Wireless networks operating in the 60 GHz band have the potential to provide very high throughput but face a number of challenges (e.g., high attenuation, beam training, and coping with mobility) which are widely accepted but often not well understood in practice. Understanding these challenges, and especially their actual impact on consumer-grade hardware is fundamental to fully exploit the high physical layer rates in the 60 GHz band. To this end, we perform an extensive measurement campaign using two commercial off-the-shelf 60 GHz routers in real-world environments. Our results allow us to revisit a range of issues and provide much deeper insights into the reasons for specific performance compared to prior work on performance characterization. Further, our study goes beyond basic link characterization and explores for the first time practical considerations such as coverage and access point deployment. While some of our observations are expected, we also obtain highly surprising insights that challenge the prevailing wisdom in the community. We derive the shortcomings of current commercial 60 GHz devices, and the fundamental problems that remain open on the way to fast and efficient 60 GHz networking.

Next-generation wireless networks promise to provide extremely high data rates, especially exploiting the so-called millimeter-wave frequency range. Gaining information from spectrum usage is becoming important to provide smart adaptation capabilities to future network protocol stacks. Issues such as deafness, misaligned antennas, or blockage may severely impact network performance, and their identification is crucial. Despite the complexity of full analytical models, machine learning techniques are progressively being considered to improve spectrum usage at higher layers. In this paper, we design a signal processing technique that uses narrowband physical layer energy traces, obtained from one or multiple channel sniffers. The proposed technique utilizes a combination of template matching and an Explicit Duration Hidden Markov Model (EDHMM) to correctly classify frames, while coping with the non-stationarity of the traces. This leads to a protocol level monitor that does not need to decode the channel at the physical layer, but just infers the type of packets that are exchanged based on sub-sampled energy traces. The performance of this framework is evaluated using off-the-shelf mm-wave wireless devices, quantifying its detection performance in the presence of one or multiple sniffers, and assessing the impact of physical layer parameters such as noise power and signal levels.

Both IEEE 802.15.3c and 802.11ad use the 60 GHz band for high datarate Wireless Personal/Local Area Network applications. These millimeter-wave communications use very directional antennas since the small wavelength allows to integrate many small antenna elements to form a beamforming antenna array, enabling very high spatial reuse as can be found in dense indoor and IoT settings. However, earlier work shows that current mmWave systems are not as directional as theory would suggest, with significant interference that may prevent spatial reuse. In this work, we propose a centralized system that allows the network to carry out the beamtraining process not only to maximize signal power, but also taking into account other stations in order to minimize interference. This system is designed to work with unmodified clients. We implement and validate our system on commercial off-the-shelf 60 GHz hardware, achieving an average throughput gain of 24.67% for TCP traffic, and up to a twofold throughput gain in specific cases.

Understanding the gap between theory and practice of 60 GHz networks is crucial. Theoretical models typically focus on the physical layer only and assume ideal hardware components. However, such assumptions do not hold in practice. For instance, the cost-effective design of commercial off-the-shelf products often results in hardware impairments, such as highly irregular beam shapes. Further, mobility and frame aggregation have a strong impact throughout the protocol stack. Taking these effects into account when designing 60 GHz networking mechanisms is critical. In this paper, we characterize the aforementioned effects and contribute a comprehensive dataset of real-world traces that capture them. Among other traces, our dataset includes measurements of actual beam patterns, the impact of the cases of the devices on propagation, exhaustive measurements on the degradation of link performance in case of mobility, and the behavior of aggregation at the physical layer. Our dataset is the most comprehensive study of this kind to date. As a result, our work allows researchers to take the aforementioned hardware impairments into account, which paves the way for comparable and realistic evaluations of 60 GHz networking mechanisms in the community.

The very large bandwidth available in the 60 GHz band allows, in principle, to design highly accurate positioning systems. Integrating such systems with standard protocols (e.g., IEEE 802.11ad) is crucial for the deployment of location-based services, but it is also challenging and limits the design choices. Another key problem is that consumer-grade 60 GHz hardware only provides coarse channel state information, and has highly irregular beam shapes due to its cost-efficient design. In this paper, we explore the location accuracy that can be achieved using such hardware, without modifying the 802.11ad standard. We consider a typical 802.11ad indoor network with multiple access points (APs). Each AP collects the coarse signal-to-noise ratio of the directional beacons that clients transmit periodically. Given the irregular beam shapes, the challenge is to relate each beacon to a set of transmission angles that allows to triangulate a user. We design a location system based on particle filters along with linear programming and Fourier analysis. We implement and evaluate our algorithm on commercial off-the-shelf 802.11ad hardware in an office scenario with mobile human blockage. Despite the strong limitations of the hardware, our system operates in real-time and achieves sub-meter accuracy in 70% of the cases.

Beamforming is vital to overcome the high attenuation in wireless millimeter-wave networks. It enables nodes to steer their antennas in the direction of communication. To cope with complexity and overhead, the IEEE 802.11ad standard uses a sector codebook with distinct steering directions. In current off-the-shelf devices, we find codebooks with generic pre-defined beam patterns. While this approach is simple and robust, the antenna modules that are typically deployed in such devices are capable of generating much more precise antenna beams. In this paper, we adaptively adjust the sector codebook of IEEE 802.11ad devices to optimize the transmit beam patterns for the current channel. To achieve this, we propose a mechanism to extract full channel state information (CSI) regarding phase and magnitude from coarse signal strength readings on off-the-shelf IEEE 802.11ad devices. Since such devices do not expose the CSI directly, we generate a codebook with phase-shifted probing beams that enables us to obtain the CSI by combining strategically selected magnitude measurements. Using this CSI, transmitters dynamically compute a transmit beam pattern that maximizes the signal strength at the receiver. Thereby, we automatically exploit reflectors in the environment and improve the received signal quality. Our implementation of this mechanism on off-the-shelf devices demonstrates that adaptive codebook optimization achieves a significantly higher throughput of about a factor of two in typical real-world scenarios.

Wireless networks operating in the 60 GHz band have the potential to provide very high throughput but face a number of challenges (e.g., high attenuation, beam training, and coping with mobility) which are widely accepted but often not well understood in practice. Understanding these challenges, and especially their actual impact on consumer-grade hardware is fundamental to fully exploit the high physical layer rates in the 60 GHz band. To this end, we perform an extensive measurement campaign using two commercial off-the-shelf 60 GHz routers in practical real-world environments. Our study is centered around two fundamental adaptation mechanisms in 60 GHz networks-beam training and rate control- whose interactions are key for performance. Understanding these interactions allows us to revisit a range of issues and provide much deeper insights into the reasons for specific performance compared to prior work on performance characterization. Further, our study goes beyond basic link characterization and explores for the first time practical considerations such as coverage and access point deployment. While some of our observations are expected, we also obtain highly surprising insights that challenge the prevailing wisdom in the community.

The use of directional antennas in millimeter-wave communication promises high spatial reuse at multi-gigabit-per-second data rates in dense wireless networks. Existing work studies such networks using commercial hardware but is limited to individual links. Moreover, such hardware typically allows for little or no control of the lower layers of the protocol stack. In this paper, We study the performance of dense millimeterwave deployments featuring up to eight stations. To this end, we use a practical IEEE 802.11ad millimeter-wave testbed that allows access to the lower layer parameters of each station. This enables us to analyze the impact of these parameters on upper layer performance. We study, for first time to our best knowledge, issues such as the impact of channel contention on the buffer size at the transport layer, the effect of frame aggregation, and the efficiency of spatial sharing. Our results show that using large buffer sizes with TCP is harmful due to channel contention despite the multi-gigabit-per-second data rates. Further, frame aggregation is only beneficial up to a certain level due to higher error rates for large frames. Finally, we also study delay, showing that the regular beacon transmission time can degrade performance.

Millimeter-wave devices must use highly directional antennas to achieve GBit/s data rates over reasonable distances due to the high path loss. As a consequence, it is important to precisely align the antenna beams between sender and receiver. Even minor movement or rotation of a device can result in beam misalignment and thus a strong performance degradation. Existing work as well as standards such as IEEE 802.11ad tackle this issue by means of antenna sector probing. This comes at the expense of a significant overhead, which may significantly reduce the performance of millimeter-wave communication, particularly in mobile scenarios. In this paper, we present a mechanism that can track both movement and rotation of 60 GHz mobile devices with zero overhead. To this end, we transmit part of the preamble of each packet using a multi-lobe beampattern. Our approach does not require any additional control messages and is backward compatible with 802.11ad. We implement our scheme on a 60 GHz testbed using phased antenna arrays, and show that we reduce the angle error to less than 5° in most cases. We also perform simulations to validate our approach in a wide range of scenarios, achieving up to 2x throughput gain.

This work conducts one of the first experimental studies of Multipath TCP (MPTCP) in dual-band 60 GHz/5 GHz WLANs using off-the-shelf hardware. We consider both uncoupled and different coupled congestion control algorithms, compare their performance and their potential to improve throughput over single path TCP, and uncover their limitations. In contrast to a recent study that reports reduced throughput with MPTCP compared to single path TCP over 60 GHz, our results show that significant performance improvements are possible, especially in the case of uncoupled congestion control. On the other hand, performance gains with coupled congestion control are lower as these algorithms often fail to fully utilize the capacity of both paths simultaneously. We also observe a pathological case that can lead to significantly reduced throughput with MPTCP regardless of the congestion control algorithm.

The directional nature of communication in millimeter-wave bands suggests that the frequency selectivity of the channel may be limited due to the absence of reflected paths. However, our measurement studies show that reflections often cause significant frequency selectivity in practical scenarios, which we can exploit to increase the otherwise limited range of 60 GHz networks. Specifically, we measure real-world indoor 60 GHz channels with a bandwidth of 2 GHz, and study their behavior with respect to techniques such as bitloading, subcarrier switch-off, and waterfilling. To this end, we consider Orthogonal Frequency-Division Multiplexing (OFDM) as defined in the IEEE 802.11ad standard and show that in contrast to common belief, these techniques are highly beneficial in millimeter-wave networks. We analyze this in practice for both horn antennas as well as an electronically-steerable phased antenna array. Most importantly, our practical results demonstrate that for the specific case of the 60 GHz band, this selectivity allows for a range extension of up to 50%. Hence, our approach enables us to alleviate one of the main limitations of millimeter-wave networks.

The lower layer characteristics of 60 GHz wireless networks impact the entire protocol stack. This includes effects such as sudden link failure due to blockage, high packet loss due to deafness, or suboptimal transmit rates due to antenna misalignment. To understand the resulting behaviors at the upper layers, it is key to have access to lower layer information. At the time of writing, 60 GHz experimentation hardware is either limited in terms of bandwidth, real-time capability, or does not support a full stack. In contrast, commercial off-the-shelf (COTS) hardware covers all of the above features but is a black box regarding the lower layers. Our demo exploits hidden monitoring capabilities of a certain COTS device which has become popular in the 60 GHz research community. The demo allows conference attendees to interact with 60 GHz links and observe how this impacts lower layer parameters. Moreover, the demo shows how such interactions relate to the performance at the upper layers.

The millimeter-wave technology is emerging as an alternative to legacy 2.4/5 GHz WiFi, offering multi-Gigabit throughput. While a lot of attention has been paid recently to analyzing the performance of the 60 GHz technology and adapting it for indoor WLAN usage, the power consumption aspect has largely been neglected. Given that mobile devices are the next target for 60 GHz, any discussion about this technology is incomplete without considering power consumption. In this work, we present the first, to our best knowledge, detailed study of the power consumption of 60 GHz commodity devices. We evaluate the power and energy consumption of two standard-compliant 60 GHz wireless adapters in different operating states and under a number of different configurations. We also compare our results against 802.11ac and discuss power-performance tradeoffs for the two technologies.

Troubleshooting millimeter-wave (mm-wave) wireless networks is complex due to the directionality of the communication. Issues such as deafness, misaligned antennas, or blockage may severely impact network performance, and identifying them is crucial to improve network deployments. To this end, access to lower-layer information is important. However, commercial off-the-shelf mm-wave wireless devices typically do not provide such information. Even if they would, detecting effects such as deafness based on information of a single node that forms part of the network is typically hard. In this paper, we present the design and evaluation of an external sniffing device that can infer the aforementioned performance issues only using narrowband physical layer energy traces. Our sniffer does not need to decode any data, resulting in a simple but effective approach which also preserves privacy and works on encrypted networks. Our key contribution is a machine learning framework which enables automated energy trace analysis while coping with the non-stationarity of the traces. We evaluate its performance in practice using off-the-shelf wireless devices operating in the 60 GHz band. Our results show that the above framework correctly infers physical layer events in virtually all cases, thus providing valuable information to troubleshoot issues in mm-wave networks.

After being widely studied in theory, physical layer security schemes are getting closer to enter the consumer market. Still, a thorough practical analysis of their resilience against attacks is missing. In this work, we use software-defined radios to implement such a physical layer security scheme, namely, orthogonal blinding. To this end, we use orthogonal frequency-division multiplexing (OFDM) as a physical layer, similarly to WiFi. In orthogonal blinding, a multi-antenna transmitter overlays the data it transmits with noise in such a way that every node except the intended receiver is disturbed by the noise. Still, our known-plaintext attack can extract the data signal at an eavesdropper by means of an adaptive filter trained using a few known data symbols. Our demonstrator illustrates the iterative training process at the symbol level, thus showing the practicability of the attack.

The impact of frame aggregation on wireless network performance increases dramatically with higher data rates. The key problem is that the transmission time of packets decreases while the medium access, preamble and packet header overhead remain the same. Recent 802.11 standards address this issue using frame aggregation, i.e., grouping multiple data frames in a single transmission to reduce the overhead. This already provides substantial efficiency gains in networks operating in the 2.4 GHz and 5 GHz bands, and for future 60 GHz networks such as 802.11ad, gains are even more pronounced due to the order-of-magnitude higher data rates. In 802.11ad, frame aggregation becomes crucial to achieve the multi-gbps data rates that are possible in theory, since medium access overhead can be 20x larger than the time required to transmit a single packet. While frame aggregation is essential, it very much depends on the traffic patterns present in the wireless network, and a node may not always have enough packets in the transmit queue to achieve a sufficiently large aggregated frame size. In this paper, we investigate in which case nodes should wait to construct a larger aggregated packet before starting the channel access procedure. We present a simple waiting policy for the uplink case that either waits for a minimum number of packets or for a maximum amount of time, whichever comes first. For the downlink case, we utilize a maximum weight scheduling policy with a maximum waiting time. Our results show that both policies significantly improve medium utilization, thus increasing throughput and reducing end-to-end delay.

The performance of current consumer-grade devices for 60 GHz wireless networks is limited. While such networks promise both high data rates and uncomplicated spatial reuse, we find that commercially available devices based on the WiHD and WiGig standards may suffer from their cost-effective design. Very similar mechanisms are used in upcoming devices based on the IEEE 802.11ad standard. Hence, understanding them well is crucial to improve the efficiency and performance of next generation millimeter wave networks. In this paper, we present the first in-depth beamforming, interference, and frame level protocol analysis of off-the-shelf millimeter wave systems with phased antenna arrays. We focus on (a) the interference due to the lack of directionality of consumer-grade antennas, and (b) the degree of data aggregation of current devices. Regarding (a), our beam pattern measurements show strong side lobes that challenge the common conception of high spatial reuse in 60 GHz networks. We also show that reflections in realistic settings worsen this effect. Further, we measure weak directionality when beamforming towards the boundary of the transmission area of an antenna array. Regarding (b), we observe that devices only aggregate data if connections require high bandwidth, thus increasing medium usage time otherwise.

Geographic routing yields scalable operation in Wireless Multihop Networks. Nodes forward packets to the destination either in a greedy manner, or along arbitrary curves from source to destination. The latter allows for higher robustness, as paths are adjusted to avoid, for instance, congested areas or non-cooperative nodes. However, determining working curves is either based on unreliable trial-and-error approaches or requires global network knowledge. We propose two local schemes that dynamically adapt the curve defining the packet forwarding at node level. Our first procedure uses cached curves, and the second one builds local maps at each node. We implement them for the case of planar graph routing with guaranteed delivery but they can also be easily adapted to greedy routing. We evaluate our schemes by means of simulation in constrained multihop networks, and show that they reduce transmissions up to 74% compared to existing solutions.

Corridor-based Routing enables advanced physical layer schemes in Wireless Multihop Networks (WMNs). It widens paths in order to span multiple nodes per hop. As a result, groups of nodes cooperate locally at each hop to forward packets. Recent theoretical work suggests using Orthogonal Frequency-Division Multiple Access (OFDMA) in combination with Corridor-based Routing to improve throughput in WMNs. However, results focus on achievable capacity and do not consider practical issues such as modulation and coding schemes. In this paper, we study OFDMA for corridors in practice and implement it on software-defined radios. We show that OFDMA corridors provide significantly larger throughput gains when considering realistic modulation and coding, achieving up to 2x throughput gain compared to traditional routing not based on corridors.

Corridor-based Routing widens traditional hop-by-hop paths to enable advanced physical layer mechanisms in Wireless Multihop Networks. The key concept is to let groups of nodes cooperate to jointly forward data. Instead of establishing a path formed by a fixed sequence of nodes, the network layer builds a "corridor" from source to destination to allow for such an approach. The corridor is divided into stages, which consist of the aforementioned groups of nodes. Existing mechanisms for Corridor-based Routing focus on the physical layer and assume a routing protocol that builds the corridor. In this paper, we design a corridor construction algorithm and a practical protocol that implements it. In particular, we address in detail the overhead introduced by our protocol, since this is crucial for performance and an open issue of Corridor-based Routing. We implement our approach both in simulation and practice. We obtain the turning point at which corridors become profitable and show that our protocol builds corridors, which enable throughput gains up to 74%.

Today, the performance of routing mechanisms in Wireless Multihop Networks (WMNs) is still limited by the lower layers. While recent cross-layer approaches take advantage of the characteristics of the medium, they are often based on traditional physical layers such as OFDM. State-of-the-art techniques used in one-hop scenarios, such as OFDMA or MIMO, pose a significant challenge in practical multihop networks, since typically Channel State Information (CSI) at the transmitter is required. Due to its volatile nature, disseminating timely CSI in the network is often infeasible. We generalize Corridor-based Routing to enable advanced physical layers in WMNs. Instead of routing packets from node to node, we forward them along fully-connected groups of nodes. As a result (1) CSI only needs to be exchanged locally to enable cooperation in a group, and (2) groups can adaptively choose the best physical layer technique according to CSI. We investigate the benefits of Corridor-based Routing and present a protocol design that enables operation of corridors in WMNs.

Physical layer techniques enable new security paradigms in wireless networks, such as node authentication based on channel characteristics. However, current approaches are often limited to one-hop scenarios. We investigate how the aforementioned authentication schemes can be extended to the multihop case. In particular, we design mechanisms to provide path signatures, which allow destination nodes to verify the path followed by packets in the network. This enables, for instance, detection of packets not being forwarded along the intended route directly at the physical layer and without the need for per-packet cryptographic schemes. We propose multiple approaches based on the accumulative effect of link CSI along multihop routes. We implement and evaluate them on a software-defined radio testbed. Our preliminary results show that we can successfully identify unauthorized nodes participating in the forwarding of packets up to a path length of six hops.

Interference alignment (IA) is often considered in the spatial domain in combination with MIMO systems. In contrast, aligning interference in the frequency domain among multiple subcarriers can also benefit single-antenna OFDM-based access networks. It allows for flexible operation on a per-subcarrier basis. We investigate the gains achievable by frequency IA in practice for a scenario with multiple access points and clients. Previous work is predominantly theoretical and focuses on idealized cases where all nodes have the same average signal-to-noise ratio (SNR). On the contrary, in practical networks, nodes typically have heterogeneous SNRs depending on channel conditions, which might have a significant impact on IA performance. We tackle this problem by designing mechanisms that adaptively choose which nodes shall perform IA on which subcarriers depending on current channel conditions. We implement and validate our approach on software-defined radios. To the best of our knowledge, this is the first practical implementation of IA in the frequency domain. Our measurements show that (1) frequency IA is feasible in practice, and (2) choosing appropriate nodes and subcarriers overcomes the main limitations due to heterogeneous SNRs. Our mechanisms enable IA in scenarios where it would be infeasible otherwise, achieving throughput gains close to the 33% theoretical maximum.

State-of-the-art wireless technologies such as LTE employ OFDMA to enable robust and efficient frequency-division multiple access, which is also promising for other scenarios, such as Wireless Mesh Networks. OFDMA can improve throughput by allocating subchannels to the users which experience the best performance on the corresponding frequencies. This has lead to a large body of work on subchannel allocation schemes which optimize throughput while guaranteeing fairness. However, most solutions are discussed at a purely theoretical and/or simulative level. In this paper, we provide a framework for prototyping such schemes on the WARP Software Defined Radio to obtain practical results with small implementation overhead. We design a mechanism which allows to analyze offline how a certain allocation technique would have performed in a previously measured scenario. Moreover, we show that our approach is stable and robust regarding practical issues. While we focus on Wireless Mesh Networks, our system can easily be adapted to infrastructure-based systems.

In multi-hop networks, conventional forwarding along a unicast route forces the data transmission to follow a fixed sequence of nodes. In previous works, it has been shown that widening this path to create a corridor of forwarding nodes and applying OFDMA to split and merge the data as it travels through the corridor towards the destination node leads to considerable gains in achievable throughput compared to the case forwarding data along a unicast route. However, the problem of selecting potential nodes to act as forwarding nodes within the corridor has not been addressed in the literature, as in general a rather homogeneous network topology with equally spaced relay clusters per hop between source and destination node has been assumed. In this paper, a more realistic heterogeneous network is considered where the nodes in the area between source and destination are randomly distributed instead of being clustered with equal distance. A node selection scheme is presented which selects the forwarding nodes within the corridor based on a given unicast route between source and destination node. In simulations, it is shown that with the proposed node selection scheme, considerable throughput gains of up to 50 % compared to forwarding along unicast route can be achieved applying corridor-based routing in heterogeneous networks especially in sparse networks.

State-of-the-art physical layers such as OFDMA are widely used in infrastructure-based networks to enhance efficiency in one-to-many transmissions. Application to wireless mesh networks is highly promising, as the diversity increases in many-to-many scenarios. While theoretical work on OFDMA for this scenario exists, it has not yet been implemented in practice. In this demonstration, we show practical measurements of OFDMA in a topology with multiple transmitters and receivers, which represents a fully-connected segment of a mesh network. Moreover, we model our system analytically and in simulation. Our measurements show the validity of these models. We demonstrate over 90% reduction of the symbol error rate and a 29% channel capacity increase.

In multi-hop networks, conventional unipath routing approaches force the data transmission to follow a fixed sequence of nodes. In this paper, we widen this path to create a corridor of forwarding nodes. Within this corridor, data can be split and joined at different nodes as the data travels through the corridor towards the destination node. To split data, decode-and-forward OFDMA is used since with OFDMA, one can exploit the benefits of opportunistically allocating different subcarriers to different nodes according to their channel conditions. To avoid interference, each subcarrier is only allocated once per hop. For the presented scheme, the problem of optimizing the network throughput by means of resource and power allocation is formulated and two suboptimal algorithms are proposed to solve this problem with feasible effort. Simulations show that in multi-hop networks corridor-based routing using opportunistic forwarding outperforms conventional unipath routing approaches in terms of achievable throughput.

Localized geographic routing schemes operating on planar graphs promise scalability for use within large multihop wireless networks. Existing schemes base routing path construction on faces defined by the planar graph. Once running on a particular planar graph, none of the existing schemes is flexible enough to adapt the sequence of faces visited by the constructed path. Thus, real-world constraints such as network congestion, limited node energy levels, or non-cooperation of nodes might severely impact the performance and the robustness of existing planar graph routing variants. To address this problem, we extend planar graph routing with one further degree of freedom: control over the sequence of visited faces. Basically, our face routing extension now follows a sequence of faces intersected by any curve we can freely adjust. We investigate basic schemes for choosing curves dealing with imperfections in the network, and derive algorithms for routing and forwarding along these curves. We analytically prove that our scheme is loop free and allows for guaranteed delivery in arbitrary planar connected graphs. We implement curve-based routing and show its feasibility by means of a simulation study. As a proof-of-concept scenario, we investigate the case of non-cooperating nodes. Our results show that curve-based routing is able to sustain the delivery of packets where traditional schemes fail.

Scalability of routing algorithms is a critical issue in large multihop wireless networks. In this sense, approaches like localized geographic routing are very promising. Existing schemes base routing path construction on faces defined by the planar graph of the network. Once running on a particular planar graph, none of the existing schemes is flexible enough to adapt the sequence of faces visited by the constructed path. To address this problem, we extend planar graph routing with one further degree of freedom: control over the sequence of visited faces. Basically, our face routing extension now follows a sequence of faces intersected by any curve we can freely adjust. We motivate our work by discussing application scenarios that benefit from our scheme and suggest basic mechanisms for choosing appropriate curves. We further present preliminary results from an implementation of our curve-based routing scheme.