Inter-domain SDN Traffic Engineering Resource Allocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In large software-defined networks (SDN), centralized controllers face challenges in efficiently managing resource provisioning due to increased computational complexity and latency, especially when predicting resource requirements for inter-domain traffic, leading to inefficient resource allocation and reduced performance.

Innovation Solution

The implementation of inter-domain SDN traffic engineering techniques that involve receiving border element information from multiple SDN controllers, computing inter-domain allocation constraints, and provisioning resources to reserve capacity for inter-domain traffic, ensuring equitable access to resources and maintaining minimum throughput rates, either through centralized or distributed methods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If centralized SDN controllers are used to manage resource provisioning in large networks, then network service management is achieved through abstraction of lower level functionality, but latencies involved with collecting network information and distributing provisioning instructions significantly delay policy implementation

Engineering Contradiction:
Improvenetwork service managementVSAvoidpolicy implementation delay
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The network is divided into multiple domains, each managed by a local SDN controller that handles resource provisioning independently. This segmentation allows parallel processing of provisioning requests across domains, eliminating the single-point bottleneck of centralized controllers and significantly reducing policy implementation latency while maintaining standardized interfaces for network-wide coordination.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The architecture transitions from a single-dimensional centralized control model to a multi-dimensional hierarchical control structure, where control functions are distributed across multiple layers (local domain controllers, inter-domain controllers, and centralized coordination). This dimensional expansion enables simultaneous local optimization and global coordination, reducing implementation delays through parallel decision-making at different hierarchical levels.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If centralized SDN controllers service large networks, then comprehensive network control is achieved, but computational complexity of optimization algorithms increases significantly as additional links are added

Engineering Contradiction:
Improvenetwork controlVSAvoidcomputational complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The computational workload of optimization algorithms is segmented and distributed to local SDN controllers within each domain. Each controller solves optimization problems only for its local domain, dramatically reducing individual computational complexity compared to a centralized solver handling the entire network. Inter-domain optimization is handled separately by inter-domain controllers, creating a divide-and-conquer approach that maintains comprehensive control while managing complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Each local SDN controller performs optimization for its specific domain rather than attempting to optimize the entire network globally. This partial action approach allows controllers to use simpler, faster optimization algorithms tailored to local conditions, while inter-domain coordination ensures overall network optimality is achieved through iterative exchange of border element information and constraints.

Inventive Principle:
Principle #16Partial or excessive action

3Adaptability or versatility

If resources are reserved for inter-domain traffic, then equitable access to resources is achieved, but available capacity for intra-domain traffic may be reduced

Engineering Contradiction:
Improveresource allocation fairnessVSAvoidintra-domain traffic throughput
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The system dynamically adjusts resource allocation parameters based on traffic patterns and service level agreements. Border element information and allocation constraints are exchanged between controllers to compute optimal resource reservations that satisfy inter-domain traffic requirements while maximizing intra-domain throughput. This parameter optimization ensures fair access for inter-domain traffic without unnecessarily reducing intra-domain capacity.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

Resource reservations for inter-domain traffic are made dynamic rather than static, allowing the system to adapt to changing network conditions. The inter-domain TE controller continuously monitors border element utilization and adjusts allocation constraints in real-time, releasing reserved capacity when inter-domain traffic demand is low and tightening reservations when demand increases, thereby maintaining both fairness and productivity across varying loads.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3090517B1Inter-domain SDN traffic engineering
Publication Date: 2018.01.24 HUAWEI TECH CO LTD
  • EP3090517B1 patent drawingFigure 1~6
  • EP3090517B1 patent drawingFigure 2
  • EP3090517B1 patent drawingFigure 3

AI summary

Constraining resource provisioning by domain controllers based on resource requirements projected for remotely-originating inter-domain traffic can improve utilization efficiency and link reliability in multi-domain software defined network (SDN) architectures. A domain controller may be required to reserve a portion of inter-domain link capacity for transporting remotely-originating traffic. This may limit the inter-domain link capacity available for transporting locally-originating traffic in a manner that ensures remotely-originating traffic flows have equitable and/or adequate access to resources of inter-domain links. Alternatively, a domain controller may be required to maintain a minimum throughput rate for remotely-originating traffic, which may cause remotely-originating traffic to be prioritized over locally-originating traffic when necessary to maintain the minimum throughput rate. Provisioning constraints can be generated in a centralized or distributed fashion.