Fast Recovery From Link Failures in Ethernet Networks

Link failures in wide area networks are common. To recover from such failures, a number of methods such as SONET rings, protection cycles, and source rerouting have been investigated. Two important considerations in such approaches are the recovery time and the needed spare capacity to complete the recovery. Usually, these techniques attempt to achieve a recovery time less than 50 ms. In this paper we introduce an approach that provides link failure recovery in a hitless manner, or without any appreciable delay. This is achieved by means of a method called diversity coding. We present an algorithm for the design of an overlay network to achieve recovery from single link failures in arbitrary networks via diversity coding. This algorithm is designed to minimize spare capacity for recovery. We compare the recovery time and spare capacity performance of this algorithm against conventional techniques in terms of recovery time, spare capacity, and a joint metric called Quality of Recovery (QoR). QoR incorporates both the spare capacity percentages and worst case recovery times. Based on these results, we conclude that the proposed technique provides much shorter recovery times while achieving similar extra capacity, or better QoR performance overall.

Resiliency to link failures in optical networks is becoming increasingly important due to the increasing data rate in the fiber. Path protection schemes attempt to guarantee a backup path for a connection upon a failure in the network, thereby reducing the recovery time for a connection. In this paper, we develop a failure dependent path protection scheme that dynamically assigns a primary path and backup paths, one for each failure that would affect the primary path. A connection established on the primary path will be re-established on its backup path only if a failure in the network affects the connection. We evaluate the performance of our developed protocol and compare with an alternative approach based on sub-graph routing that achieves high network utilization and low blocking probability at the cost of re-establishing connections even if the failure in the network does not affect the primary path of the connection. We observe that up to a factor of eight reduction in the number of reconfiguration scenarios is achieved with less than 10% reduction in effective network utilization and less than 3% reduction in fairness metrics for tolerating any single link failure in NSFNET and ARPA-2 networks. The failure dependent protection approach developed in this paper is also applicable to any general failure scenarios that are modeled as shared risk link group failures.

The paper proposes five novel protection strategies against single link failure in multicast multi-fiber WDM networks. These strategies are able to provide rapid recovery from failure by reserving sufficient preplanned spare capacity. Mathematical formulations based on integer linear programming (ILP) are derived to determine optimal multicast routing and wavelength assignment (MC-RWA) while achieving 100% restoration. Based on a sample network, it is shown that the joined-path recovery (JPR) scheme is the most cost-effective, whereas the ring-based protection strategies offer other advantages, i.e. fast and reliable recovery.

Broadcast TV distribution over an IP network requires stringent QoS constraints, such as low latency and loss. Streaming content in IPTV is typically delivered to distribution points on an IP backbone using IP multicast protocols such as Protocol Independent Multicast Source Specific Mode (PIM-SSM). Link-restoration using MPLS or layer-2 Fast Reroute (FRR) is a proven failure restoration technique. Link-based FRR creates a pseudo-wire or tunnel in parallel to the IP adjacencies (links); and thus, single link failures are transparent to the Interior Gateway Protocol (IGP) such as OSPF. Although one may choose the back-up path's IGP link weights to avoid traffic overlap during any single physical link failure, multiple failures may still cause traffic overlap with FRR. We present a cross-layer restoration approach that combines both FRR-based restoration for single link failure and “hitless” (i.e., without loss) PIM tree reconfiguration algorithms to prevent traffic overlap when multiple failures occur.

Network survivability has become a crucial requirement in all types of computer networks. It becomes even more significant for wavelength division multiplex (WDM) mesh networks due to their high speed and capacity. These networks are prone to link failures. A link failure may be a single or multiple link failure. A single-link failure is easy to locate and fix as compared to multiple link failures. A dual-link failure recovery technique has been proposed using p-cycles. This technique uses a replication method for the p-cycle circle. It is an enhancement of the original failure independent path protection p-cycle scheme. The replica properties of p-cycle have been used to protect the nodes through the same p-cycle available. Creating a new p-cycle always adds to the cost of the network, whereas using a replica of already existing p-cycle significantly reduces the network cost. The proposed technique is implemented using network simulator in three phases.

We propose a fast restoration method for bandwidth-variable paths against single link failure in large-scale networks. This method achieves 50-ms service recovery time in 10K-path-scale ring networks while suppressing required network resources.

Software-defined networks (SDN) usually require intervention of a controller to restore connectivity in case of link or node failures. To avoid this dependence, a fast failover function enables switches to locally detect a failure and deviate affected traffic. In this paper, we develop fast reroute (FRR) methods leveraging that feature and the principle of loop-free alternates (LFAs) from IP networks for network-wide configuration. We explain how LFAs, remote LFAs, and novel explicit-path LFAs may be implemented with OpenFlow so that SDN with general destination-based routing can be protected. An advanced loop detection and termination function is suggested to prevent loops that may be caused by FRR in case of severe failures. We evaluate the FRR methods on a large set of representative network topologies and compare them with existing FRR methods. Classic rLFAs from IP networks generate many loops in case of node failures so that they are not appropriate for SDN. One developed protection method is particularly suitable for SDN as it can protect all flows in a network against any single link and node failure. The solution is efficient because it secures most traffic also in case of dual failures, leads to short backup paths, and requires in most networks only a few additional forwarding entries per node. The latter is important as OpenFlow switches can accommodate only a moderate number of forwarding entries in their flow tables.

Ethernet is becoming a preferred technology to be extended to metropolitan area networks (MANs) due to its low cost, simplicity, and ubiquity. However, current Ethernet lacks a fast failure recovery mechanism as it reconstructs the spanning tree after the failure is detected, which commonly requires tens of seconds. Some fast failure-handling methods based on multiple spanning trees have been proposed in the literature, but these approaches are either centralized or require periodic message broadcasting over the entire network. In this paper, we propose a local restoration mechanism for metro Ethernet using multiple spanning trees, which is distributed and fast and does not need failure notification. Upon failure of a single link, the upstream switch locally restores traffic to preconfigured backup spanning trees. We propose two restoration approaches, connection-based and destination-based, to select backup trees. We formulate the tree preconfiguration problem that includes working spanning tree assignment and backup spanning tree configuration. We prove that the preconfiguration problem is NP-complete and develop an integer linear programming model. We also develop heuristic algorithms for each restoration approach to reduce the computation complexity. To evaluate the effectiveness of our heuristic algorithms, we carry out the simulation on grid and random networks. The simulation results show that our heuristic algorithms have comparable performance close to the optimal solutions, and both restoration approaches can efficiently utilize the network bandwidth to handle single link failures.

Empowered by the rapid advancement of fiber optic sensing techniques in recent years, network carriers are able to upgrade their network infrastructure beyond the basic communication services with extra sensing applications and services (e.g., monitoring traffic and road condition, leakage detection, etc.), thus evolving to a new era of Infrastructure-as-a-Sensor (IaaSr) or Network-as-a-Sensor (NaaSr). When network carriers upgrade their network infrastructures with distributed fiber optic sensing (DFOS) technique to provide IaaSr services, there will arise a critical challenge: how to provide survivable (or reliable) IaaSr services against network failures (e.g., fiber cut). In this work, for the first time, we investigate the problem of survivable DFOS placement against single link failure. More specifically, we study where to place the primary and backup sensors and how to assign the primary and backup fiber sensing routes, with the objective of minimizing the number of sensors used. We formulate the problem using Integer Linear Programming (ILP) to facilitate the optimal solution. In addition, we propose a set of efficient heuristic algorithms to solve the problem in a fast manner. In particular, the proposed Shared-one algorithm provides a cost-efficient shared protection, through a one-step global optimization of the assignment of primary and backup DFOS placement. We conduct extensive simulations to evaluate the performance of the proposed solutions. We find out that Shared-one can achieve a close-to-optimal performance, compared to the ILP optimal results, while outperforming the other heuristic solutions with an average performance improvement by at least 16%.

In this paper, we investigate efficient approaches and algorithms for protecting multicast sessions against any single link failure while establishing multicast sessions in WDM mesh networks. Since a single failure may affect whole nodes in a multicast group and causes severe service disruption and a lot of traffic loss, protecting critical multicast sessions against link failure such as fiber cut becomes important in WDM optical networks. One of the most efficient algorithms is optimal path pair-shared disjoint paths (OPP-SDP). In this algorithm every source-destination (SD) pair has the optimal path pair (working and protection path) between the source and destination node. Since degree of sharing among the paths is essential to reduce the total cost and blocking probability, we propose the longest path first-shared disjoint paths (LPF-SDP) algorithm which decides the priority of selection among SD pairs in a resource-saving manner. Our LPF-SDP is shown to outperform over OPP-SDP in terms of degree of sharing and blocking probability.

An efficient critical protection scheme for intra-domain routing using link characteristics

IP Fast Reroute (IPFRR) is resiliency technology to provide fast traffic recovery upon link or node failure for mission critical services. This technology is able to recover impacted traffic in 50 ms upon any single link or node failure. The key aspect of IP Fast Reroute technology is fast failure detection and pre-calculation of alternative path in advance, before the link or node failure occurs. If specific failure within the network occurs, IP Fast Reroute mechanism uses this pre-calculated alternative path to bypass failure and restores communication. In this paper, we bring survey of advanced Fast Reroute solutions for IP-based intra-domain networks. There have been several approaches proposed such as Directed LFA, Topology Independent LFA, BIER-TE FRR and others.

The current state of security and availability of the Internet is far from being commensurate with its importance. The number and strength of DDoS attacks conducted at the network layer have been steadily increasing. However, the single path (SP) routing used in today’s Internet lacks a mitigation scheme to rapidly recover from network attacks or link failure. In case of a link failure occurs, it can take several minutes until failover. In contrast, multi-path routing can take advantage of multiple alternative paths and rapidly switch to another working path. According to the level of available path control, we classfy the multi-path routing into two types, first-hop multi-path (FMP) and multi-hop multi-path (MMP) routing. Although FMP routing supported by networks, such as SD-WAN, shows marginal improvements over the current SP routing of the Internet, MMP routing supported by a global Internet architecture provides strong improvement under network attacks and link failure. MMP routing enables changing to alternate paths to mitigate the network problem in other hops, which cannot be controlled by FMP routing. To show this comparison with practical outcome, we evaluate network performance in terms of latency and loss rate to show that MMP routing can mitigate Internet hazards and provide high availability on global networks by 18 participating ASes in six countries. Our evaluation of global networks shows that, if network attacks or failures occur in other autonomous systems (ASes) that FMP routing cannot avoid, it is feasible to deal with such problems by switching to alternative paths by using MMP routing. When the global evaluation is under a transit-link DDoS attack, the loss rates of FMP that pass the transit-link are affected significantly by a transit-link DDoS attack, but the other alternative MMP paths show stable status under the DDoS attack with proper operation.

In this paper, we consider IP fast recovery from single-link failures in a given network topology. The basic idea is to replace some existing routers with a designated switch. When a link fails, the affected router will send all the affected traffic to the designated switch (through pre-configured IP tunnels), which will deliver the affected traffic to its destination without using the failed link. The goal of the approach is to achieve faster failure recovery than traditional routing protocols that employ reactive computing upon link failures. Software-Defined Networking (SDN) switches can serve as the designated switches because they can flexibly redirect affected traffic to other routes, instead of only to the shortest paths in the network. However, SDN switches are very expensive. Our objective is to minimize the number of SDN switches needed and to guarantee that the network can still recover from any single-link failure. For networks with uniform link costs, we show that using normal non-SDN switches with IP tunneling capability as designated switches can guarantee recovery from any single-link failure. For networks with general link costs, we find that not all single-link failures can be recovered by using non-SDN switches as designated switches; by using SDN switches only when necessary, we can reduce the total number of SDN switches needed compared to an existing work. We conduct extensive simulations to verify our proposed approaches.

Simultaneous link failures are common in IP networks [1]. In this paper, we develop a technique for locating multiple failures in Service Provider or Enterprise IP networks using active measurement. We propose a two-phased approach that minimizes both the additional traffic due to probe messages and the measurement infrastructure costs. In the first phase, using elements from max-plus algebra theory, we show that the optimal set of probes can be determined in polynomial time, and we provide an algorithm to find this optimal set of probes. In the second phase, given the optimal set of probes, we compute the location of a minimal set of measurement points (beacons) that can generate these probes. We show that the beacon placement problem is NP-hard and propose a constant factor approximation algorithm for this problem. We then apply our algorithms to existing ISP networks using topologies inferred by the Rocketfuel tool [2]. We study in particular the difference between the number of probes and beacons that are required for multiple and single failure(s) diagnosis.