Kai Y. Eng

A growable packet (ATM) switch architecture: design principles and application

The problem of designing a large high-performance, broadband packet of ATM (asynchronous transfer mode) switch is discussed. Ways to construct arbitrarily large switches out of modest-size packet switches without sacrificing overall delay/throughput performance are presented. A growable switch architecture is presented that is based on three key principles: a generalized knockout principle exploits the statistical behaviour of packet arrivals and thereby reduces the interconnect complexity, output queuing yields the best possible delay/throughput performance, and distributed intelligence in routing packets through the interconnect fabric eliminates internal path conflicts. Features of the architecture include the guarantee of first-in-first-out packet sequence, broadcast and multicast capabilities, and compatibility with variable-length packets, which avoids the need for packet-size standardization. As a broadband ISDN example, a 2048*2048 configuration with building blocks of 42*16 packet switch modules and 128*128 interconnect modules, both of which fall within existing hardware capabilities, is presented. >

A wireless broadband ad-hoc ATM local-area network

We describe the theory, design and ongoing prototyping of a wireless ATM LAN/PBX capable of supporting mobile users with multi-Mb/s access rates and multi-Gb/s aggregate capacities. Our proposed LAN Consists of network nodes called Portable Base Stations (PBS) providing microcell coverage. The PBSs are designed to be low-cost, compact and high-speed and can be relocated conveniently. We employ a concept ofad-hoc networking in the layout of the PBS-to-PBS interconnection. That is, the PBSs can be distributed in an arbitrary topology to form a backbone network and can be reconfigured with relative ease. The PBS-to-PBS backbone links are high-speed (Gb/s) for supporting high system capacity. Although they can either be wired or wireless, our emphasis is on wireless implementations. The user-to-PBS links, on the other hand, are primarily for mobile access (e.g., 2–20 Mb/s) and therefore are wireless. Wired connections from stationary users to PBSs are also possible. Typical mobile users are assumed to be laptops or notebook computers. Services supported include conventional data applications (e.g., over TCP/IP or SPX/IPX) as well as multimedia (video, voice and data) applications with QoS (Quality-of-Service) guarantees. A “wireless ATM” concept is proposed so as to provide seamless internetworking with other wired ATM local and wide-area net-works. Algorithms and control in our network are highly distributed for simple implementations and ease of mobility management. A new wireless VP/VC concept and a Homing Algorithm are described to provide ATM cell routing and connections in the network. PBS hardware and software architectures are discussed. Call management, network management and signaling are designed for simplicity, high performance and modular implementations. A fast network restoration scheme is proposed to cope with the potential link or node failures in the ad-hoc network. Error control is addressed taking the unreliable wireless links into consideration. Finally, a prototyping project called BAHAMA (Broadband Ad Hoc ATM Anywhere) for demonstrating this network concept is briefly outlined.

An efficient demand-assignment multiple access protocol for wireless packet (ATM) networks

In a wireless packet (ATM) network that supports an integrated mix of multimedia traffic, the channel access protocol needs to be designed such that mobiles share the limited communications bandwidth in an efficient manner: maximizing the utilization of the frequency spectrum and minimizing the delay experienced by mobiles. In this paper, we propose and study an efficient demand-assignment channel access protocol, which we call Distributed-Queueing Request Update Multiple Access (DQRUMA). The protocol can be used for a wide range of applications and geographic distances. Mobiles need to send requests to the base station only for packets that arrive to an empty buffer. For packets that arrive to a non-empty buffer, transmission requests are placed collision-free by piggybacking the requests with packet transmissions. The simulation results show that even with the “worst possible” traffic characteristics, the delay-throughput performance of DQRUMA is close to the best possible with any access protocol. In addition, explicit slot-by-slot announcement of the “transmit permissions” gives the base station complete control over the order in which mobiles transmit their packets. This important feature helps the base station satisfy diverse Quality-of-Service (QoS) requirements in a wireless ATM network.

Improving the performance of input-queued ATM packet switches

The authors propose a single way to dramatically improve the performance of input-queued ATM packet switches beyond the 82% saturation point obtained in previous work. The method is an extension of the independent output-port schedulers technique and is based on the notion of recycled time slots, i.e. reusing time slots normally wasted due to scheduling conflicts. In contrast to previous results, the technique yields a throughput improvement from 65% to 92% without speedup, trunking, or complicated hardware. If input grouping with a group size of four is also employed, then the method can yield up to 95% throughput.<<ETX>>

Mobility and connection management in a wireless ATM LAN

This paper proposes algorithms for handoff, location, and connection management in a wireless asynchronous transfer mode (ATM) local-area network (LAN). Fast handoffs while maintaining cell sequence and quality-of-service (QoS) guarantees are achieved by distributing switching functionality to base stations, and using a networking scheme based on provisioned virtual trees. A new distributed location management scheme using a minimal registration procedure and broadcasts on wired links is proposed for this LAN. The detailed signaling procedures that support the algorithms for mobility and connection management are described. Finally, an implementation of these procedures and an analysis of the measured data is presented. Measurements of service times obtained from this implementation indicate that over 100 calls/s. can be handled by each node in 50-node network with a high-percentage of mobiles (75%) relative to fixed endpoints. This is comparable to current wired ATM switch call handling throughputs, in spite of the fact that these nodes perform additional handoff and location management functions. The data also indicates handoff latency times of 1.3 ms. This validates our proposal for maintaining cell sequence while performing handoffs.

Channel access and interference issues in multi-code DS-CDMA wireless packet (ATM) networks

Multi-Code Direct-Sequence Code-Division-Multiple-Access (MC-CDMA) has been proposed as a flexible multiple access scheme for wireless packet networks that support a large variety of mobiles with different and even time-varying rates. Using MC-CDMA, traffic streams with significantly different transmission rates can be easily integrated into a unified architecture, with all the transmissions occupying the same bandwidth and having the same spread spectrum processing gain. In this paper, we address medium-access and interference issues in MC-CDMA wireless packet networks. For medium access, we propose and study Multi-Code CDMA (MC-CDMA) with Distributed-Queueing Request Update Multiple Access (DQRUMA) to form a unified bandwidth-on-demand fair-sharing platform for multi-rate wireless services. DQRUMA is an efficient demand-assignment multiple access protocol for wireless access and scheduling. Pseudo-Noise (PN) codes (primary codes) and optimal power levels are allocated to the mobiles on a slot-by-slot basis, and a Maximum Capacity Power Allocation (MCPA) criterion exploits the sub-code concatenation property of the MC-CDMA transmission. Simulation results show that the system provides close to ideal-access performance for multi-rate mobiles, both with homogeneous traffic characteristics and with a mix of heterogeneous traffic characteristics. Finally, we analyze the effects of MC-CDMA intercell interference on the reverse link (i.e., mobile to cell site) and investigate interference reduction by using the Maximum Capacity Power Allocation (MCPA) criterion. Our results show significant reduction in reverse-link MC-CDMA intercell interference is possible using the MCPA criterion.

A growable packet (ATM) switch architecture: design principles and applications

The authors consider the generic problem of designing a large N*N(N>1000) high-performance, broadband packet (or asynchronous transfer mode) switch. They provide ways to construct arbitrarily large switches out of modest-size packet switches, without sacrificing overall delay/throughput performance. They propose and study a growable switch architecture based on three key principles: (a) a generalized knockout principle which exploits the statistical behavior of packet arrivals and thereby reduces the interconnect complexity; (b) output queuing, which yields the best possible delay/throughput performance; and (c) distributed intelligence in routing packets through the interconnect fabric. Other features include the guarantee of a first-in first-out packet sequence, broadcast and multicast capabilities, and compatibility with variable-length packets. In a broadband ISDN (integrated services digital network) example, the authors show a 2048*2048 switch configuration with building blocks of 42*16 packet switch modules and 128*128 interconnect modules.<<ETX>>

Fundamental Conditions Governing TDM Switching Assignments in Terrestrial and Satellite Networks

We consider the traffic handling capability of a new switching architecture which generalizes upon the structure of a traditional time multiplex switching system. For the traditional approach, lowbandwidth end users are formed into groups, each of which shares a single high bandwidth time-division multiplexed (TDM) line into the central time multiplex switch. Each user synchronously generates packets of data in preassigned time slots at a rate consistent with its offered traffic, and the feeder for the TDM line serving a group of users merely time multiplexes the arriving packets prior to routing by the central switch; analogously, each output port of the switch feeds a demultiplexer which routes the packets to the appropriate user within its group. The generalized approach permits each user group to share some multitude of TDM lines interconnecting that group with the central switch, and the group multiplexers and demultiplexers are replaced by switches which route packets from users to TDM lines (and vice versa). For this structure, we derive a set of necessary and sufficient conditions on the user-touser offered traffic such that a valid, nonconflicting TDM assignment of packets-to-time slots exists. These conditions reveal that the constraints imposed by the three-tiered switching hierarchy do not limit the useable capacity of the switch. Consequently, with no loss of traffic bearing efficiency, it is possible to reduce the number of multiplexers used to serve the end-user population, achieve greater trunking efficiency since small user groups served by one TDM line are replaced by larger groups serving multiple lines, and modularly grow the system by adding TDM lines to each group commensurate with the traffic offered by that group. As a byproduct, it is shown that such a system designed to switch low-speed circuits of some particular data rate can, with no hardware change, switch circuits at lower rates (subrate switching). These conclusions have important ramifications for the design of terrestrial and satellite-based switching systems.

Performance of hierarchical multiplexing in ATM switch design

The authors study the delay-throughput characteristics and the buffer requirements associated with hierarchical multiplexing for random and bursty traffic models. The interface data rates to the external lines can be and usually are different from the internal core fabric speed of an asynchronous transfer mode (ATM) switch. As signals are multiplexed inside the switch to higher speeds, the required dimension of the core fabric is reduced, leading to a reduction in physical size, easing of input/output constraints, simpler control, and improved hardware efficiency. The simulation results, for random and bursty traffic models, indicate that this hierarchical multiplexing technique has little impact on the delay-throughput performance. The results also show that a small degree of multiplexing dramatically increases the buffer requirements. However, further multiplexing reduces the amount of buffering down to more acceptable levels.<<ETX>>

A framework for a national broadband (ATM/B-ISDN) network

The authors propose a hybrid packet- and circuit-switched network as a framework for a national broadband (ATM/B-ISDN) (asynchronous transfer mode/broadband-integrated services digital network). Built on the observation that transmission speeds are likely to remain much faster than switching speeds (thus packet switching at peak transmission rates needs to be avoided), the network architecture is a three-tier hierarchy composed of LANs (local area networks), network nodes and DACS (digital access and cross-connect systems). Access to the network is either through direct connection to a DACS or network node, or through the end-users LAN, which has a gateway to a network node. Each network node is a high-performance ATM packet switch, which accepts input cells at a B-ISDN rate of 150 Mb/s and serves both as LAN-to-LAN interconnect and as a packet concentrator for traffic destined to other network nodes and LANs. To minimize the delay and simplify the implementation of gigabit-per-second packet switching, the network nodes are interconnected by a backbone network of multi-gigabit-per-second fibers and DACS, which provide reconfigurable circuits between network nodes. A congestion control technique using channel sharing that dramatically reduces the required buffer size is proposed.<<ETX>>