Abstract – The Internet is rapidly growing in number of users, traffic levels, and topological complexity. At the same time it is increasingly driven by economic competition. These developments render the characterization of network usage and workloads more difficult, and yet more critical. Few recent studies have been published reporting Internet backbone traffic usage and characteristics. At MCI, we have implemented a high-performance, low-cost monitoring system that can capture traffic and perform analyses. We have deployed this monitoring tool on OC-3 trunks within internetMCI’s backbone and also within the NSF-sponsored vBNS. This paper presents observations on the patterns and characteristics of wide-area Internet traffic, as recorded by MCI’s OC-3 traffic monitors. We report on measurements from two OC-3 trunks in MCI’s commercial Internet backbone over two time ranges (24-hour and 7-day) in the presence of up to 240,000 flows. We reveal the characteristics of the traffic in terms of packet sizes, flow duration, volume, and percentage composition by protocol and application, as well as patterns seen over the two time scales. 1

Increased bandwidth in the Internet puts great demands on network routers; for example, to route minimum sized Gigabit Ethernet packets, an IP router must process about packets per second per port. Using the "rule-of-thumb" that it takes roughly 1000 packets per second for every 10 6 bits per second of line rate, an OC-192 line requires routing lookups per second; well above current router capabilities. One limitation of router performance is the route lookup mechanism. IP routing requires that a router perform a longest-prefix-match address lookup for each incoming datagram in order to determine the datagram's next hop. In this paper, we present a route lookup mechanism that when implemented in a pipelined fashion in hardware, can achieve one route lookup every memory access. With current 50ns DRAM, this corresponds to approximately packets per second; much faster than current commercially available routing lookup schemes. We also present novel schemes for performing quick updates ...

Finding the longest matching prefix from a database of keywords is an old problem with a number of applications, ranging from dictionary searches to advanced memory management to computational geometry. But perhaps today's most frequent best matching prefix lookups occur in the Internet, when forwarding packets from router to router. Internet traffic volume and link speeds are rapidly increasing; at the same time, an increasing user population is increasing the size of routing tables against which packets must be matched. Both factors make router prefix matching extremely performance critical. In this paper, we introduce a taxonomy for prefix matching technologies, which we use as a basis for describing, categorizing, and comparing existing approaches. We then present in detail a fast scheme using binary search over hash tables, which is especially suited for matching long addresses, such as the 128 bit addresses proposed for use in the next generation Internet Protocol, IPv6. We also present optimizations that exploit the structure of existing databases to further improve access time and reduce storage space.

Introduction One of the principal design considerations in a flow switching network is the definition of a flow. Flow switching uses a network of connection oriented switches that allow flows to be cut-through on their own connection. The granularity of a flow can vary over a wide range from the very fine, for example an individual application level transaction, to the very coarse, such as the aggregated traffic on the next hop between adjacent routing nodes. There are technical trade-offs to be considered together with the constraints of current technology when addressing the question of the appropriate level of aggregation. It appears that all the levels of aggregation have their place in flow switching and that the aggregation level should be a parameter available to network designers for optimization. Flows and flow switching networks Flow switching networks are characterized by the combination of conventional store-and-forward routin

We describe a WDM-based optical access network architecture for providing broadband Internet services. The architecture uses a passive collection and distribution network and a configurable Feeder network. Unlike earlier papers that concentrate on the physical layer design of the network, we focus on higher layer architectural considerations. In particular we discuss the joint design of the electronic and optical layers including: WDM Medium Access Control protocols; the choice of electronic multiplexing and switching between the IP and WDM layers; joint optical and electronic protection mechanisms; network reconfiguration algorithms that alter the logical topology of the network in response to changes in traffic; and traffic grooming algorithms to minimize the cost of electronic multiplexing. Finally we also discuss the impact of the optical topology on higher layer protocols such as IP routing, TCP flow control and multi-layer switching. 1.

. In this paper we investigate issues related to the integration of QoS capabilities in the framework of IP switching technology. First, we propose an architecture which includes the RSVP protocol with minimal impact on the original structure and protocols of IP switching. This leads us to further investigate the queueing and scheduling mechanisms adopted both at the routing and the switching level in order to guarantee effective support of Guaranteed (GS) and Best-Effort (BE) services. Based on the observation that GS service typically results in network underutilization, we then introduce a new service class (BR) that provides flows with long-term rate guarantees without impacting GS flows and while exploiting their unused bandwidth. Second, we propose a possible implementation of this new service and we analyze it with the help of simulation. Finally we briefly discuss scalability problems in an integrated services model. 1 Introduction New applications such as teleme...

Broadband networking concepts and technologies, such as open signalling, IP switching and xDSL technologies are experiencing a rapid evolution whereas the service concept in traditional ATM networks has not evolved at the same pace. As opposed to the broadband IN-based approach we have composed one vision of a flexible broadband media services and network architecture. The key issue is flexibility in the creation and deployment of new services by service providers that operate independent of the network operator. This can be achieved by introducing an architecture where an open network control interface and a Service Execution Environment (SEE) are provided for service applications. In this environment the service itself can control the network to set up media service sessions. A Java-based SEE is suggested for implementing such an environment. The Java SEE would provide the basic functionality (such as media stream setup /teardown, user and service authentication and bil...

Several strategies have been proposed for routing in a low earth orbit (LEO) satellite system. Some of them are based on the Internet Protocol (IP) and a few on asynchronous transfer mode (ATM) switching. However, issues like memory requirements of the satellites in the LEO network and the overheads involved in transmitting packets over the network have tended to be ignored. This paper focuses on developing an efficient solution to the routing issue in a LEO network. It also discusses the issue of the placement of the Network Operations and Control Centers (NOCCs) required to support the satellite system. Keywords: low earth orbit satellite network, intersatellite links, snapshot, stored program model, network operations & control centers. I. INTRODUCTION Of late, low earth orbit (LEO) and medium earth orbit (MEO) based satellite networks have become the focus of attention as they promise lower delays and better biterror rate performance for Internet and multi-media services than geo...

this paper. The "RSVP Multicast" model provides a multipoint-to-multipoint communication model with QoS. It enables flows to request several types of reservations with aggregation of resources and filtering. Receivers can dynamically access an RSVP session by joining the corresponding multicast groups. They can choose which packet and QoS they want by doing filtering and specifying reservations. So far, providing efficient multicast communication with QoS support is not achieved today in the IP world. On the other hand, the ATM technology was sold as a solution for providing a powerful set of QoS virtual circuits (VC) to satisfy all needs. Nevertheless, the ATM multicast scheme is quite simple: it only allows point-to-multipoint connections with explicit receivers when initiating the multicast tree. Point-tomultipoint VC QoS is identical for all receivers and defined only once. Again, multicast support with QoS is far to be supported in the ATM environment. The ubiquity of IP associated with the acknowledgment of ATM as a key switching technology has motivated an increasing interest towards the design of a more efficient way of operating IP over ATM networks. Different solutions have been proposed by the ATM Forum and the IETF (Internet Engineering Task Force):

This document presents a set of requirements for Traffic Engineering over Multiprotocol Label Switching (MPLS). It identifies the functional capabilities required to implement policies that facilitate efficient and reliable network operations in an MPLS domain. These capabilities can be used to optimize the utilization of network resources and to enhance traffic oriented performance characteristics.