Implementation aspects of ATM switches
نویسنده
چکیده
Since in the future, the demand on broadband services (e.g. Lan and video interconnection services) is likely to increase, research is focusing on the introduction of the Broadband ISDN. In 1989 standards on the transfer mode for B-ISDN were introduced: ATM. From then a lot of ATM switch designs were introduced, but they all assume the VLSI technology to be more developed than it actually is. This report describes which modifications should be performed on the switch designs to make them better implementable. This includes a detailed examination of general ATM switch designs, an examination of Delta, Clos and Benes networks and an overview of actual introduced ATM switch designs. Since for complex circuits maximum on-chip frequency is currently limited to about 50MHz (off-chip 25MHz) and ATM traffic has a speed of 155Mbit/s or more, parallellisation has to be performed to be able to reduce speed. Bitparallellisation means a cell is divided in several bits wide packets. It has the advantages of having only a small speed adaptation buffer delay and of guaranteeing preservation of cell sequence integrity. Cellparallellisation means cells are divided over switch copies, which reduces the load per switch copy. This has the advantage of having more evenly spread traffic. Since large switch designs can't be put on one chip and because they have to be modular, they have to be partitioned. Slicing means the large switch design (with parallellisation) is cut in a way that one copy of a switch which results from the parallellisation, is kept in one piece, i.e. the switch is divided in planes. Slicing can only be performed after parallellisation is performed, and results in all equal parts. Logical partitioning means the switch design is cut in a way that one part is a part of an ASE, one ASE or several ASE's. Logical partitioning is needed when slicing on its own is not sufficient and leads to several bits wide parts. This might be all equal parts (depending on the design). How the partitioning should be performed depends on the switch design. Small ASE's (size 2x2 or 4x4) don't have to be partitioned themselves. The MIN's these ASE's are used in, are based on Delta networks. Of these networks, the Batcher-Banyan network should be logical partitioned in a way that one part is one stage of ASE's. The other small ASE based networks should be partitioned in a way that one part is a stage of larger (than 2x2) networks. Limited size ASE's are ASE's which can't grow beyond some size, because of limitations in technology. These ASE's should be sliced, to keep high (size dependable) speed links on-chip. Unlimited size ASE's can grow without meeting limitations in technology. Because of their large complexity and their regular design, they should be logical partitioned. Both limited size and unlimited size ASE's are used in folded and unfolded Clos and Benes networks. For Benes networks the unfolded network is preferred because it uses less chips. For the Clos networks, the folded network with all equal ASE' s is preferred, because although it uses more chips than other networks, it only uses one kind of chip. Implementation aspects of ATM switches
منابع مشابه
An Efficient Location Management Scheme for Hybrid Wireless ATM Networks
Location management is one of the most important aspects in the realization of mobility in wireless networks. Following developments of ATM (Asynchronous Transfer Mode) technology, more core networks become ATM-based. Therefore, the demand of mobility on wireless networks with ATM backbones is increasing. In recent years, several location management schemes have been proposed for classical wire...
متن کاملDesign and Implementation of a Fault Tolerant ATM Switch
In this paper, we propose a new method to build a fault tolerant ATM switch. Using this method, we can build an ATM switch which has two disjoint paths between each input/ output pair. The key component in the proposed switch is a 2 ¥ 2 FTSE (Fault Tolerant Switching Element), which can be used as the basic building block for high speed ATM switches. The design of the FTSE-based fault tolerant ...
متن کاملMethodology for implementing scalable test configurations in ATM switches
In performance testing of ATM switches and network of switches a variety of connection configurations is needed. In most of the cases, these configurations require one traffic generator and/or analyzer for each switch port. Since this equipment is rather expensive, it is desirable to define scalable configurations that can be used with a limited number of generators. In this paper we present a ...
متن کاملEfficient Bandwidth Allocation and Call Admission Control for VBR Service Using UPC Parameters
Provision of Quality-of-Service (QoS) guarantees is an important and challenging issue in the design of Asynchrous Transfer Mode (ATM) networks. Call Admission Control (CAC) is an integral part of the challenge and is closely related to other aspects of network design such as traÆc characterization and QoS speci cation. Since the Usage Parameter Control (UPC) parameters are the only standardize...
متن کاملImplementation of Virtual LANs over ATM WANs
Virtual LANs (VLANs) allow to interconnect users over campus or wide area networks and gives the users the impression as they would be connected to the same local area network (LAN). The implementation of VLANs is based on ATM Forum’s LAN Emulation and LAN/ATM switches providing interconnection of emulated LANs over ATM and the LAN ports to which the user’s end systems are attached to. The pape...
متن کاملDynamic multipath routing in networks and switches carrying connection-oriented traffic
We propose a dynamic multipath routing (DMPR) scheme to improve resource utilization of a network for a given quality of service (QoS) in carrying delay sensitive traffic. The proposed scheme takes advantage of available alternate routes which can be used beneficially in communication networks or switches, without increasing the network latency. Through simulation study of circuit-switched, vir...
متن کامل