Dynamic load balancing is a popular recent technique that protects ISP networks from sudden congestion caused by load spikes or link failures. Dynamic load balancing protocols, however, require schemes for splitting traffic across multiple paths at a fine granularity. Current splitting schemes present a tussle between slicing granularity and packet reordering. Splitting traffic at the granularity of packets quickly and accurately assigns the desired traffic share to each path, but can reorder packets within a TCP flow, confusing TCP congestion control. Splitting traffic at the granularity of a flow avoids packet reordering but may overshoot the desired shares by up to 60% in dynamic environments, resulting in low end-to-end network goodputContrary to popular belief, we show that one can systematically split a single flow across multiple paths without causing packet reordering. We propose FLARE, a new traffic splitting algorithm that operates on bursts of packets, carefully chosen to avoid reordering. Using a combination of analysis and trace-driven simulations, we show that FLARE attains accuracy and responsiveness comparable to packet switching without reordering packets. FLARE is simple and can be implemented with a few KB of router state

/pdf/dynamic-load-balancing-without-packet-reordering-37fm6tjcie.pdf

Dynamic load balancing without packet reordering

Traffic engineering is an important mechanism for Internet network providers seeking to optimize network performance and traffic delivery. Routing optimization plays a key role in traffic engineering, finding efficient routes so as to achieve the desired network performance. In this survey we review Internet traffic engineering from the perspective of routing optimization. A taxonomy of routing algorithms in the literature is provided, dating from the advent of the TE concept in the late 1990s. We classify the algorithms into multiple dimensions: unicast/multicast, intra-/inter- domain, IP-/MPLS-based and offline/online TE schemes. In addition, we investigate some important traffic engineering issues, including robustness, TE interactions, and interoperability with overlay selfish routing. In addition to a review of existing solutions, we also point out some challenges in TE operation and important issues that are worthy of investigation in future research activities.

/pdf/an-overview-of-routing-optimization-for-internet-traffic-568i46uj1v.pdf

An overview of routing optimization for internet traffic engineering

The overlay approach has been widely used by many service providers for traffic engineering in large Internet backbones. In the overlay approach, logical connections are set up between edge nodes to form a full mesh virtual network on top of the physical topology. IP routing is then run over the virtual network. Traffic engineering objectives are achieved through carefully routing logical connections over the physical links. Although the overlay approach has been implemented in many operational networks, it has a number of well-known scaling issues. This paper proposes a new approach, which we call the integrated approach, to achieve traffic engineering without full-mesh overlaying. In the integrated approach, IP routing runs natively over the physical topology rather than over the virtual network. Traffic engineering objectives are realized by setting appropriate link metrics in IP routing protocols. We first illustrate our approach with a simple network, then present a formal analysis of the integrated approach and a method for deriving the appropriate link weights. Our analysis shows that for any given set of optimal routes of the overlay approach with respect to a set of traffic demands, the integrated approach can achieve exactly the same result by reproducing them as shortest paths. We further extend the result to a more generic one: for any arbitrary set of routes, as long as they are not loopy, they can be converted to shortest-paths with respect to some set of positive link weights. A theoretical insight of our result is that the optimal routing (with respect to any objective function) is always shortest path routing with respect to some appropriate positive link weights.

Internet traffic engineering without full mesh overlaying

Multiprotocol label switching (MPLS) is rapidly emerging as an Internet Engineering Task Force (IETF) standard intended to enhance the speed, scalability, and service provisioning capabilities in the Internet. MPLS uses the technique of packet forwarding based on labels, to enable the implementation of a simpler high-performance packet forwarding engine. This also decouples packet forwarding from routing, facilitating the provision of varied routing services independent of the packet forwarding paradigm. The authors track the evolution of this technology in relation to other existing technologies. Then an overview of the MPLS architecture and design is provided. In addition, some of the work that was a precursor to MPLS is discussed, as well as related issues and debates.

Evolution of multiprotocol label switching

Unlike traditional routing schemes that route all traffic along a single path, multipath routing strategies split the traffic among several paths in order to ease congestion. It has been widely recognized that multipath routing can be fundamentally more efficient than the traditional approach of routing along single paths. Yet, in contrast to the single-path routing approach, most studies in the context of multipath routing focused on heuristic methods. We demonstrate the significant advantage of optimal (or near optimal) solutions. Hence, we investigate multipath routing adopting a rigorous (theoretical) approach. We formalize problems that incorporate two major requirements of multipath routing. Then, we establish the intractability of these problems in terms of computational complexity. Finally, we establish efficient solutions with proven performance guarantees.

/pdf/multipath-routing-algorithms-for-congestion-minimization-5g4wa06e8f.pdf

Multipath routing algorithms for congestion minimization

This paper considers explicit routing algorithms for Internet traffic engineering. Explicit routing is seen to be a much more capable solution for improving network utilization than the current destination-based routing and the multi-protocol label switching (MPLS) standard has made explicit routes implementable. ISP can now have fine granularity control over the traffic distribution across their backbones by carefully overlaying explicit routes over the physical network. The basic traffic engineering problem is how to set up explicit routes to meet bandwidth demands between the edge nodes of the network and at the same time to optimize the network performance. We model the traffic engineering problem as an optimization problem with the objective of minimizing congestion and maximizing potential for traffic growth. We present two mathematical formulations, one linear programming for the case of allowing demand bifurcation and one integer programming for the case of disallowing demand bifurcation. While the bifurcation case can be solved to optimality, we show that the non-bifurcation case is NP-hard. Four heuristic schemes are proposed for the non-bifurcation case, with the most sophisticated one being based on re-routing of split demands in the optimal solution of the bifurcation case. The performance of these heuristic schemes are tested in a large backbone topology. Our results show that shortest-path and minimum hop algorithms, although widely used in current routing protocols, perform poorly, white the re-routing approach performs best.

Yufei Wang

Papers

Explicit routing algorithms for Internet traffic engineering