[Paper Review] Finding Alternate Paths in the Internet:A Survey of Techniques for End-to-End Path Discovery
This paper surveys techniques for discovering end-to-end alternate paths in the Internet to improve resilience and Quality of Service, focusing on methods like path monitoring, disjoint path selection, BGP enhancements, multi-homing, detour routing, and Fast Re-Route mechanisms. It identifies key challenges in exploiting path diversity and outlines open research issues for future work.
The Internet provides physical path diversity between a large number of hosts, making it possible for networks to use alternative paths when one path fails to deliver the required Quality of Service. However, for various reasons, many established protocols (e.g. de facto Internet inter-domain routing protocol, Border-Gateway Protocol - BGP) do not fully exploit such alternate paths. This paper surveys research into techniques for discovering end-to-end alternate paths, including those based on monitoring path performance, choosing paths that are maximally disjoint, and in routing across multiple paths. It surveys proposals for making BGP better able to exploit multiple paths and how multi-homing can create alternate paths. It also describes how alternate paths can be realized through detour routing (application layer mechanisms) and routing deflections (network layer mechanisms). It also discusses Fast Re-Route techniques for construction of backup routes. It concludes by surveying open research issues into the discovery and use of alternate paths in the Internet.
Motivation & Objective
- To identify the limitations of existing protocols like BGP in exploiting physical path diversity across the Internet.
- To examine techniques that enable discovery of alternate end-to-end paths for improved network resilience and Quality of Service (QoS).
- To analyze how multi-homing and network-layer or application-layer mechanisms can create and utilize alternate paths.
- To evaluate the effectiveness of Fast Re-Route and routing deflection techniques in constructing backup routes.
- To highlight open research challenges in discovering and dynamically utilizing alternate paths in production networks.
Proposed method
- Surveying existing research on end-to-end path discovery using path monitoring to assess real-time performance and availability.
- Evaluating path selection strategies that maximize disjointness (e.g., link, node, or SRLG disjoint paths) to enhance fault tolerance.
- Analyzing enhancements to BGP to support multi-path routing and enable load balancing or failover across multiple paths.
- Examining application-layer detour routing mechanisms that reroute traffic around failures without network-layer changes.
- Reviewing network-layer routing deflection techniques that redirect traffic via alternate paths using control-plane signaling.
- Assessing Fast Re-Route (FRR) techniques that precompute backup paths for rapid recovery after link or node failures.
Experimental results
Research questions
- RQ1Why do established protocols like BGP fail to fully utilize available physical path diversity in the Internet?
- RQ2What are the most effective techniques for discovering maximally disjoint alternate paths to improve fault tolerance?
- RQ3How can multi-homing be leveraged to create and utilize alternate end-to-end paths?
- RQ4In what ways can application-layer detour routing and network-layer deflection mechanisms improve path resilience?
- RQ5What are the key open challenges in dynamically discovering and deploying alternate paths in real-world Internet deployments?
Key findings
- Many existing protocols, including BGP, do not fully exploit the physical path diversity available in the Internet, leading to suboptimal resilience and QoS.
- Techniques based on path monitoring and performance measurement can effectively identify viable alternate paths in real time.
- Maximally disjoint path selection—especially in terms of links, nodes, or shared risk link groups—significantly improves fault tolerance.
- Enhancements to BGP to support multi-path routing enable better load balancing and failover, though deployment remains limited.
- Detour routing at the application layer and routing deflection at the network layer offer viable alternatives for fast recovery without requiring full protocol changes.
- Fast Re-Route mechanisms provide sub-second recovery times by pre-computing backup paths, but their scalability and deployment complexity remain open challenges.
Better researchstarts right now
From reading papers to final review, dramatically reduce your research time.
No credit card · Free plan available
This review was created by AI and reviewed by human editors.