Introduction
Invitation
Conference Committee
General Info
Call for Papers
Program At-A-Glance
Guest Speakers and Panels
Tutorials
Technical Program

Other Conference Events

OPENARCH'98
Gigabit

Request for Info

Registration

Gateway to the 21st Century
29 March - 2 April 1998 Hotel Nikko, San Francisco, USA

Tutorials

---

Sunday, 29 March 1998 - half day
2:30 - 6:30 pm

---

T1 Broadband Satellite Network
Marie-Jose Montpetit, Teledesic

This tutorial will address various issues in broadband satellite network. It will focus on the next generation satellite systems and will present the reasons why they are more and more considered as an important part of the portfolio of broadband system solutions. After a review of satellite fundamentals such as orbits, RF issues, space environment issues, major subsystems and constraints and advantages necessary to fully understand the challenges of broadband satellite systems, the tutorial will present the different proposed systems. In particular, the tutorial will describe geostationary (GEO) systems with no on-board processing (OBP) like Eutelsat and Cadenza, GEO systems with OBP like Advanced Satcom, Spaceway and Cyberstar, Low Earth Orbit (LEO) system without OBP like Skybridge and LEO systems with OBP like Celestri and Teledesic. The remainder of the tutorial will focus on the technical challenges and risk areas. Network architectures, resource management, routing, on-board processing, terminal concept and architecture as well as protocol impact will be discussed in depth.

Location: Nikko I

Operators of high-speed networks, such as wide area carriers and Internet backbone service providers, are concerned to implement charging schemes by which they can fairly recover costs from their customers and also effectively control the allocation of resources in their networks. Well-designed tariffs not only generate income for the network, but also introduce feedback and control by providing the right incentives for network usage to the customers, and hence improve performance by reducing unnecessary congestion. An important aspect of charging schemes is their technological feasibility. Since one of the major operating costs of a network is accounting and billing, there is an important trade-off between the complexity of the charging model in terms of measurements and the resulting accuracy of the model for expressing effective usage and providing fair charging. Furthermore, such models should comply and take advantage of existing technology standards for broadband equipment and services. This course will address the above issues, and propose a framework that can be used for constructing simple but yet effective usage-based charging schemes. In particular we will address the following topics.

• Review of basic concepts in broadband networks (ATM, Internet) related to resource allocation and traffic management

• Introduction of basic tools for assessing resource usage in networks with a high degree of statistical multi-plexing (effective bandwidths).

• A methodology for constructing simple usage based charging schemes for guaranteed services, and its implications for user behavior (traffic shaping, choice of traffic contract).

• Charging best-effort services and its relationship to the fair allocation of network resources (max-min fairness and proportional fairness).

• Issues related to Internet charging, the implication of charging to call acceptance control, the advantages of renegotiating traffic contracts, etc.

Location: Pink Pearl

Multicast is becoming an increasingly important service for many network applications, including teleconferencing, telelecturing, software distribution, push-based information dissemination, and distributed games. The dizzying range of requirements imposed by these applications present many challenges. This tutorial examines the network protocols, services, and technology for supporting multicast communication. Throughout the tutorial, we focus on both existing solutions as well as on-going research efforts.

We begin by discussing the unique application requirements for multicast, and illustrate its use by several existing Internet and intranet applications. We then examine the Internet IP multicast model, various multicast routing protocols and the MBone. We next turn our attention to the higher-level concerns of reliable data transfer and congestion control in a multicast environment, highlighting the many challenges imposed by a potentially large, heterogeneous set of multicast users and the wide range of possible solutions. We conclude with a brief discussion of multicast ATM networks and, more generally, the directions in which multicast is continuing to evolve.

Location: Pink Pearl

As use of the Internet diversifies and expands at an exceptional rate the issue of how to provide necessary quality of service (QoS) for a wide variety of different user applications is gaining increasing importance. The broadband ISDN is an alternative multi-service network platform with a greater potential for QoS guarantees although there remains considerable confusion about the way users and network operators can optimally exploit the standardized ATM service classes. In this tutorial we aim to clarify the quality of service issue and to critically examine the effectiveness of proposed networking solutions.

The term quality of service in fact covers a range of network performance requirements which must be fulfilled conjointly by the definition of an appropriate service model and by traffic engineering. We classify QoS requirements according to whether they relate to: transparency (temporal or semantic integrity), throughput (the realized rate of information transfer of a flow) and accessibility (the availability of required network resources to handle expressed demand). The service model, by defining, different services classes and priorities for access to network resources, allows the provision of transparency and throughput guarantees for some users; traffic engineering, consisting essentially in sizing the network to meet estimated demand, ensures that such quality of service is available for a population of users.

The tutorial will cover both aspects of QoS provision and is organized according to the following chapter headings: the nature of traffic demand in a multi-service network; proposed multi-service network service models; performance of network multi-plexers and schedulers; methods for admission control; multi-service network traffic engineering; network evolution perspective.

Location: Nikko III

This course will cover the emerging area of XDSL, Protocols and Transmission. The latest trends of standardization, including ATM over ADSL standards efforts in ADSL Forum ,the ANSI/ITU T1.413 Issue II standard, G.lite and G.dmt. At the protocol-stack level, ATM over xDSL protocol stacks and protocol conversion options, +IP over PPP over ATM over ADSL (ADSL Forum 97-127), and ATM vs packet mode backbones. QoS Support issues related to voice/ telephony applications, video conferencing, motion video (MPEG or JPEG), high resolution imaging, Circuit emulation service (CES), Connection Admission Control - QOS based path selection when setting up, traffic scheduling, usage parameter control, and synchronization. Multimedia support will also be considered. Latest Trends of Telco Market, including Digital Loop Carrier support of ADSL, and CLECs vs ILECs vs private ADSL based enterprise networks will also be considered.

At the transmission layer, the difficult transmission challenges of DSL will be reviewed and then transmission methods to meet the challenges will be outlined and investigated. Some detail focus using the T1.413 ADSL issue-2 standard and the VDSL Alliance standards proposal will be discussed. At the end, a little time will be spent on other technologies such as high speed POTS modems and cable modem technologies.

Location: Nikko I

With the increased penetration of distributed (network) computing and the use of networks to support critical business applications, (network) security has become a key concern to most users. This tutorial will review major security issues and technologies, starting with basic cryptographic primitives for authentication, encryption,trust and key management, and their use in IP and ATM standards to provide various security capabilities. Existing technologies such as firewalls will also be covered. Throughout the presentation, we will cover existing standards and future directions. We begin by discussing the threat model for the Intranet, Extranet, and Internet and outline the security requirements (e.g., authentication, integrity) for each of these environments. We outline why both communication and platform security are required to satisfy security objectives. Then we outline the fundamental cryptographic techniques available for securing communication including symmetric and public key ciphers, and hash functions. Then,we proceed to discuss the trust management using certificate infrastructure (e.g., X.509). We next turn our attention to communication security protocols including IPSEC, SSL/TLS, Secure RPC, S/MIME and ATM security. We will detail the protocols, their security properties,performance, strengths and weaknesses and their status as standards. Then, we describe the challenges with the deployment of the security protocols. Finally, we conclude with a brief discussion of multicast security and more generally, the directions in which secure communication protocols continue to evolve.

Location: Grey Pearl

This full-day tutorial will cover recent advances in network transport protocols, both for ``data'' and real-time services, as well as the interaction of transport protocols with lower-layer protocols such as ATM, resource reservation and adaptive multimedia services.

The tutorial will start by reviewing the basic mechanism of TCP, indicate how its connection setup can be modified to reduce latency for short transaction (transaction TCP), and then investigate new flow and congestion control mechanism, including new-Reno, selective acknowledgement TCP (SACK TCP). Recent research work such as TCP Vegas and rate-based TCP will follow.

Even though TCP is independent of lower layers, TCP cannot be evaluated or improved in isolation. We will cover the interaction of TCP Vegas and Reno with ATM UBR and available bit rate (ABR) services. For real-time services, protocols and algorithms have been developed that allow real-time delivery over the Internet, with protocols such as RTP providing information about the quality of service visible to the end systems. This feedback can be used to dynamically adjust the sending rate of sources or regulate the transmission of enhancement layers. In the Internet, TCP-based and streaming multimedia services must co-exist, even if there are no resource reservations in place to separate different traffic classes. The tutorial will summarize recent work on the coexistence of TCP- and RTP-based flows.

Initial ideas on flow and congestion control for multicast services will be discussed. Finally, interactions between TCP and other types of networks, such as wireless and satellite will also be discussed.