Mobile network operators collect a humongous amount of network measurements. Among those, sector Key Performance Indicators (KPIs) are used to monitor the radio access, i.e., the "last mile" of mobile networks. Thresholding mechanisms and synthetic combinations of KPIs are used to assess the network health, and rank sectors to identify the underperforming ones. It follows that the available monitoring methodologies heavily rely on the fine grained tuning of thresholds and weights, currently established through domain knowledge of both vendors and operators. In this paper, we study how to bridge sector KPIs to reflect Quality of Experience (QoE) groundtruth measurements, namely throughput, latency and video streaming stall events. We leverage one month of data collected in the operational network of mobile network operator serving more than 10 million subscribers. We extensively investigate up to which extent adopted methodologies efficiently capture QoE. Moreover, we challenge the current state of the art by presenting data-driven approaches based on Particle Swarm Optimization (PSO) metaheuristics and random forest regression algorithms, to better assess sector performance. Results show that the proposed methodologies outperforms state of the art solution improving the correlation with respect to the baseline by a factor of 3, and improving visibility on underperforming sectors. Our work opens new areas for research in monitoring solutions for enriching the quality and accuracy of the network performance indicators collected at the network edge.

To manage and maintain large-scale cellular networks, operators need to know which sectors underperform at any given time. For this purpose, they use the so-called hot spot score, which is the result of a combination of multiple network measurements and reflects the instantaneous overall performance of individual sectors. While operators have a good understanding of the current performance of a network and its overall trend, forecasting the performance of each sector over time is a challenging task, as it is affected by both regular and non-regular events, triggered by human behavior and hardware failures. In this paper, we study the spatio-temporal patterns of the hot spot score and uncover its regularities. Based on our observations, we then explore the possibility to use recent measurements' history to predict future hot spots. To this end, we consider tree-based machine learning models, and study their performance as a function of time, amount of past data, and prediction horizon. Our results indicate that, compared to the best baseline, tree-based models can deliver up to 14% better forecasts for regular hot spots and 153% better forecasts for non-regular hot spots. The latter brings strong evidence that, for moderate horizons, forecasts can be made even for sectors exhibiting isolated, non-regular behavior. Overall, our work provides insight into the dynamics of cellular sectors and their predictability. It also paves the way for more proactive network operations with greater forecasting horizons.

Tremendous effort has gone into the ongoing battle to make webpages load faster. This effort has culminated in new protocols (QUIC, SPDY, and HTTP/2) as well as novel content delivery mechanisms. In addition, companies like Google and SpeedCurve investigated how to measure "page load time" (PLT) in a way that captures human perception. In this paper we present Eyeorg [12], a platform for crowdsourcing web quality of experience measurements. Eyeorg overcomes the scaling and automation challenges of recruiting users and collecting consistent user-perceived quality measurements. We validate Eyeorg's capabilities via a set of 100 trusted participants. Next, we showcase its functionalities via three measurement campaigns, each involving 1,000 paid participants, to 1) study the quality of several PLT metrics, 2) compare HTTP/1.1 and HTTP/2 performance, and 3) assess the impact of online advertisements and ad blockers on user experience. We find that commonly used, and even novel and sophisticated PLT metrics fail to represent actual human perception of PLT, that the performance gains from HTTP/2 are imperceivable in some circumstances, and that not all ad blockers are created equal.

Video streaming on mobile devices is prone to a multitude of faults and although well established video Quality of Experience (QoE) metrics such as stall frequency are a good indicator of the problems perceived by the user, they do not provide any insights about the nature of the problem nor where it has occurred. Quantifying the correlation between the aforementioned faults and the users' experience is a challenging task due the large number of variables and the numerous points-of-failure.

Online Behavioral Advertising (OBA) is an important revenue source for online publishers and content providers. However, the extensive user tracking required to enable OBA raises valid privacy concerns. Existing and proposed solutions either block all tracking, therefore breaking OBA entirely, or require significant changes on the current advertising infrastructure, making adoption hard. We propose Web Identity Translator (WIT), a new privacy service running as a proxy or middlebox. WIT stops the original tracking cookies from being set on the browser of users and instead substitutes them by private cookies it controls. Manipulating the mapping between tracking and private cookies WIT maintains permits transparent OBA to continue while simultaneously protecting the identity of users from attacks based on behavioral analysis of browsing patterns.

A significant fraction of Internet traffic is now encrypted and HTTPS will likely be the default in HTTP/2. However, Transport Layer Security (TLS), the standard protocol for encryption in the Internet, assumes that all functionality resides at the endpoints, making it impossible to use in-network services that optimize network resource usage, improve user experience, and protect clients and servers from security threats. Re-introducing in-network functionality into TLS sessions today is done through hacks, often weakening overall security.

Distributing long-tail content is a difficult task due to the low amortization of bandwidth transfer costs as such content has limited number of views. Two recent trends are making this problem harder. First, the increasing popularity of user-generated content and online social networks create and reinforce such popularity distributions. Second, the recent trend of geo-replicating content across multiple points of presence spread around the world, done for improving quality of experience (QoE) for users. In this paper, we analyze and explore the tradeoff involving the “freshness” of the information available to the users and WAN bandwidth costs, and we propose ways to reduce the latter through smart update propagation scheduling, by leveraging on the knowledge of the mapping between social relationships and geographic location, the timing regularities and time differences in end user activity. We first assess the potential of our approach by implementing a simple social-aware scheduling algorithm that operates under bandwidth budget constraints and by quantifying its benefits through a trace-driven analysis. We show that it can reduce WAN traffic by up to 55 percent compared to an immediate update of all replicas, with a minimal effect on information freshness and latency. Second, we build TailGate, a practical system that implements our social-aware scheduling approach, which distributes on the fly long-tail content across PoPs at reduced bandwidth costs by flattening the traffic. We evaluate TailGate by using traces from an OSN and show that it can decrease WAN bandwidth costs by as much as 80 percent and improve QoE. We deploy TailGate on PlanetLab and show that even in the case when imprecise social information is available, it can still decrease by a factor of 2 the latency for accessing long-tail YouTube videos.

Through their normal operation, cellular networks are a repository of continuous location information from their subscribed devices. Such information, however, comes at a coarse granularity both in terms of space, as well as time. For otherwise inactive devices, location information can be obtained at the granularity of the associated cellular sector, and at infrequent points in time, that are sensitive to the structure of the network itself, and the level of mobility of the device. In this paper, we are asking the question of whether such sparse information can help to identify the paths followed by mobile connected devices throughout the day. If such a task is possible, then we would not only enable continuous mobility path estimation for smartphones, but also for the millions of future connected "things".

Increased user concern over security and privacy on the Internet has led to widespread adoption of HTTPS, the secure version of HTTP. HTTPS authenticates the communicating end points and provides confidentiality for the ensuing communication. However, as with any security solution, it does not come for free. HTTPS may introduce overhead in terms of infrastructure costs, communication latency, data usage, and energy consumption. Moreover, given the opaqueness of the encrypted communication, any in-network value added services requiring visibility into application layer content, such as caches and virus scanners, become ineffective.

Mobile phones in the 3G/4G era enable us to stay connected not only to the voice network, but also to online services like social networks. In this paper, we study the energy and network costs of mobile applications that provide continuous online presence (e.g. WhatsApp, Facebook, Skype). By combining measurements taken on the mobile and the cellular access network, we reveal a detailed picture of the mechanisms selected to implement online presence, along with their effect on handset energy consumption and network signaling traffic.

Content caching is a fundamental building block of the Internet. Caches are widely deployed at network edges to improve performance for end-users, and to reduce load on web servers and the backbone network. Considering mobile 3G/4G networks, however, the bottleneck is at the access link, where bandwidth is shared among all mobile terminals. As such, per-user capacity cannot grow to cope with the traffic demand. Unfortunately, caching policies would not reduce the load on the wireless link which would have to carry multiple copies of the same object that is being downloaded by multiple mobile terminals sharing the same access link.

The co-existence of cellular and wired networks has been exploited almost exclusively in the direction of OffLoading traffic from the former onto the latter. In this paper we claim that there exist cases that call for the exact opposite, i.e, use the cellular network to assist a fixed wired network. In particular, we show that by "OnLoading'' traffic from the wired broadband network onto the cellular network we can usefully speedup wired connections, on the downlink or the uplink. We consider the technological challenges pertaining to this idea and implement a prototype 3G OnLoading service that we call 3GOL, that can be deployed by an operator providing both the wired and cellular network services. By strategically OnLoading a fraction of the data transfers to the 3G network, one can significantly enhance the performance of particular applications. In particular we demonstrate non-trivial performance benefits of 3GOL to two widely used applications: video-on-demand and multimedia upload. We also consider the case when the operator that provides wired and cellular services is different, adding the analysis on economic constraints and volume cap on cellular data plans that need to be respected. Simulating 3GOL}over a DSLAM trace we show that 3GOL can reduce video pre-buffering time by at least 20% for 50% of the users while respecting data caps and we design a simple estimator to compute the daily allowance that can be used towards 3GOL while respecting caps. Our prototype is currently being piloted in 30 households in a large European city by a large network provider.

The large-scale collection and exploitation of personal information to drive targeted online advertisements has raised privacy concerns. As a step towards understanding these concerns, we study the relationship between how much information is collected and how valuable it is for advertising. We use HTTP traces consisting of millions of users to aid our study and also present the first comparative study between aggregators. We develop a simple model that captures the various parameters of today's advertising revenues, whose values are estimated via the traces. Our results show that per aggregator revenue is skewed (5% accounting for 90% of revenues), while the contribution of users to advertising revenue is much less skewed (20% accounting for 80% of revenue). Google is dominant in terms of revenue and reach (presence on 80% of publishers). We also show that if all 5% of the top users in terms of revenue were to install privacy protection, with no corresponding reaction from the publishers, then the revenue can drop by 30%.

The popularity of smartphones, cloud computing, and the app store model have led to cellular networks being used in a completely different way than what they were designed for. As a consequence, mobile applications impose new challenges in the design and efficient configuration of constrained networks to maximize application's performance. Such difficulties are largely caused by the lack of cross-layer under- standing of interactions between different entities -applications, devices, the network and its management plane. In this paper, we describe RILAnalyzer, an open-source tool that provides mechanisms to perform network analysis from within a mobile device. RILAnalyzer is capable of recording low-level radio information and accurate cellular net- work control-plane data, as well as user-plane data. We demonstrate how such data can be used to identify previously overlooked issues. Through a small user study across four cellular network providers in two European countries we infer how different network configurations are in reality and explore how such configurations interact with application logic, causing network and energy overheads.

Voice and data traffic growth over the last several years has become a major challenge for cellular operators with a direct impact on revenues, infrastructure investments, and end-user performance. The economics of these operators depend on various incentives used to attract users in the form of unlimited, buffet-like voice/sms/data packages. However, our understanding of the effects of user behavior under these offerings on operator revenues/costs remains poor. Using two years of detailed usage information of ~1 million users across three services, voice, sms and data, combined with payment and cost information, we study how user behavior affects the economics of cellular operators. We discover that around 20% of the users consume more resources than what they pay for and hence are non-profitable. In addition to the individual user behavior, we study how the user interactions in the call graph affect the operator's revenues and cost, drawing on tools from social network analysis. We develop a framework that incorporates both the individual and social user behavior for studying how volume caps influence the revenues and the traffic costs. Using this framework we empirically show that volume caps can increase the difference between the revenues and the traffic costs of the studied operator by a factor of 2, while affecting only 16% of the existing user base.

The wide channels feature combines two adjacent channels to form a new, wider channel to facilitate high-data-rate transmissions in multiple-input-multiple-output (MIMO)-based IEEE 802.11n networks. Using a wider channel can exacerbate interference effects. Furthermore, contrary to what has been reported by prior studies, we find that wide channels do not always provide benefits in isolation (i.e., one link without interference) and can even degrade performance. We conduct an in-depth, experimental study to understand the implications of wide channels on throughput performance. Based on our measurements, we design an auto-configuration framework called ACORN for enterprise 802.11n WLANs. ACORN integrates the functions of user association and channel allocation since our study reveals that they are tightly coupled when wide channels are used. We show that the channel allocation problem with the constraints of wide channels is NP-complete. Thus, ACORN uses an algorithm that provides a worst-case approximation ratio of O(1/Δ + 1), with Δ being the maximum node degree in the network. We implement ACORN on our 802.11n testbed. Our evaluations show that ACORN: 1) outperforms previous approaches that are agnostic to wide channels constraints; it provides per-AP throughput gains ranging from 1.5 × 6×; and 2) in practice, its channel allocation module achieves an approximation ratio much better than the theoretically predicted O(1/Δ + 1).

The assumption of rationality is fundamental to large part of network economics literature. In this paper, we use a simple definition of rationality based on economic self-interest and test for such behavior using real data on how users purchase and consume mobile network services. If users acted in their best (optimal) interest, then they would opt for the tariff that best suits their demands. However, that need not be the case, as users can fall prey to biases that can lead them to make seemingly sub-optimal choices. Such biases are hard to characterize and in this paper we empirically study how end-users purchase and use network services.

Access link can often be the bottleneck for application performance. In this paper, we propose to augment wired connections using cellular ones, that we term "3G onloading (3GOL)". 3GOL utilizes available mobile devices and already-paid-for data volumes to augment and improve performance of applications on wired network. We motivate 3GOL by understanding bottlenecks present in the wired and the cellular networks. In order to understand the potential benefits of 3GOL, we conduct active experiments using mobile devices. We show that capacity gains can scale linearly with the number of devices on the downlink while also seeing improvements on the uplink. Using real traces we show how video on demand can benefit with 3GOL, even when volume caps are in place. We design 3GOL as an over the top service, and highlight research challenges.

Mobile phones and tablets can be considered as the first incarnation of the post-PC era. Their explosive adoption rate has been driven by a number of factors, with the most signifcant influence being applications (apps) and app markets. Individuals and organizations are able to develop and publish apps, and the most popular form of monetization is mobile advertising.

Distributing long-tail content is an inherently difficult task due to the low amortization of bandwidth transfer costs as such content has limited number of views. Two recent trends are making this problem harder. First, the increasing popularity of user-generated content (UGC) and online social networks (OSNs) create and reinforce such popularity distributions. Second, the recent trend of geo-replicating content across multiple PoPs spread around the world, done for improving quality of experience (QoE) for users and for redundancy reasons, can lead to unnecessary bandwidth costs. We build TailGate, a system that exploits social relationships, regularities in read access patterns, and time-zone differences to efficiently and selectively distribute long-tail content across PoPs. We evaluate TailGate using large traces from an OSN and show that it can decrease WAN bandwidth costs by as much as 80% as well as reduce latency, improving QoE. We deploy TailGate on PlanetLab and show that even in the case when imprecise social information is available, TailGate can still decrease the latency for accessing long-tail YouTube videos by a factor of 2.

We consider the problem of providing mobility support for Voice-over-IP (VoIP) traffic in small-scale enterprises. There is considerable interest in providing on-the-go support for VoIP through the use of WiFi-enabled smart phones. However, existing solutions either do not support client mobility or require client modifications, making them difficult to deploy in practice.

Multiple communication channels are common in today's consumer and enterprise networks. For example, a high bandwidth but unreliable wireless network might co-exist with a reliable wired link (EWLANs and neighborhood networks). In this paper, we present a system that uses this reliable wired communication channel to boost the bandwidth of the lossy wireless link. Specifically, we propose a new, efficient partial packet recovery (PPR) technique and adaptive feedback mechanism specially designed to correct partial packets on an 802.11 wireless network using a wired backhaul. Our initial experiments demonstrate up to a 3x improvement over standalone 802.11 and upto a 30% improvement over existing PPR techniques.

Community wireless networks (CWNs) have been proposed to spread broadband network access to underprivileged, underprovisioned, and remote areas. Research has focused on optimizing network performance through intelligent routing and scheduling, borrowing solutions from mesh networks. Surprisingly, however, there has been no work on how to make efficient use of multiple channels in CWNs in the presence of multiple gateways and a single radio per device. In fact, today's deployments in underprivileged areas are primarily single-radio and do operate on a single channel. Frequency selection in such CWNs is very complex because it does not only determine the nodes' channel of operation, but also the gateway and the routing tree to the gateway-a rather computationally intensive task. In this paper, we propose, design, implement, and evaluate SWARM, a practical system that allows a CWN to make effective use of the available wireless channels in order to offer globally optimal performance. SWARM improves performance versus current single-channel protocols by up to 7.7 × in our experiments. Moreover, while we should be expecting performance gains due to channel diversity, we clearly demonstrate that up to 3.7 × improvement is attributed to the network organization into efficient traffic distribution structures.

Channel Bonding (CB) combines two adjacent frequency bands to form a new, wider band to facilitate high data rate transmissions in MIMO-based 802.11n networks. However, the use of a wider band with CB can exacerbate interference effects. Furthermore, CB does not always provide benefits in interference-free settings, and can even degrade performance in some cases. We conduct an in-depth, experimental study to understand the implications of CB. Based on this study we design an auto-configuration framework, ACORN, for enterprise 802.11n WLANs. ACORN integrates the functions of user association and channel allocation, since our study reveals that they are tightly coupled when CB is used. We show that the channel allocation problem with the constraints of CB is NP-complete. Thus, ACORN uses an algorithm that provides a worst case approximation ratio of [EQUATION] with Δ being the maximum node degree in the network. We implement ACORN on our 802.11n testbed. Our experiments show that ACORN (i) outperforms previous approaches that are agnostic to CB constraints; it provides per-AP throughput gains from 1.5x to 6x and (ii) in practice, its channel allocation module achieves an approximation ratio much better than [EQUATION].

Peer-to-peer has emerged in recent years as a promising approach to providing Video-on-Demand streaming. The design space, however, is vast and still not well understood---yet choosing the right approach is critical to system performance. This paper takes a fresh look at the p2p VoD design space using a simple analytical model that focuses on the allocation of uplink bandwidth resource for different chunks across peers. We describe a fundamental tradeoff that exists between system throughput, sequentiality of downloaded content and robustness to heterogeneous network conditions and node capacities, and we prove that no system can achieve all three simultaneously. Empirical results from Emulab confirm the analysis and show how one might implement efficient peer-to-peer VoD streaming with an appropriate balance of the tradeoff.

Recent work demonstrates that directional antennas have significant potential to improve wireless network capacity in indoor environments. This paper provides a broader exploration of the design space of indoor directional antenna systems along two main dimensions: antenna configuration and antenna control. Studying a number of alternative configurations, we find that directionality on APs and clients can significantly improve performance, even over other configurations with stronger directionality. Moreover, it is sufficient to have a small number of narrow beam antennas to achieve such gains, thus making such a solution practical for actual deployment. Designing systems with directional APs and clients for increased spatial reuse comes, however, with a number of challenges in the way the directional antennas are controlled. Antenna control needs to encompass antenna orientation algorithms, an appropriate MAC layer protocol, and novel client-AP association solutions. To overcome these challenges, we propose Speed, a distributed directional antenna control system that is easy to deploy and significantly improves network capacity over existing solutions.

Energy management is a critical issue for mobile devices, with network activity often consuming a significant portion of the total system energy. In this paper, we propose Catnap, a system that reduces energy consumption of mobile devices by allowing them to sleep during data transfers. Catnap exploits high bandwidth wireless interfaces -- which offer significantly higher bandwidth compared to available bandwidth across the Internet -- by combining small gaps between packets into meaningful sleep intervals, thereby allowing the NIC as well as the device to doze off. Catnap targets data oriented applications, such as web and file transfers, which can afford delay of individual packets as long as the overall transfer times do not increase. Our evaluation shows that for small transfers (128kB to 5MB), Catnap allows the NIC to sleep for up to 70% of the total transfer time and for larger transfers, it allows the whole device to sleep for a significant fraction of the total transfer time. This results in battery life improvement of up to 2-5x for real devices like Nokia N810 and Thinkpad T60.

Data-intensive applications that operate on large volumes of data have motivated a fresh look at the design of data center networks. The first wave of proposals focused on designing pure packet-switched networks that provide full bisection bandwidth. However, these proposals significantly increase network complexity in terms of the number of links and switches required and the restricted rules to wire them up. On the other hand, optical circuit switching technology holds a very large bandwidth advantage over packet switching technology. This fact motivates us to explore how optical circuit switching technology could benefit a data center network. In particular, we propose a hybrid packet and circuit switched data center network architecture (or HyPaC for short) which augments the traditional hierarchy of packet switches with a high speed, low complexity, rack-to-rack optical circuit-switched network to supply high bandwidth to applications. We discuss the fundamental requirements of this hybrid architecture and their design options. To demonstrate the potential benefits of the hybrid architecture, we have built a prototype system called c-Through. c-Through represents a design point where the responsibility for traffic demand estimation and traffic demultiplexing resides in end hosts, making it compatible with existing packet switches. Our emulation experiments show that the hybrid architecture can provide large benefits to unmodified popular data center applications at a modest scale. Furthermore, our experimental experience provides useful insights on the applicability of the hybrid architecture across a range of deployment scenarios.

Rate adaptation is a critical component that impacts the performance of IEEE 802.11 wireless networks. In congested networks, traditional rate adaptation algorithms have been shown to choose lower data-rates for packet transmissions, leading to reduced total network throughput and capacity. A primary reason for this behavior is the lack of real-time congestion measurement techniques that can assist in the identification of congestion-related packet losses in a wireless network. In this work, we first propose two real-time congestion measurement techniques, namely an active probe-based method called Channel Access Delay, and a passive method called Channel Busy Time. We evaluate the two techniques in a testbed network and a large WLAN connected to the Internet. We then present the design and evaluation of Wireless cOngestion Optimized Fallback (WOOF), a rate adaptation scheme that uses congestion measurement to identify congestion-related packet losses. Through simulation and testbed implementation we show that, compared to other well-known rate adaptation algorithms, WOOF achieves up to 300 percent throughput improvement in congested networks.

Wireless networks have evolved into an important technology for connecting users to the Internet. As the utility of wireless technology grows, wireless networks are being deployed in more widely varying conditions. The monitoring of wireless networks continues to reveal key implementation deficiencies that need to be corrected in order to improve protocol operation and end-to-end network performance. In wireless networks, where the medium is shared, unwanted traffic can pose significant overhead and lead to suboptimal network performance. Much of the previous analyses of unwanted traffic in wireless networks focus on malicious traffic. However, another major contributor of unwanted traffic is incorrect link layer behavior. Using data we collected from the 67th Internet Engineering Task Force (IETF) meeting held in November 2006, we show that a significant portion of link layer traffic stems from mechanisms that initiate, maintain, and change client-AP associations. We further show that under conditions of high medium utilization and packet loss rate, handoffs are initiated incorrectly. We analyze the traffic to understand when handoffs occur and whether the handoffs were beneficial or should have been avoided.

Dense deployments of WLANs suffer from increased interference and, as a result, reduced capacity. There are three main functions used to improve the overall network capacity: 1) intelligent frequency allocation across access points (APs); 2) load-balancing of user affiliations across APs; and 3) adaptive power control for each AP. Several algorithms have been proposed in each category, but so far, their evaluation has been limited to: a) each approach in isolation; and b) simulations or small-scale testbeds. In this paper, we ask the question: What is the best way to combine these different functions? Our focus is to fully explore the interdependencies between the three functions in order to understand when and how to deploy them on a network. We follow a measurement-driven study to quantify the effects of three previously proposed optimization schemes (one for each category) on a relatively large testbed and in many different scenarios. Surprisingly, we find that blindly applying all the three optimization schemes is not always preferable; it can sometimes degrade the performance by as much as 24% compared to using only two of the schemes. We discover that there are explicit conditions that are conducive for applying specific combinations of the optimization schemes. We capture these conditions within a comprehensive framework, which we call measurement-driven guidelines (MDG). While we derive such guidelines based on measurements on one experimental testbed, we test their applicability and efficacy on a second testbed in a different location. We show that our framework improves network capacity consistently across both testbeds, with improvements ranging from 22% to 142% with 802.11a, and 103% to 274% with 802.11g.

We propose a powerful MAC/PHY cross-layer approach to measuring IEEE 802.11 transmission opportunities in WLAN networks on a per-link basis. Our estimator can operate at a single station and it is able to: 1) classify losses caused by noise, collisions, and hidden nodes; and 2) distinguish between these losses and the unfairness caused by both exposed nodes and channel capture. Our estimator provides quantitative measures of the different causes of lost transmission opportunities, requiring only local measures at the 802.11 transmitter and no modification to the 802.11 protocol or in other stations. Our approach is suited to implementation on commodity hardware, and we demonstrate our prototype implementation via experimental assessments. We finally show how our estimator can help the WLAN station to improve its local performance.

Community wireless networks (CWNs) have been proposed to spread broadband network access to underprivileged, under-provisioned and remote areas. Research has focused on optimizing network performance through intelligent routing and scheduling, borrowing solutions from mesh networks. Surprisingly, however, there has been no work on how to make efficient use of multiple channels in CWNs in the presence of multiple gateways, and a single radio per device. In fact, today's deployments in under-privileged areas are primarily single radio and do operate on a single channel [20]. Frequency selection in such CWNs is very complex because it does not only determine the nodes' channel of operation but also the gateway and the routing tree to the gateway - a rather computationally intensive task. In this paper, we propose, design, implement, and evaluate SWARM, a practical system that allows a CWN to make effective use of the available wireless channels in order to offer globally optimal performance. SWARM improves performance versus current single channel protocols by up to 7.7× in our experiments. Moreover, while we should be expecting performance gains due to channel diversity, we clearly demonstrate that up to 3.7 x improvement is attributed to the network organization into efficient traffic distribution structures.

Enterprise WLANs have made a dramatic shift towards centralized architectures in the recent past. The reasons for such a change have been ease of management and better design of various control and security functions. The data path of WLANs, however, continues to use the distributed, random-access model, as defined by the popular DCF mechanism of the 802.11 standard. While theoretical results indicate that a centrally scheduled data path can achieve higher efficiency than its distributed counterpart, the likely complexity of such a solution has inhibited practical consideration. In this paper, we take a fresh, implementation and deployment oriented, view in understanding data path choices in enterprise WLANs. We perform extensive measurements to characterize the impact of various design choices, like scheduling granularity on the performance of a centralized scheduler, and identify regions where such a centralized scheduler can provide the best gains.Our detailed evaluation with scheduling prototypes deployed on two different wireless testbeds indicates that DCF is quite robust in many scenarios, but centralization can play a unique role in 1) mitigating hidden terminals - scenarios which may occur infrequently, but become pain points when they do and 2) exploiting exposed terminals - scenarios which occur more frequently, and limit the potential of successful concurrent transmissions. Motivated by these results, we design and implement CENTAUR - a hybrid data path for enterprise WLANs, that combines the simplicity and ease of DCF with a limited amount of centralized scheduling from a unique vantage point. Our mechanisms do not require client cooperation and can support legacy 802.11 clients.

The demand for wireless bandwidth in indoor environments such as homes and offices continues to increase rapidly. Although wireless technologies such as MIMO can reach link throughputs of 100s of Mbps (802.11n) for a single link, the question of how we can deliver high throughput to a large number of densely-packed devices remains an open problem. Directional antennas have been shown to be an effective way to increase spatial reuse, but past work has focused largely on outdoor environments where the interactions between wireless links can usually be ignored. This assumption is not acceptable in dense indoor wireless networks since indoor deployments need to deal with rich scattering and multipath effects. In this paper we introduce DIRC, a wireless network design whose access points use phased array antennas to achieve high throughput in dense, indoor environments. The core of DIRC is an algorithm that increases spatial reuse and maximizes overall network capacity by optimizing the orientations of a network of directional antennas. We implemented DIRC and evaluated it on a nine node network in an enterprise setting. Our results show that DIRC improves overall network capacity in indoor environments, while being flexible enough to adapt to node mobility and changing traffic workloads.

Increased AP density in enterprise WLANs leads to increasing RF interference and decreasing performance. An important step towards mitigating this problem is to construct precise RF maps in the form of a conflict graph. Prior work on conflict graph construction, mostly using bandwidth tests [17], suffers from two problems: a) It is limited to static settings and cannot support mobility, and b) It incurs significant measurement overhead and must be performed offline (e.g. overnight). An alternative to bandwidth tests is "micro-probing" [4] that operates on millisecond-level time scales. Micro-probing rapidly constructs the conflict graph even while the network is in use (i.e. online). While interesting in principle, micro-probing has only been evaluated in simulation. In this work, we empirically study micro-probing on a 40-node wireless testbed. In doing so, we not only show that micro-probing is in fact practically realizable, but also present key insights that drive the design choices for our implementation. We benchmark micro-probing against bandwidth tests and find that micro-probing is just as accurate but with up to a 400 times reduction in overhead. Finally, we argue that a successful implementation of micro-probing opens up the space for further innovations in real-time WLAN adaptation and optimization.

In this paper we take a look at the health of a typical enterprise network via a new metric based on the fraction of useful flows generated by endhosts. Flows considered non-useful are those that explicitly fail or else do not elicit a response from the intended destination. Examining traces collected from a large number of mobile hosts in an enterprise network, we find that about 34% of the flows are not useful. Through our study that combines data analysis and ongoing interactions with our IT department, we learn that these non-useful flows arise from several causes. Our mobile hosts frequently change environments, by either moving in and out of the corporate environment, or by switching the point and means of attachment to the corporate network. We find that many of the failures occur due to the hosts' lack of environment awareness, which results in attempts to discover services that are not present in all environments. Other causes include misconfiguration, unnecessary broadcast traffic, and excessive connection retries. Understanding this ever present noise in endhost communication is important for a variety of reasons including the fact that it complicates anomaly detection design and wastes resources, the latter of which is particularly crucial for wireless and mobile environments. Finally, we discuss possible means to design applications and services that can significantly improve the health of the network.

Residential Internet connectivity is growing at a phenomenal rate. A number of recent studies have attempted to characterize this connectivity - measuring coverage and performance of last-mile broadband links - from a various vantage points on the Internet, via wireless APs, and even with user cooperation. These studies, however, sacrifice accuracy or require substantial human time. In this work, we present a novel two-pass method to characterize neighborhood networks. We demonstrate that the two pass method dramatically reduces the time spent in active measurement while retaining accuracy. A case study on two neighborhoods in Pittsburgh provide new and accurate insights into broadband connectivity, including throughput, broadband coverage (DSL vs. cable vs. fiber), NAT configurations, DHCP, DNS usage. The results further characterize 802.11 connectivity in the neighborhood.

Traditional rate adaptation solutions for IEEE 802.11 wireless networks perform poorly in congested networks. Measurement studies show that congestion in a wireless network leads to the use of lower transmission data rates and thus reduces overall network throughput and capacity. The lack of techniques to reliably identify and characterize congestion in wireless networks has prevented development of rate adaptation solutions that incorporate congestion information in their decision framework. To this end, our main contributions in this paper are two-fold. First, we present a technique that identifies and measures congestion in an 802.11 network in real time. Second, we design Wireless congestion Optimized Fallback (WOOF), a measurement-driven rate adaptation scheme for 802.11 devices that uses the congestion measurement to identify congestion related packet losses. Through experimental evaluation, we show that WOOF achieves up to 300% higher throughput in congested networks, compared to other well-known adaptation algorithms.

Community wireless networks have been proposed as a powerful technique to spread broadband network access to underprivileged, under-provisioned and remote areas. These networks consist of a few Internet gateways which are reached by homes using multi-hop wireless links between wireless routers. The benefits of such networks include low costs for deployment due to reduced wiring needs, low maintenance and increased flexibility. Current practice in routing protocols for such networks (e.g. LQSR, OLSR and SrcRR) is for routing protocols to obtain information about the link quality (via some metric such as ETT, ETX) and select a gateway to whom a route minimizes the cost of the metric. All nodes operate on the same known frequency to maintain connectivity.

This paper addresses the design challenge of interference mitigation in the emerging high density (HD) wireless LAN. It is proposed to differentiate interference according to theirenergyandtimingrelative to desired signal, and measure packet error rate (PER) locally at transmitter for each type of interference. Then,self-adaptationalgorithms are designed to adjust a) clear channel assessment (CCA) threshold, aka physical carrier sensing threshold, to leverage spatial reuse for achieving higher aggregate throughput, and b) transmit power (TP) to compensate location difference among links, and prevent individual links from starving. Compared to an end-to-end (E2E) feedback loop, ours has negligible complexity and zero over-the-air overhead. Extensive OPNET simulations are used to compare the performance of our solutions against the legacy and the ideal.

Multi-hop wireless mesh networks are increasingly being deployed for last-mile Internet access. Typically, network algorithms such as routing, channel assignment and topology control for such networks rely heavily on metrics that intend to capture link "quality" across the network. However, the underlying dynamics of the proposed link metrics themselves have not yet been studied in detail. In this paper, we study the dynamics of the most popular link metrics in real network deployments. Using two wireless mesh testbeds, we measure a number of link metrics across different hardware platforms and network environments. The collected measurements allow us to study the stability and sensitivity of the different metrics to various conditions. Our study provides several insights and future research directions on how network algorithms need to adapt to link dynamics as well as how popular and widely used link metrics can be improved.

As the utility of wireless technology grows, wireless networks are being deployed in more widely varying conditions. The monitoring of these networks continues to reveal key implementation deficiencies that need to be corrected in order to improve protocol operation and end-to-end performance. Using data we collected from the 67th Internet Engineering Task Force (IETF) meeting held in November 2006, we show that under conditions of high medium utilization and packet loss, handoffs can be incorrectly initiated. Using the notion of persistence and prevalence for the association of a client to an Access Point (AP), we show that although the clients were predominantly static, the handoff rate is surprisingly high. Through the analysis of the data set, we show that unnecessary handoff events not only increase the amount of management traffic in the network, but also severely impact client performance.

The popularity of IEEE 802.11 WLANs has led to dense deployments in urban areas. High density leads to sub-optimal performance unless the interfering networks learn how to optimally use and share the spectrum. This paper proposes two fully distributed algorithms that allow (i) multiple interfering 802.11 access points to select their operating frequency in order to minimize interference, and (ii) users to choose the access point they attach to, in order to get their fair share of the whole network bandwidth. The proposed algorithms rely on Gibbs sampler, and do not require explicit coordination among the wireless devices. They only require the participating wireless nodes to measure local quantities such as interference and transmission delay. The algorithms are shown to lead to optimal bandwidth sharing, where optimality is defined according to the minimal potential delay. We analytically prove the convergence of the proposed algorithms, and study their performance by simulation.

The low cost and the ease of deployment of WiFi devices, as well as the need to support high bandwidth applications over 802.11 WLANs has led to the emergence of high density 802.11 networks in urban areas and enterprises. High density wireless networks, by design, face significant challenges due to increased interference resulting from the close proximity of co-channel cells. We demonstrate that power control can be used to mitigate interference in such an environment. It is well-known that variable transmit powers result in asymmetric links in the network, and can potentially lead to throughput starvation of some nodes. We first show that in order to perform starvation-free power control in 802.11 networks, across-layerapproach is required, whereby the transmit powers and the carrier sensing parameter of the MAC layer of the nodes should bejointlytuned. We then propose a framework that determines optimum settings for these parameters with the objective of maximizing the network-wide throughput for elastic traffic. Within this framework, we devise a distributed power control algorithm that uses a Gibbs sampler. OPNET simulations and experiments over a proof of concept testbed demonstrate that in a dense network the proposed power control algorithm yields significant improvement in client throughput.

Wireless LANs are commonplace installations in enterprise environments. Their ease of use and deployment, however, are accompanied by a difficulty in their management and security. Proposed solutions to these problems are based on centralization; in the control plane through centralized authentication and allocation of channels and power levels, and in the data plane through time slotted medium access using centralized scheduling for interference mitigation. While centralization of some control plane tasks has been shown to be feasible, centralization on the data plane is significantly harder to realize. This is because it needs to take into account the inherent variability of the wireless medium while offering bounds on delay and jitter on the control paths. In this work, we present a study of the various problems that arise in centralization of the data plane in an enterprise WLAN. We believe that a pragmatic solution for data plane centralization is the key approachto provisioning an enterprise WLAN consisting of a dense deployment of APs.

Dense deployments of WLANs suffer from increased interference and as a result, reduced capacity. There are three main functions used to improve the overall network capacity: a) intelligent frequency allocation across APs, b) load-balancing of user affiliations across APs, and c) adaptive power-control for each AP. Several algorithms have been proposed in each category, but so far, their evaluation has been limited to: (a) each approach in isolation and, (b)simulations or small-scale testbeds. In this paper, we ask the question: what is the best way to combine these different functions? Our focus is to fully explore the interdependencies between the three functions in order to understand when and how to deploy them on a network. We follow a measurement-driven study to quantify the effects of three previously proposed optimization schemes (one for each category) on a relatively large testbed and in many different scenarios. Surprisingly, we find that blindly applying all the three optimization schemes is not always preferable; it can sometimes degrade the performance by as much as 24% compared to using only two of the schemes. We discover that there are explicit conditions that are conducive for applying specific combinations of the optimization schemes. We capture those conditions within a comprehensive framework, which we call MDG (Measurement-Driven Guidelines). While we derive suchguidelines based on measurements on one experimental testbed, we test their applicability and efficacy on a second testbed in a different location. We show that our framework improves network capacity consistently acrossboth testbeds, with improvements ranging from 22% to 142% with 802.11a, and 103% to 274% with 802.11g.

The provisioning of high throughput performance infrastructure wireless networks necessitates the deployment of a high density of Access Points. While the latter improves wireless link quality to the clients, it can also introduce additional interference unless the network is carefully planned and tuned. It has been shown that the reaction of CSMA/CA protocol to interference is unnecessarily conservative in high density environments. In this work, we study the problem of infrastructure wireless network design, and the interaction between high density and MAC parameter tuning. Through analysis and numerical results, we provide recommendations on (i) optimum dimensioning of high density networks, and (ii) optimum tuning of their MAC parameters. We demonstrate that 802.11a networks are inherently noise-dominated, while 802.11g networks are interference-dominated, thus requiring different network design approaches. In sharp contrast to previous work, we establish that MAC parameter tuning has limited benefit in properly planned 802.11a networks. On the other hand, analytical results on the optimal tuning of MAC parameters in interference-dominated 802.11g deployments show substantial throughput improvements. Using the insight gained through our analysis, we propose an algorithm for the optimal tuning of MAC parameters in unstructured high density environments. Opnet simulations show that the proposed algorithm results in up to 260% improvement in network throughput.

The low price of commodity wireless LAN cards and access points (APs) has resulted in the rich proliferation of high density WLANs in enterprise, academic environments, and public spaces. In such environments wireless clients have a variety of affiliation options that ultimately determine the quality of service they receive from the network. The state of the art mechanism behind such a decision typically relies on received signal strength, associating clients to that access point (AP) in their neighborhood that features the strongest signal. More intelligent algorithms have been further proposed in the literature. In this work we take a step back and look into the fundamental metrics that determine end user throughput in 802.11 wireless networks. We identify three such metrics pertaining to wireless channel quality, AP capacity in the presence of interference, and client contention. We modify the low level software functionality (firmware and microcode) of a commercial wireless adaptor to measure the necessary quantities. We then test, in a real testbed, the ability of each metric to capture end user throughput through a range of diverse network conditions. Our experimental results indicate that user affiliation decisions should be based on metrics that do not only reflect physical layer performance, or network occupancy, but also concretely capture MAC layer behavior. Based on the acquired insight, we propose a new metric that is shown to be highly accurate across all tested network scenarios.

Abstract — Anecdotal evidence suggests that home wireless networks may be unpredictable despite their limited size. In this work, we deploy six-node wireless testbeds in three houses in the United States and the United Kingdom. We examine the quality of links in home wireless networks and the effect of (i) transmission rate, (ii) transmission power, (iii) node location, (iv) type of house, and (v) 802.11 technology. We provide empirical evidence suggesting that homes are challenging environments for wireless communication. Wireless links in the home are highly asymmetric and heavily influenced by precise node location, transmission power, and encoding rate, rather than physical distance between nodes. In our measurements, many links were unable to utilize the maximum transmission rate of the deployed 802.11 technology, and a few provided no connectivity at all. These results suggest that creating an AP-based topology with maximum coverage and throughput in this environment is challenging. Our findings have implications on the design of future home wireless networks and requirements for future wifi-enabled consumer electronic devices. We show that coverage and performance can be improved using a multi-hop topology, implying that mesh capabilities may actually be needed in consumer electronics for seamless connectivity across the home. I.

The handoff algorithms in the current generation of 802.11 networks are primarily reactive in nature, because they wait until the link quality degrades substantially to trigger a handoff. They further rely on instantaneous signal strength measurements when choosing the best AP. This approach leads to handoff delays on the order of 1-2 seconds that are unacceptable for delay sensitive applications such as VoIP. We propose a fundamentally new approach to handoffs that is based on continuous monitoring of wireless links. In our approach, a client measures the beacon strengths of all the APs operating on the current, and the overlapping channels, and makes its handoff decisions based on the long-term, and short-term trends in these signals. We show through experiments in a campus wireless network that our proposed algorithms result in more than 50% reduction in average handoff delays, while having the potential to improve overall user performance. Our algorithms have been implemented in today's hardware, and unlike other proposed roaming algorithms in the literature, need no infrastructure support.

We present a fundamentally different approach to classifying traffic flows according to the applications that generate them. In contrast to previous methods, our approach is based on observing and identifying patterns of host behavior at the transport layer. We analyze these patterns at three levels of increasing detail (i) the social, (ii) the functional and (iii) the application level. This multilevel approach of looking at traffic flow is probably the most important contribution of this paper. Furthermore, our approach has two important features. First, it operates in the dark, having (a) no access to packet payload, (b) no knowledge of port numbers and (c) no additional information other than what current flow collectors provide. These restrictions respect privacy, technological and practical constraints. Second, it can be tuned to balance the accuracy of the classification versus the number of successfully classified traffic flows. We demonstrate the effectiveness of our approach on three real traces. Our results show that we are able to classify 80%-90% of the traffic with more than 95% accuracy.

International audience

We introduce a methodology to predict when and where link additions/upgrades have to take place in an Internet protocol (IP) backbone network. Using simple network management protocol (SNMP) statistics, collected continuously since 1999, we compute aggregate demand between any two adjacent points of presence (PoPs) and look at its evolution at time scales larger than 1 h. We show that IP backbone traffic exhibits visible long term trends, strong periodicities, and variability at multiple time scales. Our methodology relies on the wavelet multiresolution analysis (MRA) and linear time series models. Using wavelet MRA, we smooth the collected measurements until we identify the overall long-term trend. The fluctuations around the obtained trend are further analyzed at multiple time scales. We show that the largest amount of variability in the original signal is due to its fluctuations at the 12-h time scale. We model inter-PoP aggregate demand as a multiple linear regression model, consisting of the two identified components. We show that this model accounts for 98% of the total energy in the original signal, while explaining 90% of its variance. Weekly approximations of those components can be accurately modeled with low-order autoregressive integrated moving average (ARIMA) models. We show that forecasting the long term trend and the fluctuations of the traffic at the 12-h time scale yields accurate estimates for at least 6 months in the future.

The traffic matrix of a telecommunications network is an essential input for any kind of network design and capacity planning decision. In this paper we address a debate surrounding traffic matrix estimation, namely whether or not the costs of direct measurement are too prohibitive to be practical. We examine the feasibility of direct measurement by outlining the computation, communication and storage overheads, for traffic matrices defined at different granularity levels. We illustrate that today's technology, that necessitates a centralized solution, does indeed incur prohibitive costs. We explain what steps are necessary to move towards fully distributed solutions, that would drastically reduce many overheads. However, we illustrate that the basic distributed solution, in which flow monitors are on all the time, is excessive and unnecessary. By discovering and taking advantage of a key stability property underlying traffic matrices, we are able to propose a new scheme that is distributed and relies only on a limited use of flow measurement data. Our approach is simple, accurate and scalable. Furthermore, it significantly reduces the overheads above and beyond the basic distributed solution. Our results imply that direct measurement of traffic matrices should become feasible in the near future.

We perform a detailed analysis of point-to-point packet delay in an operational tier-1 network. The point-to-point delay is the time between a packet entering a router in one PoP (an ingress point) and its leaving a router in another PoP (an egress point). It measures the one-way delay experienced by packets from an ingress point to an egress point across an ISP's network and provides the most basic information regarding the delay performance of the ISP's network. Using packet traces captured in the operational network, we obtain precise point-to-point packet delay measurements and analyze the various factors affecting them. Through a simple, step-by-step, systematic methodology and careful data analysis, we identify the major network factors that contribute to point-to-point packet delay and characterize their effect on the network delay performance. Our findings are: 1) delay distributions vary greatly in shape, depending on the path and link utilization; 2) after constant factors dependent only on the path and packet size are removed, the 99th percentile variable delay remains under 1 ms over several hops and under link utilization below 90% on a bottleneck; 3) a very small number of packets experience very large delay in short bursts.

A common traffic engineering design principle is to select a small set of flows, that account for a large fraction of the overall traffic, to be differentially treated inside the network so as to achieve a specific performance objective. We illustrate that one needs to be careful in implementing such an approach because there are tradeoffs to be addressed that arise due to traffic dynamics. We demonstrate that Internet flows are very volatile in terms of volume, and may substantially change the volume of traffic they transmit as time evolves. Currently proposed schemes for flow classification, although attractive due to their simplicity, face challenges due to this property of flows. Bandwidth volatility impacts the amount of load captured in a set of flows, which usually drops both significantly and quickly after flow classification is performed. Thus if the goal is to capture a large fraction of traffic consistently over time, flows will need to be reselected often. Our first contribution is in understanding the impact of flow volatility on the classification schemes employed in a traffic engineering context. Our second contribution is to propose a classification scheme that is capable of addressing the issues identified above by incorporating historical flow information. Using actual Internet data we demonstrate that our scheme outperforms previously proposed schemes, and reduces both the impact of flow volatility on the load captured by the selected set of flows and the required frequency for its reselection.

This paper provides an authoritative knowledge of through-router packet delays and therefore a better understanding of data network performance. Thanks to a unique experimental setup, we capture all packets crossing a router for 13 hours and present detailed statistics of their delays. These measurements allow us to build the following physical model for router performance: each packet experiences a minimum router processing time before entering a fluid output queue. Although simple, this model reproduces the router behaviour with excellent accuracy and avoids two common pitfalls. First we show that in-router packet processing time accounts for a significant portion of the overall packet delay and should not be neglected. Second we point out that one should fully understand both link and physical layer characteristics to use the appropriate bandwidth value.Focusing directly on router performance, we provide insights into system busy periods and show precisely how queues build up inside a router. We explain why current practices for inferring delays based on average utilization have fundamental problems, and propose an alternative solution to directly report router delay information based on busy period statistics.

Network traffic arises from the superposition of Origin-Destination (OD) flows. Hence, a thorough understanding of OD flows is essential for modeling network traffic, and for addressing a wide variety of problems including traffic engineering, traffic matrix estimation, capacity planning, forecasting and anomaly detection. However, to date, OD flows have not been closely studied, and there is very little known about their properties.We present the first analysis of complete sets of OD flow time-series, taken from two different backbone networks (Abilene and Sprint-Europe). Using Principal Component Analysis (PCA), we find that the set of OD flows has small intrinsic dimension. In fact, even in a network with over a hundred OD flows, these flows can be accurately modeled in time using a small number (10 or less) of independent components or dimensions.We also show how to use PCA to systematically decompose the structure of OD flow timeseries into three main constituents: common periodic trends, short-lived bursts, and noise. We provide insight into how the various constitutents contribute to the overall structure of OD flows and explore the extent to which this decomposition varies over time.

In order to control and manage highly aggregated Internet traffic flows efficiently, we need to be able to categorize flows into distinct classes and to be knowledgeable about the different behavior of flows belonging to these classes. In this paper we consider the problem of classifying BGP level prefix flows into a small set of homogeneous classes. We argue that using the entire distributional properties of flows can have significant benefits in terms of quality in the derived classification. We propose a method based on modeling flow histograms using Dirichlet Mixture Processes for random distributions. We present an inference procedure based on the Simulated Annealing Expectation Maximization algorithm that estimates all the model parameters as well as flow membership probabilities - the probability that a flow belongs to any given class. One of our key contributions is a new method for Internet flow classification. We show that our method is powerful in that it is capable of examining macroscopic flows while simultaneously making fine distinctions between different traffic classes. We demonstrate that our scheme can address issues with flows being close to class boundaries and the inherent dynamic behaviour of Internet flows.

SNMP statistics are usually collected over intervals of 5 minutes and correspond to average activity of IP links and network elements for the duration of the interval. Nevertheless, reports of traffic performance across periods of minutes can mask out performance degradation due to short-lived events, such as micro-congestion episodes, that manifest themselves at smaller time scales. In this paper we perform a measurement study of packet traces collected inside the Sprint IP network to identify the time scales over which micro-congestion episodes occur. We characterize these episodes with respect to their amplitude, frequency and duration. We define a new performance metric that could be easily computed by a router and reported every 5 minutes through SNMP to shed light into the micro-behavior of the carried traffic. We show that the proposed performance metric is well suited to track the time scales over which micro-congestion episodes occur, and may be useful for a variety of network provisioning tasks.

We introduce a methodology to predict when and where link additions/upgrades have to take place in an IP backbone network. Using SNMP statistics, collected continuously since 1999, we compute aggregate demand between any two adjacent PoPs and look at its evolution at time scales larger than one hour. We show that IP backbone traffic exhibits visible long term trends, strong periodicities, and variability at multiple time scales. Our methodology relies on the wavelet multiresolution analysis and linear time series models. Using wavelet multiresolution analysis, we smooth the collected measurements until we identify the overall long-term trend. The fluctuations around the obtained trend are further analyzed at multiple time scales. We show that the largest amount of variability in the original signal is due to its fluctuations at the 12 hour time scale. We model inter-PoP aggregate demand as a multiple linear regression model, consisting of the two identified components. We show that this model accounts for 98% of the total energy in the original signal, while explaining 90% of its variance. Weekly approximations of those components can be accurately modeled with low-order autoregressive integrated moving average (ARIMA) models. We show that forecasting the long term trend and the fluctuations of the traffic at the 12 hour time scale yields accurate estimates for at least six months in the future.

We measure and analyze the single-hop packet delay through operational routers in the Sprint Internet protocol (IP) backbone network. After presenting our delay measurements through a single router for OC-3 and OC-12 link speeds, we propose a methodology to identify the factors contributing to single-hop delay. In addition to packet processing, transmission, and queueing delay at the output link, we observe the presence of very large delays that cannot be explained within the context of a first-in first-out output queue model. We isolate and analyze these outliers. Results indicate that there is very little queueing taking place in Sprint's backbone. As link speeds increase, transmission delay decreases and the dominant part of single-hop delay is packet processing time. We show that if a packet is received and transmitted on the same linecard, it experiences less than 20 μs of delay. If the packet is transmitted across the switch fabric, its delay doubles in magnitude. We observe that processing due to IP options results in single-hop delays in the order of milliseconds. Milliseconds of delay may also be experienced by packets that do not carry IP options. We attribute those delays to router idiosyncratic behavior that affects less than 1% of the packets. Finally, we show that the queueing delay distribution is long-tailed and can be approximated with a Weibull distribution with the scale parameter a=0.5 and the shape parameter b=0.6 to 0.82.

No abstract available.

We measure and analyze the single-hop packet delay through operational routers in a backbone IP network. First we present our delay measurements through a single router. Then we identify step-by-step the factors contributing to single-hop delay. In addition to packet processing, transmission, and queueing delays, we identify the presence of very large delays due to non-work-conserving router behavior. We use a simple output queue model to separate those delay components. Our step-by-step methodology used to ohtain the pure queueing delay is easily applicable to any single-hop delay measurements. After obtaining the queueing delay, we analyze the tail of its distribution, and find that it is long tailed and fits a Weihull distrihution with the scale parameter, a = 0.5, and the shape parameter, b = 0.58 to 0.6. The measured average queueing delay is larger than predicted by M/M/l, M/G/l, and FBM models when the link utilization is below 70%, but its absolute value is quite small.