Mark Stemm

Vertical Handoffs in Wireless Overlay Networks

No single wireless network technology simultaneously provides a low latency, high bandwidth, wide area data service to a large number of mobile users. Wireless Overlay Networks - a hierarchical structure of room-size, building-size, and wide area data networks - solve the problem of providing network connectivity to a large number of mobile users in an efficient and scalable way. The specific topology of cells and the wide variety of network technologies that comprise wireless overlay networks present new problems that have not been encountered in previous cellular handoff systems. We have implemented a vertical handoff system that allows users to roam between cells in wireless overlay networks. Our goal is to provide a user with the best possible connectivity for as long as possible with a minimum of disruption during handoff. Results of our initial implementation show that the handoff latency is bounded by the discovery time, the amount of time before the mobile host discovers that it has moved into or out of a new wireless overlay. This discovery time is measured in seconds: large enough to disrupt reliable transport protocols such as TCP and introduce significant disruptions in continuous multimedia transmission. To efficiently support applications that cannot tolerate these disruptions, we present enhancements to the basic scheme that significantly reduce the discovery time without assuming any knowledge about specific channel characteristics. For handoffs between room-size and building-size overlays, these enhancements lead to a best-case handoff latency of approximately 170 ms with a 1.5% overhead in terms of network resources. For handoffs between building-size and wide-area data networks, the best-case handoff latency is approximately 800 ms with a similarly low overhead.

A network architecture for heterogeneous mobile computing

This article summarizes the results of the BARWAN project, which focused on enabling truly useful mobile networking across an extremely wide variety of real-world networks and mobile devices. We present the overall architecture, summarize key results, and discuss four broad lessons learned along the way. The architecture enables seamless roaming in a single logical overlay network composed of many heterogeneous (mostly wireless) physical networks, and provides significantly better TCP performance for these networks. It also provides complex scalable and highly available services to enable powerful capabilities across a very wide range of mobile devices, and mechanisms for automated discovery and configuration of localized services. Four broad themes arose from the project: (1) the power of dynamic adaptation as a generic solution to heterogeneity, (2) the importance of cross-layer information, such as the exploitation of TCP semantics in the link layer, (3) the use of agents in the infrastructure to enable new abilities and to hide new problems from legacy servers and protocol stacks, and (4) the importance of soft state for such agents for simplicity, ease of fault recovery, and scalability.

A network measurement architecture for adaptive applications

The quality of network connectivity between a pair of Internet hosts can vary greatly. Adaptive applications can cope with these differences in connectivity by choosing alternate representations of objects or streams or by downloading the objects from alternate locations. In order to effectively adapt, applications must discover the condition of the network before communicating with distant hosts. Unfortunately, the ability to predict or report the quality of connectivity is missing in todays suite of Internet services. To address this limitation, we have developed SPAND (shared passive network performance discovery), a system that facilitates the development of adaptive network applications. In each domain, applications make passive application specific measurements of the network and store them in a local centralized repository of network performance information. Other applications may retrieve this information from the repository and use the shared experiences of all hosts in a domain to predict future performance. In this way, applications can make informed decisions about adaptation choices as they communicate with distant hosts. In this paper, we describe and evaluate the SPAND architecture and implementation. We show how the architecture makes it easy to integrate new applications into our system and how the architecture has been used with specifics types of data transport. Finally, we describe LookingGlass, a WWW mirror site selection tool that uses SPAND. LookingGlass meets the conflicting goals of collecting passive network performance measurements and maintaining good client response times. In addition, LookingGlasss server selection algorithms based on application level measurements perform much better than techniques that rely on geographic location or route metrics.

Analyzing stability in wide-area network performance

The Internet is a very large scale, complex, dynamical system that is hard to model and analyze. In this paper, we develop and analyze statistical models for the observed end-to-end network performance based on extensive packet-level traces (consisting of approximately 1.5 billion packets) collected from the primary Web site for the Atlanta Summer Olympic Games in 1996. We find that observed mean throughputs for these transfers measured over 60 million complete connections vary widely as a function of end-host location and time of day, confirming that the Internet is characterized by a large degree of heterogeneity. Despite this heterogeneity, we find (using best-fit linear regression techniques) that we can express the throughput for Web transfers to most hosts as a random variable with a log-normal distribution. Then, using observed throughput as the control parameter, we attempt to quantify the spatial (statistical similarity across neighboring hosts) and temporal (persistence over time) stability of network performance. We find that Internet hosts that are close to each other often have almost identically distributed probability distributions of throughput. We also find that throughputs to individual hosts often do not change appreciably for several minutes. Overall, these results indicate that there is promise in protocol mechanisms that cache and share network characteristics both within a single host and amongst nearby hosts.

TCP behavior of a busy Internet server: analysis and improvements

We analyze the way in which Web browsers use TCP connections based on extensive traffic traces obtained from a busy Web server (the official Web server of the 1996 Atlanta Olympic games). At the time of operation, this Web server was one of the busiest on the Internet. We first describe the techniques used to gather these traces and reconstruct the behavior of the TCP on the server. We then present a detailed analysis of the TCPs loss recovery and congestion control behavior from the recorded transfers. Our two most important results are: (1) short Web transfers lead to poor loss recovery performance for TCPs, and (2) concurrent connections are overly aggressive users of the network. We then discuss techniques designed to solve these problems. To improve the data-driven loss recovery performance of short transfers, we present a new enhancement to the TCPs loss recovery. To improve the congestion control and loss recovery performance of parallel TCP connections, we present a new integrated approach to congestion control and loss recovery that works across the set of concurrent connections. Simulations and trace analysis show that our enhanced loss recovery scheme could have eliminated 25% of all timeout events, and that our integrated approach provides greater fairness and improved startup performance for concurrent connections.