a high probability, but bursty traffic may encounter a bottleneck on the link during high periods
Solution Approach 1:
The patent implements dynamic token bucket rates that adapt to traffic conditions and service priorities. Instead of fixed CIR values, the system adjusts token generation rates based on current link utilization, traffic burst patterns, and SLA requirements, allowing the network to respond dynamically to changing conditions and avoid bottlenecks while maintaining reliability
Solution Approach 2:
The system changes key parameters including token bucket rates, drop precedence levels, and timescale configurations based on traffic analysis and network conditions. By modifying these parameters dynamically, the network can accommodate bursty traffic patterns while maintaining packet transmission guarantees for critical services
2Productivity
If statistical multiplexing is used to achieve capacity gains, then capacity efficiency improves, but it does not achieve significant capacity gains when the number of services is relatively small
Solution Approach 1:
The patent segments traffic into multiple priority levels with different drop precedence values (DP1, DP2, DP3, etc.). This segmentation allows the system to apply different token bucket rates and protection levels to different traffic types, enabling effective statistical multiplexing even with fewer services by carefully managing each segment's resource allocation and burst behavior
Solution Approach 2:
The token bucket mechanism acts as an intermediary between traffic sources and the link capacity. It mediates traffic bursts by controlling the rate at which packets can be transmitted, smoothing out variations in traffic demand and enabling more reliable statistical multiplexing gains even with a limited number of services
3Ease of operation
If conventional token bucket techniques are used, then traffic regulation is implemented, but there is no formal approach for setting token rates, leading to inefficiencies and bottlenecks
Solution Approach 1:
The system implements feedback mechanisms that monitor link utilization, traffic patterns, and packet drop rates. This feedback is used to continuously adjust token bucket rates and configurations, ensuring optimal network performance and preventing bottlenecks. The feedback loop enables the system to learn from past performance and adapt token rate settings accordingly
Solution Approach 2:
The patent performs preliminary analysis of traffic characteristics, service requirements, and link capacity before configuring token bucket parameters. By pre-calculating optimal token rates based on expected traffic patterns and SLA requirements, the system avoids inefficiencies that would arise from arbitrary or reactive parameter settings
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Embodiments include exemplary methods and/or procedures for regulating transmission of data packets between a first network and a second network over a datalink. Embodiments include determining a first plurality of token bucket rate (TBR) parameters, each TBR parameter corresponding to a one of a first plurality of packet drop precedence (DP) levels and one of a first plurality of timescales (TS). The determination of the first plurality of bucket rate parameters is based on a peak rate requirement, the data link capacity, and a nominal speed requirement associated with the data link. Embodiments also include determining a second plurality of TBR parameters based on the first plurality of TBR parameters and a guaranteed rate requirement, the second plurality comprising a further DP level than the first plurality. Embodiments also include regulating data packets sent between the first network and the second network via the data link based on the second plurality of TBR parameters. Embodiments also include network nodes configured to perform the exemplary methods and/or procedures.