Modern enterprise networks are of sufficient complexity that even simple faults can be difficult to diagnose - let alone transient outages or service degradations. Nowhere is this problem more apparent than in the 802.11-based wireless access networks now ubiquitous in the enterprise. In addition to the myriad complexities of the wired network, wireless networks face the additional challenges of shared spectrum, user mobility and authentication management. Not surprisingly, few organizations have the expertise, data or tools to decompose the underlying problems and interactions responsible for transient outages or performance degradations. In this paper, we present a set of modeling techniques for automatically characterizing the source of such problems. In particular, we focus on data transfer delays unique to 802.11 networks - media access dynamics and mobility management latency. Through a combination of measurement, inference and modeling we reconstruct sources of delay - from the physical layer to the transport - layer as well as the interactions among them. We demonstrate our approach using comprehensive traces of wireless activity in the UCSD Computer Science building.

The combination of unlicensed spectrum, cheap wireless interfaces and the inherent convenience of untethered computing have made 802.11 based networks ubiquitous in the enterprise. Modern universities, corporate campuses and government offices routinely de-ploy scores of access points to blanket their sites with wireless Internet access. However, while the fine-grained behavior of the 802.11 protocol itself has been well studied, our understanding of how large 802.11 networks behave in their full empirical complex-ity is surprisingly limited. In this paper, we present a system called Jigsaw that uses multiple monitors to provide a single unified view of all physical, link, network and transport-layer activity on an 802.11 network. To drive this analysis, we have deployed an infrastructure of over 150 radio monitors that simultaneously capture all 802.11b and 802.11g activity in a large university building (1M+ cubic feet). We describe the challenges posed by both the scale and ambiguity inherent in such an architecture, and explain the algorithms and inference techniques we developed to address them. Finally, using a 24-hour distributed trace containing more than 1.5 billion events, we use Jigsaw's global cross-layer viewpoint to isolate performance artifacts, both explicit, such as management inefficiencies, and implicit, such as co-channel interference. We believe this is the first analysis combining this scale and level of detail for a production 802.11 network.

In this paper a passive methodology for TCP performance evaluation over general packet radio service (GPRS) networks is presented that relies on traffic monitoring at the GPRS ingress/egress router interface (Gi). Based on the IP and TCP headers of the packets we estimate the end-to-end performance of TCP connections such as connection setup behavior and data transfer goodput. In order to identify the effects behind the measured performance the introduced algorithms estimate round trip delays, packet loss ratios, available channel rates, throughput and cany out bottleneck analysis. Large-scale GPRS measurements in seven countries are presented to analyze TCP performance and demonstrate the applicability of the method. The effects of different TCP parameters such as maximum segment size, selective acknowledgements, timestamp usage and receiver window size are also quantified. GPRS measurement results are compared to a wireline dial-up network to identify the effects specific to the wireless environment

This paper presents a passive end-to-end loss monitoring method, which relies on traffic monitoring at a core or ingress router interface. The monitoring node captures the packets of TCP connections generated by end-hosts. Based on the seen sequence number pattern the loss ratios are estimated for the two segments of end-to-end path divided by the monitor. This feature is especially useful if the monitoring node is placed at the border of an autonomous system, e.g., at the egress point of an ISP. When applied to mobile Internet access, packet losses on the radio interface can be distinguished from losses in the core network or in the public Internet. The packet loss estimation algorithm is rigorously validated using simulations and testbed measurements. Additionally, measurement results and an application example in a dial-up ISP are presented.

A basic network setup is investigated via measurements and simulation of TCP traffic. Based on our measurements we found that the packet losses occur in the backbone network, in spite of the fact that the majority of packets mainly stays in the routers' queues of the access network. Our conclusion is that the packet loss process is independent of the TCP flows' congestion window processes. The packet loss process depends on the aggregate behaviour of a large number of end-users, therefore depends -say-on the time of the day. Based on the knowledge of the packet loss process we are able to estimate the round trip delay that is caused by permanent TCP flows. This estimation can be done by both analytically or simulations as shown in the paper.