Optik lif ortamında dağılmış veri arabağı modellemesi

Abstract

Bu tez Uç ana bölümden oluşmaktadır : Genel bilgisayar ağları kavramları, Optik Lif Ortamında Dağılmış Veri Arabağı (FDDI) standartlarının incelenmesi ve FDDI tasarım çalışması. İlk bölümünde genel bilgisayar ağları kavramlarından söz edilmektedir. Bu bölümde tüm bilgisayar ağları için geçerli olan ortak terimler sıralanmış, OSI modeli yeni hatlarıyla belirlenmiş ve son olarak yerel bilgisayar ağlarından söz edilmiştir. ikinci bölüm FDDI konusuna giriş bilgilerini içermektedir. Bu bölümde önce FDDI ağlarına ilişkin tanımlar ve kullanıcı açısından özellikleri, daha sonra ise FDDI protokoller ile gelen ana kavramlar verilmektedir. üç, dört ve beşinci bölümlerde ise ANSI ve OSI standartları yardımıyla, FDDI konusundaki protokoller incelenmekte, protokollarda verilmesi gereken tanımlar ve kavramlar üzerinde durulmaktadır. Altıncı bölümde bir FDDI istasyonunun yürütmesi gereken diğer protokollerden bahsedilirken, yedinci bölümde FDDI protokollerinin SDL diyagramları ile modellenmesi verilmiştir. Son ana bölümü içeren sekizinci bölümde SUPERNET yonga seti tanıtılmakta ve bu set ile gerçeklenebilecek bir FDDI arabirim tasarımına ilişkin çalışmalar sunulmaktadır. Ek bölümlerde ise diğer bölümlerde kullanılan CRC kodlaması ve SDL diyagramlarına ilişkin bilgiler verilmektedir.

The Fiber Distributed Data Interface (FDDI), is a proposed American National Standard for a i 00 lib/s token ring using an optical fiber medium. FDDI was originally proposed as a packet switching network with two primary areas of applications first, as a high performance interconnection among mainframes, and among mainframes and their associated mass storage subsystems and other peripheral equipment and second, as a backbone network for use with lower speed LAN's such as IEEE 802.3, 802.4, and 802.5. An enhancement to FDDI adds a circuit switching capability, expanding the field of application of FDDI to include those requiring the integration of voice, video, and sensor data streams. FDDI will use optical fiber with Light [Emitting Diodes (LED's) transmitting at a nominal wave length of İ3C0 nanometers. Connections between stations will be made with a dual fiber cable employing a polarized duplex connector. The data transmission rate is 100 Mb/s" The four out of five code used on the optical fiber medium, requires a 125 megabaud transmission rate. The nature of the clocking limits frames to 4500 octets maximum. Multiple frames may however be transmitted on the same access opportunity. A total of 1000 physical connections and a total fiber path of 200 km has been used as the basis for calculation of the default values of the recovery times. Considering reconfiguration requirements, these choices allow a maximum configuration of 500 stations (each station represents two physical connections) linked by 100 km of cable. There is no minimum configuration requirement. FDDI-II is initialised identically to the basic FDDI and is fully interoperable with it prior to the assignment of any isochronous channels. The presence of any nan FDDI-II capable stations in a ring prevents the assignment of any isochronous channels and thus the activation of the FDDI-II mode of operation. All stations making a direct physical attachment to the FDDI ring provide connection to two separate counter rotating rings. Such connections are referred vi 1 1 to as Class A and require two duple;-; cables, one to each of the adjacent stations- Simpler connections, referred to as Class B, require only one duple;-; cable, but may only be made through a wiring concentrator which itself is connected to the ring via a Class A connection» FDDI uses three techniques for improving rel iab i 1 i ty « A Station Bypass Switch with the capability to bypass any station is specified. This solves the problem of known broken or powered down stations. Counter Rotating Ring Connections are required of all stations directly attached in the ring» The second ring may be only a standby ring, or it may be used for concurrent transmission» In either case, if a station fails, or a link fails, the two rings are folded into one ring, which is appro;; i ma tel y twice as long, maintaining full connectivity» Concentrators may be used to attach stations to the ring. Each station has a direct link to the concentrator. This allows any combination of stations to be switched out of the ring while retaining full connectivity for the remaining stations» The use of all three techniques allows FDDI networks to be configured to tolerate a variety of station or link failures and physical network configurations without any catastrophic consequences. When any of these occurs, the network automatically reconfigures, thus eliminating any failing element and maintaining ring operation. Continous monitoring of the failed link, or station, allows the network to automatically reconfigure and restore normal operation when repair is effected» Any of these reconfigurations may result in the loss of individual frames» Information on the medium is in a 4 to 5 group code with each group being called a symbol» Of the 32 member symbol set, 16 are data symbols each representing 4 bits of ordered binary data, 3 are used for starting and ending delimiters, 2 are used as control indicators, and 3 are used for line-state signalling which is recognised by the physical layer hardware. The remaining 8 symbols of the symbol set are not used since they violate code run length and DC balance requirements. Note that QUITE!, the symbol which repredents no line activty, is a necessary member of the line-state symbol set since it is used to detect the absence of a functioning link. Component entities for an FrDDI station conforms to both the IEEE 802 structure and the OSI concept of - i ;< ~ Layering. These ares Station Management (SMT), a part of Network Management which resides in each station to control the overall action of the station to insure proper operation as a member of the ring. Physical Protocol (PHY) and tledium Dependent (PMD) which provide the hardware connection to an adjacent station. - Media Access Control (MAC), which controls access to t h e med i a, t r an sm i 1 1 i n g f r ames to, and receiving frames from the MAC ' s of other stations. While the actual FDDI document structure includes a Physical Medium Dependent (PMD) document dealing with the actual optical fiber hardware components, for the sake of simplicity, PMD is considered part of PHY herein. The FDDI MAC is being specified to be compatible with what is a superset of the LCC developed by IEEE 802» PHY provides the protocols and optical fiber hardware components that support a link from one FDDI station to another. PHY simultaneously receives and transmits. The transmitter accepts symbols from MAC, converts these to five- bit code groups, and transmits the encoded serial data stream on the medium. The receiver receives the encoded serial data stream from the medium, establishes symbol boundaries based on the recognition of a Start Delimeter, and forwards decoded symbols to MAC. Additional symbols (QUITE, IDLE, and HALT) are interpreted by PHY and used to support SMT fuctions. A major function of any station is deciding which station has control of the medium and what is to be placed on that medium. Thus, MAC schedules and performs data transfers on the ring. The basic concept of a ring is that each station repeats the frame it receives to its downstream neighbour. If the destination Address (DA) of the frame matches that MAC ' s address, then the frame is copied into a local buffer and MAC notifies LLC (or SMT) of the frame's arrival. MAC marks any frame, by setting indicator symbols in the FS field of a frame, to indicate the recognition of its address, the copying of the frame, or the detection of a frame in error. The frame propagates around the ring to the station that originally placed it on the ring. The transmitting station examines the indicator symbols in the FS field to determine the success of the transmission. The MAC of this transmitting station is responsible for removing from the ring all of the frames that it has placed on the ring (a process termed 'stripping'). MAC recognises these by the fact that the Source Address (SA) contained in them is its own address, IDLE symbols are placed on the medium d u ring s t r i p ^ p i n g. If MAC has a frame from LCC (or SMT) to transmit then it may only do so after a token has been captured- A token is a special sequence which indicates that the medium is available for use» Priority requirements, necessary to assure the proper handling of frames, are implemented in the rules of token capture. Under these rules, if this station is not a\l lowed to capture the token then it must repeat it (or in certain cases, reissue a token) to the next station. Having captured the token, removing it from the ring, MAC transmits a frame, or frames, and when done issues a new token, as a notification that the medium is available for use by another station. SMT monitors activity and exercises overall control of station activity. These functions include control and management within a station for such purposes as initialization, activation, performance monitoring, maintenance, and error control. Additionally, SMT communicates with other SMT entities on the network for the purpose of controlling network operation. Examples of these SMT functions include the administration of addressing, allocation of network bandwidth, and network control and c: on f i gura t i on. The FDDI as developed in ASC X3T9 will satisfy a variety of requirements. It will be an efficient backbone network suitable for use with the IEEE 802 and other lower speed LANs. As a back end network it will satisfy the varied requirements of interconnections among high speed digital processors and their shared high speed peripheral subsystems, communication systems and workstations. Both packet and circuit switching modes of operation will be supplied, thus supporting a wide range of applications requiring the integration of voice, video, sensor data streams, and packet data in a distributed processing environment » Chapter 8 is concerned with detailing the functions and characteristics of the set of semiconductor devices known collectively as the SUPERNET family which AMD has developed to provide a cost-effective implementation of the FDDI standard. The 5-chip family meets the proposed standard and offers a variety of additional system features. Features of SUPERNET may be summarised as fol lows? İ-- 200 Mb i t /sec buffer memory bandwidth 2- Serial interface to fiber optic tranceiver 3- Up to 64K long word buffer memory address range 4- Total buffer memory management 5- Parity path from buffer memory to encoder/decoder (ENDEC) section and vice versa» Parity generation / checking (32 bit data, 4 bit parity) 6- Error/Status updates and interrupt reporting via SUPERNET internal registers and receive frames stored in buffer memory. 7~ Diagnostic capabilities for the medium and system interface. 8- Implements the media access control layer and the physical layer protocols of ANSI Standard X3T9"5 FDDI. A key element of any FDDI station is the buffer management. In SUPERNET, this is partitioned into two devices? the 79C82A RAM buffer controller (RBC) and the 79C82A data path controller (DPC). The RBC provides DMA channels and arbitrates access to the network buffer memory. The DPC controls the data path between the buffer memory and the medium. Functions specific to the FDDI MAC Layer ai^e performed by the 79C83 fiber optic ring media access controller (FORMAC). The tasks of the physical layer arts performed by a two-chip encoder/decoder pairs the 7984A encoder/decoder (ENDEC) and the 7985A ENDEC data separator (EDS). Connection to the actual medium is performed by an fiber optic tranceiver (FOX) electronics external to the SUPERNET family.