John D. DeHart

Supercharging planetlab: a high performance, multi-application, overlay network platform

In recent years, overlay networks have become an important vehicle for delivering Internet applications. Overlay network nodes are typically implemented using general purpose servers or clusters. We investigate the performance benefits of more integrated architectures, combining general-purpose servers with high performance Network Processor (NP) subsystems. We focus on PlanetLab as our experimental context and report on the design and evaluation of an experimental PlanetLab platform capable of much higher levels of performance than typical system configurations. To make it easier for users to port applications, the system supports a fast path/slow path application structure that facilitates the mapping of the most performance-critical parts of an application onto an NP subsystem, while allowing the more complex control and exception-handling to be implemented within the programmer-friendly environment provided by conventional servers. We report on implementations of two sample applications, an IPv4 router, and a forwarding application for the Internet Indirection Infrastructure. We demonstrate an 80x improvement in packet processing rates and comparable reductions in latency.

A scalable high-performance active network node

Active networking in environments built to support link rates up to several gigabits per second poses many challenges. One such challenge is that the memory bandwidth and individual processing power of the routers microprocessors limit the total available processing power of a router. In this article we identify and describe three components, which promise a high-performance active network solution. This implements the key features typical to active networking, such as automatic protocol deployment and application specific processing, and it is suitable for a gigabit environment. First, we describe the hardware of the active network node (ANN), a scalable high-performance platform based on off-the-shelf CPUs connected to a gigabit ATM switch backplane. Second, we introduce the ANNs modular, extensible, and highly efficient operating system (NodeOS). Third, we describe an execution environment running on top of the NodeOS, which implements a novel large-scale active networking architecture called distributed code caching.

The open network laboratory

The Open Network Laboratory (ONL) is a remotely accessible network testbed of high performance routers which has been designed with an eye towards ease of use for users from the naïve to the expert. The system is built around a set of high-performance routers that are extendible and easily configurable through the Remote Laboratory Interface (RLI), an intuitive graphical interface. The RLI also makes it easy to configure packet filters in the routers, assign flows or flow aggregates to separate queues with configurable QoS and attach hardware monitoring points to real-time charts. The RLIs real-time charts and user data facility make it easy to directly view the effects of traffic as it moves through a router, allowing the user to gain better insight into system behavior and create compelling demonstrations. Each port of the router is equipped with an embedded processor that supports software plugins which allow users to extend the systems functionality. This paper describes the ONL and how it can be used in networking education. Our web site onl.arl.wustl.edu includes a short video and a tutorial.

Multipoint connection management in high speed networks

A connection management access protocol (CMAP) for managing multipoint connections in high-speed packet switched networks is described. The protocol is targeted to broadband integrated services digital networks (BISDN) using the asynchronous transfer mode (ATM) communication standard. To establish a multipoint connection, a client first creates a call between itself and the network. The client creating a call is designated the owner of the call. Immediately after creation, the owner is the only endpoint participating in the call. Additional endpoints are added either by invitation from the owner, invitation from another client of the network, or by explicitly requesting to be added. These three modes are sufficient for supporting point-to-point communication (for example, a telephone call), many-to-many communication (for example, a conference call), one-to-many communication (for example, broadcast video), and many-to-one communication (for example, error logging).<<ETX>>

Design and evaluation of a high-performance dynamically extensible router

This paper describes the design, implementation and performance of an open, high performance, dynamically extensible router under development at Washington University in St. Louis. This router supports the dynamic installation of software and hardware plug-ins in the data path of application data flows. It provides an experimental platform for research on programmable networks, protocols, router software and hardware design, network management, quality of service and advanced applications. It is designed to be flexible without sacrificing performance. It supports gigabit links and uses a scalable architecture suitable for supporting hundreds or even thousands of links. The systems flexibility makes it an ideal platform for experimental research on dynamically extensible networks that implement higher level functions in direct support of individual application sessions.

A call model for multipoint communication in switched networks

A call model designed for switched asynchronous transfer mode (ATM) networks is described. The model provides network clients with multipoint, multiconnection communication channels, which are denoted as calls. Calls are allowed to change dynamically during their lifetime, in terms of the number of participants, the number of connections and the bandwidth of the connections. Clients create, manage and manipulate calls using the connection management access protocol. The call model provides basic interconnection services suitable for local and wide area networks, where more sophisticated services can be layered over this substrate.<<ETX>>

The Smart Port Card: An Embedded Unix Processor Architecture for Network Management and Active Networking

This paper describes the architecture of the Smart Port Card (SPC) designed for use with the Washington University Gigabit Switch. The SPC uses an embedded Intel Pentium processor running open-source NetBSD to support network management and active networking applications. The SPC physically connects between a switch port and a normal link adapter, allowing cell streams to be processed as they enter or leave the switch. In addition to the hardware architecture, this paper describes current and future applications for the SPC.

Connection Management Access Protocol (CMAP) Specification

This document specifies a Connection Management Access Protocol (CMAP) for managing multipoint connectinos in high-speed packet switched networks. We target CMAP to networks employing the Asynchronous Transfer Mode (ATM) communication standard. We define a multipoint connection as a communication channel between two or more clients of the network, where all data sent by one client is received by all other clients who have elected to receive. A point-to-point connection is a special case of a multipoint connection involving only two clients. CMAP specifies the access procedures exercised by clients to create, modify and delete multipoint connections. once a connection... Read complete abstract on page 2.

The open network laboratory: a resource for networking research and education

The Open Network Laboratory (ONL) is a remotely accessible network testbed designed to enable networking faculty, students and researchers to conduct experiments using high performance routers and applications. The system is built around a set of extensible, high-performance routers and has a graphical interface that enables users to easily configure and run experiments remotely. ONLs Remote Laboratory Interface (RLI) allows users to easily configure a network topology, configure routes and packet filters in the routers, assign flows or flow aggregates to separate queues with configurable QoS and attach hardware monitoring points to real-time charts. The remote visualization features of the RLI make it easy to directly view the effects of traffic as it moves through a router, allowing the user to gain better insight into system behavior and create compelling demonstrations. Each port of the router is equipped with an embedded processor that provides a simple environment for software plugins allowing users to extend the systems functionality. This paper describes the general facilties and some networking experiments that can be carried out. We hope that you and your collegues and students will check out the facility and register for an account at our web site <u>onl.arl.wustl.edu</u>

Vaudeville: A High Performance, Voice Activated TeleconferencingApplication

We present Vaudeville, a voice-activated, hands-free, ATM-based video conferencing application. This system is scalable; although video bandwidth is normally a limiting factor in the number of conferences participants, the bandwidth attributed to the video is not a function of conference size. This is achieved through an automatic, distributed floor control mechanism that gives the appearance of an open floor. Audio and video are encoded in hardware using a platform-independent, ATM hardware multimedia interface. Vaudeville features digitally transmitted NTSC video, voice-activated audio transmission, audio bridging of two audio streams, and voice-activated video switching. Multiple simultaneous multiparty conferences are supported. Users can move freely among conferences without knowledge of the underlying network structure. We describe how Vaudeville was built using a component-based distributed programming environment. We also describe the algorithms used to control the audio and video of the application.