The QUIC protocol, designed to reduce latency and improve internet security, faces goodput performance challenges in high-speed networks, particularly with single-thread implementations. This poster extends an existing userspace Multipath QUIC (MPQUIC) implementation to enhance the goodput of a single connection by pinning different network path logics to different cores. Our solution, mcMPQUIC, achieves a goodput of up to 20 Gbps with ten paths/cores, surpassing the baseline MPQUIC performance by more than five times.

A growing number of Internet applications require low latency. Unfortunately, most of these applications cannot use the rich features of the QUIC protocol since it only uses retransmissions to cope with packet losses. We propose, implement and evaluate QUIRL, a revisit of the QUIC loss recovery mechanism. QUIRL relies on Forward Erasure Correction (FEC) only if it is needed by the application’s latency requirements and uses classical retransmissions otherwise. We implement QUIRL and evaluate its performance for real-time video and HTTP/3. Compared to previous works adding FEC to QUIC, QUIRL is the first to be evaluated with and obtain significant performance improvements for popular applications over real lossy networks. Our evaluation shows that for video QUIRL improves the video quality while meeting strict delay requirements. For HTTP/3 transfers, QUIRL efficiently reduces the tail latency when packet losses occur without causing harm when there are no losses. We confirm these results using emulation over a wide ranges of bandwidth, delays and loss scenarios. We release our QUIRL implementation to encourage other researchers and industry to explore in more details the use of FEC in QUIC.

Packet losses are common events in today’s networks. They usually result in longer delivery times for application data since retransmissions are the de facto technique to recover from such losses. Retransmissions is a good strategy for many applications but it may lead to poor performance with latency-sensitive applications compared to network coding. Although different types of network coding techniques have been proposed to reduce the impact of losses by transmitting redundant information, they are not widely used. Some niche applications include their own variant of Forward Erasure Correction (FEC) techniques, but there is no generic protocol that enables many applications to easily use them. We close this gap by designing, implementing and evaluating a new Flexible Erasure Correction (FlEC) framework inside the newly standardized QUIC protocol. With FlEC, an application can easily select the reliability mechanism that meets its requirements, from pure retransmissions to various forms of FEC. We consider three different use cases:$(i)$bulk data transfer,$(ii)$file transfers with restricted buffers and$(iii)$delay-constrained messages. We demonstrate that modern transport protocols such as QUIC may benefit from application knowledge by leveraging this knowledge in FlEC to provide better loss recovery and stream scheduling. Our evaluation over a wide range of scenarios shows that the FlEC framework outperforms the standard QUIC reliability mechanisms from a latency viewpoint.

With new Low Earth Orbit satellite constellations such as Starlink, satellite-based Internet access is becoming an alternative to traditional fixed and wireless technologies with comparable throughputs and latencies. In this paper, we investigate the user-perceived performance of Starlink. Our measurements show that latency remains low and does not vary significantly under idle or lightly loaded links. Compared to another commercial Internet access using a geostationary satellite, Starlink achieves higher TCP throughput and provides faster web browsing. To avoid interference from performance enhancing proxies commonly used in satellite networks, we also use QUIC to assess performance under load and packet loss. Our results indicate that delay and packet loss increase slightly under load for both upload and download.

Researchers often face the lack of data on large operational networks to understand how they are used, how they behave, and sometimes how they fail. This data is crucial to drive the evolution of Internet protocols and develop techniques such as traffic engineering, DDoS detection and mitigation. Companies that have access to measurements from operational networks and services leverage this data to improve the availability, speed, and resilience of their Internet services. Unfortunately, the availability of large datasets, especially collected regularly over a long period of time, is a daunting task that remains scarce in the literature.

The Transmission Control Protocol (TCP) is one of the key Internet protocols. It is used by a broad range of applications. TCP was designed when there was typically a single path between a client and a server. Today’s networks provide higher path diversity, yet TCP still only uses the single path selected by the network layer. This limits the ability of TCP to react to events such as interdomain failures or highly congested peering links. We propose the TCP Path Changer (TPC), a set of eBPF programs that are incorporated into the Linux TCP/IP stack to make it more agile. To illustrate the benefits of our approach, we first demonstrate that TPC can quickly reroute an ongoing TCP connection around a failure. We then show that TPC can also monitor the round-trip-time of active TCP connections and automatically reroute them if it becomes too high. Our evaluation of TPC in emulated networks evidences the significant performance benefits of a path-aware transport protocol.

TCP and TLS are among the essential protocols in today's Internet. TCP ensures reliable data delivery while TLS secures the data transfer. Although they are very often used together, they have been designed independently following the Internet layered model. This paper demonstrates the various benefits that a closer integration between TCP and TLS would bring.

TCP and TLS are among the most essential protocols in today's Internet. TCP ensures reliable delivery of data while TLS secures the data transfer. Following the layered model, TLS was designed to be as independent as possible from the underlying transport protocol.

Thanks to the standardization of routing protocols such as BGP, OSPF or IS-IS, Internet Service Providers (ISP) and enterprise networks can deploy routers from various vendors. This prevents them from vendor-lockin problems. Unfortunately, this also slows innovation since any new feature must be standardized and implemented by all vendors before being deployed.

Routing protocols such as BGP and OSPF are key components of Internet Service Provider (ISP) networks. These protocols and the operator's requirements evolve over time, but it often takes many years for network operators to convince their different router vendors and the IETF to extend routing protocols. Some network operators, notably in enterprise and datacenters have adopted Software Defined Networking (SDN) with its centralised control to be more agile. We propose a new approach to implement routing protocols that enables network operators to innovate while still using distributed routing protocols and thus keeping all their benefits compared to centralised routing approaches. We extend a routing protocol with a virtual machine that is capable of executing plugins. These plugins extend the protocol or modify its underlying algorithms through a simple API to meet the specific requirements of operators. We modify the OSPF and BGP implementations provided by FRRouting and demonstrate the applicability of our approach with several use cases.

Segment Routing (SR) is a powerful tool to solve traffic engineering in large networks. It enables steering the traffic along any arbitrary network path while limiting scalability issues as routers do not need to maintain a global state. Mathematical programming approaches proposed so far for SR either do not scale well with the size of topology or impose a strong limit on the number of possible detours (typically at most one). Moreover they do not support Segment Routing fully by ignoring the adjacency segments. This paper leverages column generation, a widely used technique for solving large scale linear programs, combined with a novel dynamic program for solving the pricing problem. Our approach reaches near optimal solutions with gap guarantees by also computing a strong lower-bound tighter than the multi-commodity flow relaxation. It scales even on large topologies and exploits the full expressiveness of SR including adjacency segments. Our experiments compared with existing traffic engineering techniques on various topologies and demand matrices demonstrate the advantages of our approach in terms of scalability, any-time behavior and quality of the solutions.

Originally implemented by Google, QUIC gathers a growing interest by providing, on top of UDP, the same service as the classical TCP/TLS/HTTP/2 stack. The IETF will finalise the QUIC specification in 2019. A key feature of QUIC is that almost all its packets, including most of its headers, are fully encrypted. This prevents eavesdropping and interferences caused by middleboxes. Thanks to this feature and its clean design, QUIC is easier to extend than TCP. In this paper, we revisit the reliable transmission mechanisms that are included in QUIC. More specifically, we design, implement and evaluate Forward Erasure Correction (FEC) extensions to QUIC. These extensions are mainly intended for high-delays and lossy communications such as In-Flight Communications. Our design includes a generic FEC frame and our implementation supports the XOR, Reed-Solomon and Convolutional RLC error-correcting codes. We also conservatively avoid hindering the loss-based congestion signal by distinguishing the packets that have been received from the packets that have been recovered by the FEC. We evaluate its performance by applying an experimental design covering a wide range of delay and packet loss conditions with reproducible experiments. These confirm that our modular design allows the protocol to adapt to the network conditions. For long data transfers or when the loss rate and delay are small, the FEC overhead negatively impacts the download completion time. However, with high packet loss rates and long delays or smaller files, FEC allows drastically reducing the download completion time by avoiding costly retransmission timeouts. These results show that there is a need to use FEC adaptively to the network conditions.

The Transmission Control Protocol (TCP) is one of the most important protocols in today's Internet. Its specification and implementations have been refined for almost forty years. The Linux TCP stack is one of the most widely used TCP stacks given its utilisation on servers and Android smartphones and tablets. However, TCP and its implementations evolve very slowly. In this paper, we demonstrate how to leverage the eBPF virtual machine that is part of the recent versions of the Linux kernel to make the TCP stack easier to extend. We demonstrate a variety of use cases where the eBPF code is injected inside a running kernel to update or tune the TCP implementation. We first implement the TCP User Timeout Option. Then we propose a new option that enables a client to request a server to use a specific congestion control scheme. Our third extension is a TCP option that sets the initial congestion window. We then demonstrate how eBPF code can be used to tune the acknowledgment strategy.

Application requirements evolve over time and the underlying protocols need to adapt. Most transport protocols evolve by negotiating protocol extensions during the handshake. Experience with TCP shows that this leads to delays of several years or more to widely deploy standardized extensions. In this paper, we revisit the extensibility paradigm of transport protocols.

Motivated by conversations with operators and by possibilities to unlock future Internet-based applications, we study how to enable Internet Service Providers (ISPs) to reliably offer connectivity through disjoint paths as an advanced, value-added service. As ISPs are increasingly deploying Segment Routing (SR), we focus on implementing such service with SR. We introduce the concept of robustly disjoint paths, pairs of paths that are constructed to remain disjoint even after an input set of failures, with no external intervention (e.g., configuration change). We extend the routing theory, study the problem complexity, and design efficient algorithms to automatically compute SR-based robustly disjoint paths. Our algorithms enable a fully automated approach to offer the disjoint-path connectivity, based on configuration synthesis. Our evaluation on real topologies shows that such an approach is practical, and scales to large ISP networks.

With the advent of Software Defined Networks (SDN), Network Function Virtualisation (NFV) or Service Function Chaining (SFC), operators expect networks to support flexible services beyond the mere forwarding of packets. The network programmability framework which is being developed within the IETF by leveraging IPv6 Segment Routing enables the realisation of in-network functions.

Quick UDP Internet Connection (QUIC) is a recent protocol initiated by Google that combines the functions of HTTP/2, TLS, and TCP directly over UDP, with the goal to reduce the latency of client-server communication. It can replace the traditional HTTP/TLS/TCP stack and the IETF has chartered a working group to standardize it. QUIC encrypts all data and most protocol headers to prevent interferences from middleboxes.

Multipath TCP is a recent TCP extension that enables the utilization of different paths for a single connection. This provides various benefits including bandwidth aggregation and fast handovers on mobiles. A Multipath TCP connection starts with a single TCP connection called subflow and other subflows are added later to increase bandwidth or support failover. One drawback of Multipath TCP is that it is not currently compatible with stateless load balancers which rely on the five-tuple for their forwarding decision. This hinders the deployment of Multipath TCP. We show that this limitation can be circumvented with a small change to the handling of the initial subflow. Clients use this subflow to discover the load-balanced server and the additional Multipath TCP subflows are terminated at a unique address associated to each physical server. With this small change, Multipath TCP becomes compatible with existing stateless load balancers. Furthermore, we show that the same approach enables anycast Multipath TCP services, a major benefit given the difficulty of deploying anycast TCP services. We implement this modification in the Linux kernel and demonstrate its benefits with several micro benchmarks.

MultiFath TCP (MPTCP) is a recent TCP extension that enables hosts to send data over multiple paths for a single connection. It is already deployed for various use cases, notably on smartphones. In parallel with this, there is a growing deployment of encryption and authentication techniques to counter various forms of security attacks. Tcpcrypt and TLS are some of these security solutions. In this paper, we propose MPTCPsec, a MultiPath TCP extension that closely integrates authentication and encryption inside the protocol itself. Our design relies on an adaptation for the multipath environment of the ENO option that is being discussed within the IETF tcpinc working group. We then detail how MultiPath TCP needs to be modified to authenticate and encrypt all data and authenticate the different TCP options that it uses. Finally, we implement our proposed extension in the reference implementation of MultiPath TCP in the Linux kernel and we evaluate its performance.

The support for safe network updates, i.e., live modification of device behavior without service disruption, is a critical primitive for current and future networks. Several techniques have been proposed by previous works to implement such a primitive. Unfortunately, existing techniques are not generally applicable to any network architecture, and typically require high overhead (e.g., additional memory) to guarantee strong consistency (i.e., traversal of either initial or final paths, but never a mix of them) during the update. In this paper, we deeply study the problem of computing operational sequences to safely and quickly update arbitrary networks. We characterize cases, for which this computation is easy, and revisit previous algorithmic contributions in the new light of our theoretical findings. We also propose and thoroughly evaluate a generic sequence-computation approach, based on two new algorithms that we combine to overcome limitations of prior proposals. Our approach always finds an operational sequence that provably guarantees strong consistency throughout the update, with very limited overhead. Moreover, it can be applied to update networks running any combination of centralized and distributed control-planes, including different families of IGPs, OpenFlow or other SDN protocols, and hybrid SDN networks. Our approach therefore supports a large set of use cases, ranging from traffic engineering in IGP-only or SDN-only networks to incremental SDN roll-out and advanced requirements (e.g., per-flow path selection or dynamic network function virtualization) in partial SDN deployments.

To guarantee correct operation of their networks, operators have to promptly detect and diagnose data-plane issues, like broken interface cards or link failures. Networks are becoming more complex, with a growing number of Equal Cost MultiPath (ECMP) and link bundles. Hence, some data-plane problems (e.g. silent packet dropping at one router) can hardly be detected with control-plane protocols or simple monitoring tools like ping or traceroute. In this paper, we propose a new technique, called SCMon, that enables continuous monitoring of the data-plane, in order to track the health of all routers and links. SCMon leverages the recently proposed Segment Routing (SR) architecture to monitor the entire network with a single box (and no additional monitoring protocol). In particular, SCMon uses SR to (i) force monitoring probes to travel over cycles; and (ii) test parallel links and bundles at a per-link granularity. We present original algorithms to compute cycles that cover all network links with a limited number of SR segments. Further, we prototype and evaluate SCMon both with simulations and Linux-based emulations. Our experiments show that SCMon quickly detects and precisely pinpoints data-plane problems, with a limited overhead.

Multipath TCP was designed and implemented as a backward compatible replacement for TCP. For this reason, it exposes the standard socket API to the applications that cannot control the utilisation of the different paths. This is a key feature for applications that are unaware of the multipath nature of the network. On the contrary, this is a limitation for applications that could benefit from specific knowledge to use multiple paths in a way that fits their needs. As the specific knowledge of an application can not be known in advance, we propose a Multipath TCP path manager that delegates the management of the paths to the applications. This path manager enables applications to control how the different paths are used to transfer data. We implement this path manager above the Linux Multipath TCP kernel. It is composed of a kernel part that exposes events and commands to an userspace application that controls the key functions of Multipath TCP such as the creation/suppression of subflows or reactions to retransmissions. We demonstrate the benefits of this path manager on different use cases.

Network operators can and do deploy multiple routing control-planes, e.g., by running different protocols or instances of the same protocol. With the rise of SDN, multiple control-planes are likely to become even more popular, e.g., to enable hybrid SDN or multi-controller deployments. Unfortunately, previous works do not apply to arbitrary combinations of centralized and distributed control-planes. In this paper, we develop a general theory for coexisting control-planes. We provide a novel, exhaustive classification of existing and future control-planes (e.g., OSPF, EIGRP, and Open-Flow) based on fundamental control-plane properties that we identify. Our properties are general enough to study centralized and distributed control-planes under a common framework. We show that multiple uncoordinated control-planes can cause forwarding anomalies whose type solely depends on the identified properties. To show the wide applicability of our framework, we leverage our theoretical insight to (i) provide sufficient conditions to avoid anomalies, (ii) propose configuration guidelines, and (iii) define a provably-safe procedure for reconfigurations from any (combination of) control-planes to any other. Finally, we discuss prominent consequences of our findings on the deployment of new paradigms (notably, SDN) and previous research works.

Smartphones are the most popular mobile multihomed devices. End-user expects that thanks to their WiFi and cellular interfaces, they are able to seamlessly use all available networks. Unfortunately, reality tells us that seamless coexistence between cellular and WiFi is not as simple as what the user expect. Several cellular/WiFi coexistence technologies have been proposed during the last years. Some of them have been deployed. Recently, Multipath TCP received a lot of attention when it was selected by Apple to support its voice recognition (Siri) application. As of this writing, Siri is the only deployed smartphone application that uses Multipath TCP. and there is no public information about the benefits of using Multipath TCP with it. Multipath TCP is a TCP extension that allows to send data from one end-to-end connection over different paths. On a smartphone, Multipath TCP allows the applications to simultaneously send and receive data over both WiFi and cellular interfaces. It achieves this objective by establishing one TCP connection, called subflow, over each interface. Once the subflows have been established, data can be sent over any of the subflows. Researchers have analyzed the performance of Multipath TCP in such hybrid networks. However, these analyses have been performed with bulk transfers between laptops and servers. As of this writing, no detailed analysis of the performance of real smartphone applications with Multipath TCP has been published. We fill this gap in this paper by proposing a framework that automates user actions on Android smartphone applications to perform network measurements. We use it to analyze how eight popular smartphone applications interact with Multipath TCP.

Ultra-low latency is a key component of safety-critical operations such as robot-assisted remote surgery or financial applications where every single millisecond counts. In this paper, we show how network operators can build upon the recently proposed Segment Routing architecture to provide a traffic duplication service to better serve the users of such demanding applications. We propose the first implementation of Segment Routing in the Linux kernel and leverage it to provide a traffic duplication service that sends packets over disjoint paths. Our experiments show that with such a service existing TCP stacks can preserve latency in the presence of packet losses. We also propose and evaluate an efficient algorithm that computes disjoint paths that can be realised by using segments. Our evaluation with real and synthetic network topologies shows that our proposed algorithms perform well in large networks.

SDN simplifies network management by relying on declarativity (high-level interface) and expressiveness (network flexibility). We propose a solution to support those features while preserving high robustness and scalability as needed in carrier-grade networks. Our solution is based on (i) a two-layer architecture separating connectivity and optimization tasks; and (ii) a centralized optimizer called framework, which translates high-level goals expressed almost in natural language into compliant network configurations. Our evaluation on real and synthetic topologies shows that framework improves the state of the art by (i) achieving better trade-offs for classic goals covered by previous works, (ii) supporting a larger set of goals (refined traffic engineering and service chaining), and (iii) optimizing large ISP networks in few seconds. We also quantify the gains of our implementation, running Segment Routing on top of IS-IS, over possible alternatives (RSVP-TE and OpenFlow).

Simultaneously providing flexibility, evolvability and correctness of routing is one of the basic and still unsolved problems in networking. Route redistribution provides a tool, used in many enterprise networks, to either partition a network into multiple routing domains or merge previously independent networks. However, no general technique exists for changing a live network's route redistribution configuration without incurring packet losses and service disruptions. In this paper, we study the problem of how to safely transition between route redistribution configurations. We investigate what anomalies may occur in the reconfiguration process, showing that many long-lasting forwarding loops can and do occur if naive techniques are applied. We devise new sufficient conditions for anomaly-free reconfigurations, and we leverage them to build provably safe and practical reconfiguration procedures. Our procedures enable seamless network re-organizations to accomplish both short-term objectives, such as local repair or traffic engineering, and long-term requirement changes.

Traffic between an edge network and the rest of the Internet can be represented as a dynamic loop-free graph. Understanding in depth the dynamics in time and space (spatial structure, topological breadth, destination persistency, traffic dominating paths) of this graph provides significant insight on the Internet internal architecture and capabilities. This paper analyzes inter-domain traffic from a large campus network based on one month by way of Netflow measurements. Our analysis reveals the topological properties and structure of the traffic graph (breadth, depth, volume), the stability of contacted destinations and the relationship between their popularity and their path length. Based on the observed traffic, we explore the suitability of a simple mathematical model to describe the structure of the outgoing traffic graph.

To configure, troubleshoot and operate their networks, operators often have no alternatives than relying on error-prone manual procedures. The emerging Software Defined Networking paradigm opens new possibilities for more structured networking methodologies.We argue that provably-effective practices can be borrowed from more developed engineering fields, especially software engineering. In this paper, we propose an adaptation of test-driven software development methodologies to software defined networks (SDNs). To support our methodological guidelines, we propose an expressive requirement formalization language. Further, we describe a prototype tool able to check the compliance of an SDN controller with requirements expressed in the proposed language. Our evaluation of the prototype shows promising results on the practical viability of our approach.

Multipath TCP is a new extension to TCP that enables a host to transmit the packets from a given connection by using several interfaces. We propose mptcptrace, a software that enables a detailed analysis of Multipath TCP packet traces.

The use of real-time multimedia or mission-critical applications over IP networks puts strong pressure on service providers to operate disruption-free networks. However, after any topological change, link-state Interior Gateway Protocols (IGPs), such as IS-IS or OSPF, enter a convergence phase during which transient forwarding loops may occur. Such loops increase the network latency and cause packet losses. In this paper, we propose and evaluate an efficient algorithm aimed at avoiding such traffic disruptions without modifying these IGPs. In case of an intentional modification of the weight of a link (e.g., to shut it down for maintenance operations or to perform traffic engineering), our algorithm iteratively changes this weight, splitting the modification into a sequence of loop-free transitions. The number of weight increments that need to be applied on the link to reach its target state is minimized in order to remain usable in existing networks. Analysis performed on inferred and real Internet service provider (ISP) topologies shows that few weight increments are required to handle most link shutdown events (less than two intermediate metrics for more than 85% of the links). The evaluation of our implementation also reveals that these minimal sequences can be computed in a reasonable time.

Many scientific disciplines rely on "Experimental Design" to study various types of systems. Experimental design refers to a planned approach to experimentation that tries to provide statistical evidence to the outcome of experiments. The networking community rarely relies on such approaches, especially for real protocol implementations. Many improvements to protocols like TCP, including the recently proposed Multipath TCP, have been evaluated by considering a relatively limited set of simulations or experiments. Multipath TCP increases the goodput of a data transfer by simultaneously using multiple interfaces. It also improves load balancing thanks to dedicated congestion control. By applying experimental design, we conduct a large set of measurements inside Mininet with the Linux kernel Multipath TCP implementation, to measure its bandwidth aggregation and load balancing. Thanks to the experimental design approach, we are able to highlight several limitations of this implementation. We identify heuristics that lead to lower than expected performance and propose improvements.

Middleboxes such as firewalls, NAT, proxies, or Deep Packet Inspection play an increasingly important role in various types of IP networks, including enterprise and cellular networks. Recent studies have shed the light on their impact on real traffic and the complexity of managing them. Network operators and researchers have few tools to understand the impact of those boxes on any path. In this paper, we propose tracebox, an extension to the widely used traceroute tool, that is capable of detecting various types of middlebox interference over almost any path. tracebox sends IP packets containing TCP segments with different TTL values and analyses the packet encapsulated in the returned ICMP messages. Further, as recent routers quote, in the ICMP message, the entire IP packet that they received, tracebox is able to detect any modification performed by upstream middleboxes. In addition, tracebox can often pinpoint the network hop where the middlebox interference occurs. We evaluate tracebox with measurements performed on PlanetLab nodes. Our analysis reveals various types of middleboxes that were not expected on such an experimental testbed supposed to be connected to the Internet without any restriction.

Network upgrades, performance optimizations and traffic engineering activities often force network operators to adapt their IGP configuration. Recently, several techniques have been proposed to change an IGP configuration (e.g., link weights) in a disruption-free manner. Unfortunately, none of these techniques considers the impact of IGP changes on BGP correctness. In this paper, we show that known reconfiguration techniques can trigger various kinds of BGP anomalies. First, we illustrate the relevance of the problem by performing simulations on a Tier-1 network. Our simulations highlight that even a few link weight changes can produce long-lasting BGP anomalies affecting a significant part of the BGP routing table. Then, we study the problem of finding a reconfiguration ordering which maintains both IGP and BGP correctness. Unfortunately, we show examples in which such an ordering does not exist. Furthermore, we prove that deciding if such an ordering exists is NP-hard. Finally, we provide sufficient conditions and configuration guidelines that enable graceful operations for both IGP and BGP.

The network infrastructure of Internet service providers (ISPs) undergoes constant evolution. Whenever new requirements arise (e.g., the deployment of a new Point of Presence or a change in the business relationship with a neighboring ISP), operators need to change the configuration of the network. Due to the complexity of the Border Gateway Protocol (BGP) and the lack of methodologies and tools, maintaining service availability during reconfigurations that involve BGP is a challenge for operators. In this paper, we show that the current best practices to reconfigure BGP do not provide guarantees with respect to traffic disruptions. Then, we study the problem of finding an operational ordering of BGP reconfiguration steps that guarantees no packet loss. Unfortunately, finding such an operational ordering, when it exists, is computationally hard. To enable lossless reconfigurations, we propose a framework that extends current features of carrier-grade routers to run two BGP control planes in parallel. We present a prototype implementation and show the effectiveness of our framework through a case study.

Internal BGP (iBGP) is used to distribute interdomain routes within a single ISP. The interaction between iBGP and the underlying IGP can lead to routing and forwarding anomalies. For this reason, several research contributions aimed at defining sufficient conditions to guarantee anomaly-free configurations and providing design guidelines for network operators. In this paper, we show several anomalies caused by defective dissemination of routes in iBGP. We define the dissemination correctness property, which models the ability of routers to learn at least one route to each destination. By distinguishing between dissemination correctness and existing correctness properties, we show counterexamples that invalidate some results in the literature. Further, we prove that deciding whether an iBGP configuration is dissemination correct is computationally intractable. Even worse, determining whether the addition of a single iBGP session can adversely affect dissemination correctness of an iBGP configuration is also computationally intractable. Finally, we provide sufficient conditions that ensure dissemination correctness, and we leverage them to both formulate design guidelines and revisit prior results.

Network-wide migrations of a running network, such as the replacement of a routing protocol or the modification of its configuration, can improve the performance, scalability, manageability, and security of the entire network. However, such migrations are an important source of concerns for network operators as the reconfiguration campaign can lead to long, service-disrupting outages. In this paper, we propose a methodology that addresses the problem of seamlessly modifying the configuration of link-state Interior Gateway Protocols (IGPs). We illustrate the benefits of our methodology by considering several migration scenarios, including the addition and the removal of routing hierarchy in a running IGP, and the replacement of one IGP with another. We prove that a strict operational ordering can guarantee that the migration will not create any service outage. Although finding a safe ordering is NP-complete, we describe techniques that efficiently find such an ordering and evaluate them using several real-world and inferred ISP topologies. Finally, we describe the implementation of a provisioning system that automatically performs the migration by pushing the configurations on the routers in the appropriate order while monitoring the entire migration process.

Multipath routing strategies such as Equal-Cost MultiPath (ECMP) are widely used in IP and data-center networks. Most current methods to balance packets over the multiple next hops toward the destination base their decision on a hash computed over selected fields of the packet headers. Because of the non-invertible nature of hash functions, it is hard to determine the values of those fields so as to make the packet follow a specific path in the network. However, several applications might benefit from being able to choose such a path. Therefore, we propose a novel next-hop selection method based on an invertible function. By encoding the selection of successive routers into common fields of packet headers, the proposed method enables end hosts to force their packets to follow a specific path.

Nowadays Internet is ubiquitous resulting in an increasing path diversity and content duplication. However, while content can be retrieved from many different places, the paths to those places are not equivalent. Indeed, some paths offer better bandwidth while others are less expensive or more stable. In addition, a new range of applications is sensitive to the performance of the paths that carry their traffic. To support this evolution of the Internet, we propose ISP-Driven Informed Path Selection (IDIPS). Any ISP can easily deploy IDIPS to help its customers to select the paths that best meet their requirements in order to reach their content. IDIPS helps in this selection through pro-active measurements and ISP-defined policies. IDIPS is scalable and can support thousands of clients. IDIPS is also flexible and can thus be used by the ISP to optimize its routing decisions to take the performance of its inter-domain links into account.

Network-wide migrations of a running network, such as the replacement of a routing protocol or the modification of its configuration, can improve the performance, scalability, manageability, and security of the entire network. However, such migrations are an important source of concerns for network operators as the reconfiguration campaign can lead to long and service-affecting outages.

The Internet topology data collected through traceroute exploration has been extensively studied in the past. In particular, a remarkable property of the Internet, the power-law shape of node degree distribution, drew the attention of the research community. Several studies have since questioned this property.

It is becoming the norm for small mobile devices to have access to multiple technologies for connecting to the Internet. This gives researchers an increasing interest for solutions allowing to use efficiently several communication mediums. We propose a demonstration of our Multipath TCP implementation for Linux, that allows spreading a single TCP flow across multiple Internet paths, without requiring any change to applications. The demonstration will involve a real Internet communication with MPTCP, with simultaneous use of several paths, as well as a demonstration of MPTCP failover capability.

During the last years, some operators have expressed concerns about the continued growth of the BGP routing tables in the default-free zone. Several proposed solutions for this issue are centered around the idea of separating the network node's identifier from its topological location. Among the existing proposals, the Locator/ID Separation Protocol (LISP) has seen important development and implementation effort. LISP relies on a mapping system to provide bindings between locators and identifiers. The mapping system is a critical protocol component, and its design is still an open issue. In this paper we present a new mapping system: LISP-TREE. It is based on DNS and has a similar hierarchical topology: blocks of identifiers are assigned to the levels of the hierarchy by following the current IP address allocation policies. We also present measurement-driven simulations of mapping systems' performance, assuming a deployment of LISP in the current Internet.

Internet Service Providers design their network with resiliency in mind, having multiple paths towards external IP subnets available at the borders of their network. However, with the current internal Border Gateway Protocol, BGP routers and route reflectors only propagate their (unique) best path over their iBGP sessions. As a result, at the BGP router level, path diversity tends to be poor. Such lack of path diversity can lead to MED oscillations, prevents an efficient use of multipath BGP and does not allow for a fast and local recovery upon nexthop failure. Advertising multiple paths over iBGP sessions with BGP Add-Paths solves those issues, depending on the way the additional paths are selected. In this paper, we analyze the various options for the selection mode of the paths to be advertised. We show that these modes differently fulfill the needs of Add-Paths applications such as fast recovery upon failure and MED oscillation avoidance. We also show in our analysis that the costs and benefits bound with these modes depend on the connectivity of the AS where it is deployed. To support the analysis, we developed a tool allowing to measure the scaling of these modes in a given network. We illustrate the utilization of this tool on synthetic Internet topologies, and provide some recommendations for the choice of an Add-Paths selection mode.

In wireless roaming a mobile device obtains a service from some foreign network while being registered for the similar service at its own home network. However, recent proposals try to keep the service provider role behind the home network and let the foreign network create a tunnel connection through which all service requests of the mobile device are sent to and answered directly by the home network. Such Wireless Roaming via Tunnels (WRT) offers several (security) benefits but states also new security challenges on authentication and key establishment, as the goal is not only to protect the end-to-end communication between the tunnel peers but also the tunnel itself.

Redundant connectivity (or multiconnectivity) between adjacent autonomous systems (ASes) is important for interdomain traffic engineering and fast recovery in case of failures. However, the redundancy of ASes business relationship links has not been quantitatively studied, mainly due to the difficulty of obtaining relevant data. In this paper, we show that the mrinfo multicast monitoring tool can provide useful data about the Internet topology and such redundant links in particular. Our analysis relies on more than four years of daily queries to about ten thousand routers mapped into more than two hundred ASes. We demonstrate that peering links between ASes are frequently redundant. In particular, our analysis shows that more than half of the studied ASes pairs are connected through multiple physical links. We then refine our analysis by considering the different types of ASes and their business relationships. A particular result of our analysis is that at least 75% of the peer-to-peer relationships between adjacent Tier-1 ASes are redundant, i.e., the connections between these ASes involve several physical links. Our analysis is conservative, providing so a lower bound, as some links might not be seen by mrinfo due to ISPs filtering policies.

The growth of the Internet is causing some stress on the Internet control plane and BGP in particular. The IRTF Routing Research Group has been chartered to explore architectural alternatives to the current Internet routing architecture. Several of these alternatives rely on separating the identifier and locator roles of IP addresses. In this talk, wepsilall review the key principles of these proposals and explain the main challenges that need to be solved.

Very recent activities in the IETF and in the Routing Research Group (RRG) of the IRTG focus on defining a new Internet architecture, in order to solve scalability issues related to interdo-main routing. The approach that is being explored is based on the separation of the end-systems' addressing space (the identifiers) and the routing locators' space. This separation is meant to alleviate the routing burden of the Default Free Zone, but it implies the need of distributing and storing mappings between identifiers and locators on caches placed on routers. In this paper we evaluate the cost of maintaining these caches when the distribution mechanism is based on a pull model. Taking as a reference the LISP protocol, we base our evaluation on real Netflow traces collected on the border router of our campus network. We thoroughly analyze the impact of the locator/ID separation, and related cost, showing that there is a trade-off between the dynamism of the mapping distribution protocol, the demand in terms of bandwidth, and the size of the caches.

This paper describes and briefly evaluates a distributed and secure mechanism that allows IP addresses used as locators to be automatically distributed and assigned to routers inside an IP network. The routers then are responsible for the suballocation of these locators to their locally connected endsystems and customers.

During the past ten years, we have seen the emergence of a set of applications requiring more and more quality of service (QoS). For instance, IPTV needs large bandwidth and delays as lows as possible. Further, while previously a content was located in a single place, it is, nowadays, frequent that the content is replicated among a set of servers located anywhere on five continents or even among users themselves. Perfect examples of this are peer-to-peer (P2P) applications and FTP mirrors. In addition, multihoming, i.e., the ability of having different connections to Internet potentially through different providers, is becoming more and more popular [1, 2]. Finally, network level protocols such as SHIM6 or LISP must often choose the best path among a list of highly disparate paths according to traffic engineering or policies considerations (see discussions on IETF mailing lists). This situation leads to more complexity in ensuring QoS to the user. We have now to maximize the QoS perceived by the user based on a set of highly disparate paths. In addition, in many new applications, path selection and QoS are ensured at the application layer resulting in an ineffective network level resources use. Today, although many of measurements techniques have been developed within the IPPM working group of the IETF, an application that needs to select a path or a server must implement its own measurement system to obtain data to perform its selection. Thus, several applications running on the same host or in the same campus will probably perform almost the same kind of measurements. Duplicating those measurements is not the appropriate solution. This work was partially supported by the Europeanfounded

A few modifications to software and/or hardware of routers have been proposed recently to avoid the transient micro loops that can occur during the convergence of link-state interior gateway protocols like IS-IS and OSPF. We1 propose in this paper a technique that does not require modifications to ISIS and OSPF, and that can be applied now by ISPs. Roughly, in the case of a manual modification of the state of a link, we progressively change the metric associated with this link to reach the required modification by ensuring that each step of the progression will be loop-free. The number of changes that are applied to a link to reach the targeted state by ensuring the transient consistency of the forwarding inside the network is minimized. Analysis performed on real regional and tier-1 ISP topologies show that the number of required transient changes is small. The solution can be applied in the case of link metric updates, manual set up, and shut down of links.

This paper presents a solution aimed at avoiding losses of connectivity when an eBGP peering link is shut down by an operator for a maintenance. Currently, shutting down an eBGP session can lead to transient losses of connectivity even though alternate path are available at the borders of the network. This is very unfortunate as ISPs face more and more stringent service level agreements, and maintenance operations are predictable operations, so that there is time to adapt to the change and preserve the respect of the service level agreement.

We first show by measurements that BGP peering links fail as frequently as intradomain links and usually for short periods of time. We propose a new fast-reroute technique where routers are prepared to react quickly to interdomain link failures. For each of its interdomain links, each router precomputes a protection tunnel, i.e. an IP tunnel to an alternate nexthop which can reach the same destinations as via the protected link. We propose a BGP-based auto-discovery technique that allows each router to learn the candidate protection tunnels for its links. Each router selects the best protection tunnels for its links and when it detects an interdomain link failure, it immediately encapsulates the packets to send them through the protection tunnel. Our solution is applicable for the links between large transit ISPs and also for the links between multi-homed stub networks and their providers. Furthermore, we show that transient forwarding loops (and thus the corresponding packet losses) can be avoided during the routing convergence that follows the deactivation of a protection tunnel in BGP/MPLS VPNs and in IP networks using encapsulation.

No abstract available.

No abstract available.

When the topology of an IP network changes due to a link failure or a link metric modification, the routing tables of all the routers must be updated. Each of those updates may cause transient loops. In this paper, we prove that by ordering the updates of the routing tables on the routers, it is possible to avoid all transient loops during the convergence of ISIS or OSPF after a planned link failure, an unplanned failure of a protected link and after a link metric modification. We then propose a protocol that allows the routers to order the update of their routing tables to avoid transient loops without requiring any complex computation.

Very few measurements have been done on wide area ATM networks. In this paper, we first look at how a compliant traffic can be generated with standard ATM equipment. We show that in our environment, even with a 23.42 Mbps VP out of 34 Mbps links, it is necessary to use some spacing in the ATM adapters to generate a compliant traffic. Our measurements show that with a large window TCP is able to achieve a throughput of 19 Mbps when the ATM level traffic is compliant with the traffic contract. However, if the ATM level traffic is not entirely compliant the segment loss rate increases quickly, and the TCP throughput collapses. A 3% change in the cell rate of the ATM level traffic can cause the TCP throughput to drop from 19 Mbps down to 1 Mbps.