Given by Roman Markowski at Lectures at Xi'an Jaotong University on Sept 1998. Foils prepared Dec. 6 98
Outside Index
Summary of Material
Describes in detail 6 different approaches to LANs |
Ethernet, IsoEthernet, Fast and Giga Ethernet |
Token ring and Token Bus |
100 VG -- AnyLAN |
FDDI/CDDI |
Wireless Networks |
ATM for LANs |
Outside Index
Summary of Material
Roman Markowski |
IS Manager |
Northeast Parallel Architectures Center |
September 1998 |
http://www.npac.syr.edu/users/roman/ |
Ethernet, IsoEthernet, Fast, Giga |
Token Ring, Token Bus |
100 VG- AnyLAN |
FDDI / CDDI |
Wireless |
ATM |
High-speed, fault-tolerant data network that covers a relatively small geographic area |
LAN protocols function at the 2 lowest layers of OSI RM (Physical and Data Link Layers) |
Typical access methods: CSMA/CD (carrier sense multiple access collision detection) and Token Passing |
Transmission methods: unicast, multicast and broadcast; in these methods a single packet is sent to one or more nodes |
LAN topologies: bus, star, ring |
Devices: repeaters, hubs, bridges, switches, routers |
LAN - Local Area Network. A network that interconnects PCs, terminals, workstations, servers, printers, and other peripherals at high speed over short distances (ethernet, token ring, FDDI, ATM, wireless) |
WAN - Wide Area Network. A network that covers long-haul areas and usually utilizes public phone companies (T1/T3, ISDN, xDSL, Frame Relay, SMDS, ATM) |
Wireless (802.11) 2-10 Mbps |
Ethernet (802.3) 10 Mbps |
Iso Enet (802.9) 16 Mbps |
Token Bus/Ring (802.4,802.5) 4, 16 Mbps |
100VG-AnyLAN (802.12) 96 Mbps |
Fast Ethernet (802.3u) 100 Mbps |
FDDI (X3T9.5) 100 Mbps |
Fiber Channel 133, 266, 531, 1062 Mbps |
Giga Ethernet (802.3z) 1000 Mbps |
ATM 25, 100, 140, 155, 622 Mbps |
HiPPI 800, 1600, 6400 Mbps |
This term refers to family of Local Area Network implementations
|
Characteristics
|
standard: IEEE 802.3 |
media: coaxial cable, twisted pair, fiber, microwave |
media access: CSMA/CD (Carrier Sense Multiple Access/Collision Detection) |
bandwidth: 10 Mbps, full duplex-20Mbps |
topology: bus, star |
frame size: 64 - 1500 Bytes |
max nodes: 1024 |
max interstation distance: fiber - 2 km; twisted pair - 70 m; coax - 1.5 km; microwave - 7.2 km |
max coverage: 2.8 km |
connection between hubs: crossover cable |
frame formats: Ethernet_II, Ethernet_802.3, Ethernet_802.2 |
Differences between Ethernet and IEEE 802.3
|
IEEE 802.3 Physical Layer (i.e. 10base5)
|
10base5
|
10base2
|
10baseT
|
10baseF
|
AUI - attachment unit interface; MAU - media attachment unit |
By John Mazza http://jmazza.shillsdata.com/tech/ |
By John Mazza http://jmazza.shillsdata.com/tech/ |
By John Mazza http://jmazza.shillsdata.com/tech/ |
specification: IEEE 802.9 (it is not Ethernet!) |
isoENET is a hybrid of a normal 10baseT Ethernet and up to 96x 64 Kbps communication channels |
speed: 16.144 Mbps (10 Mbps - Ethernet data and isochronous 6.144 Mbps- time sensitive applications) |
access method: CSMA/CD |
applications: integrated voice and data networks |
characteristics: backward-compatible relationship between isoENET and Ethernet; |
topology: star (requires special isoENET hub and special NIC cards) |
wiring : twisted pair |
specification: IEEE 802.3u |
speed: 100 Mbps |
access method: CSMA/CD; |
topology: star; applications: enterprise servers, backbone |
frame size : 64-1500 Bytes |
Three media types: 100base-TX, 100base-FX, 100base-T4 |
100base-TX
|
100base-FX
|
100base-T4
|
100VG-AnyLAN
|
Ethernet Hub |
10 Mbps Segments |
100 Mbps Backbone |
specification: IEEE 802.3z |
speed: 1 Gbps |
access method: CSMA/CD |
frame size: 64-1500 bytes |
flow control: 802.3x |
applications: enterprise servers |
implementation: intelligent adapters that minimize host involvement |
segment length:
|
1000base-LX
|
1000base-SX
|
1000base-CX
|
1000base-T
|
The Gigabit ethernet standard adopts both the original IEEE 802.3 ethernet specifications and Fiber Channel ANSI X3T11 for physical interface. Gigabit ethernet retains ethernet's link layer protocol, maximum frame size, and frame format. Gigabit includes both full and half-duplex operating modes |
standard: IEEE 802.5; developed by IBM; very popular |
media: optical fiber, twisted pair STP or UTP |
bandwidth: 4, 16 Mbps; also 100 Mbps High-Speed Toke Ring (IBM) and Gigabit Token Ring 1Gbps (Cisco) |
max frame size: (4 Mbps) no limit; (16 Mbps) 18,000 Bytes |
max nodes: 260 for STP and 72 for UTP |
max interstation distance: 300 m to MSAU |
topology: physical star, logical ring |
hardware:
|
By John Mazza http://jmazza.shillsdata.com/tech/ |
media access: Token Passing - information is traveling from station to station and is regenerated by each station; no collisions; possession of the token grants the right to transmit data |
priority system: token ring networks use a sophisticated priority system that permits certain users and high priority stations to use network more frequently (frame contains priority field and reservation field) |
One selected station works as active monitor: it acts as a centralized source of timing information; removal of faulty frames circling the ring |
beaconing algorithm detects and tries to repair certain network faults |
frames:
|
standard: IEEE 802.4 |
media: 75 ohm CATV coaxial cable |
topology: bus and star |
media access: Token Passing - information is traveling from station to station and is regenerated by each station; the sequence is defined by station addresses |
bandwidth: up to 10 Mbps |
100VG AnyLAN is a 100 Mbps high speed networking standard that was originally developed to transmit Ethernet or Token Ring packets over existing wiring (UTP cat 3,4,5 or fiber); |
VG stands for Voice Grade |
standard: IEEE 802.12, developed by Hewlett-Packard |
actual throughput: up to 96 Mbps |
frames: standard 802.3 ethernet format or 802.5 Token Ring format |
distance (hub to node, or hub to hub): UTP cat3 - 100m; UTP cat5 - 200m; fiber - 2000m |
access method: Demand Priority (MAC layer protocol), no collisions |
compatible with existing software; easy to implement |
number of nodes: up to 1024 unbridged nodes |
ideal for multimedia applications (traffic can be prioritized) |
comparison between CSMA/CD and Demand Priority:
|
topology: cascade to 3 levels of hubs (total network span=1200m) |
hardware: NICs PCI-VG10/100, EISA-VG10/100 , ISA-VG10/100 with 2 RJ45 (one for ethernet, one for VG); VG-HUB with RJ45 VG ports, HUB with ethernet-to-100VG Bridge |
vendor: Ragula Systems, Cisco, Cabletron, Hewlett Packard, etc |
By John Mazza http://jmazza.shillsdata.com/tech/ |
FDDI is an excellent medium for building backbones of LANs. |
standard: ANSI X3T9.5 developed by ANSI in mid-1980s |
media: optical fiber (FDDI), copper (CDDI) |
media access: Timed Token Passing Dual Ring |
bandwidth: 100 Mbps (FDDI network actually has a 125 Mbps signaling rate, however 4B/5B (Differential Manchester) encoding takes up to 25 Mbps) |
max frame size: 4500 Bytes; max nodes: 500 |
max inter-station distance: 2 km; max distance: 100 km (60 miles) |
topology: dual ring (for redundancy); during normal operation, the primary ring is used for data transmission, the secondary ring remains idle; traffic on each ring flows in opposite directions (called counter-rotating);stations act as repeaters |
DAS - Dual Attached Station is attached to both rings |
SAS - Single Attached Station is attached only to the primary ring |
connectors: ST, MIC, SMA906; status: mature, reliable and proven |
transition modes: asynchronous - traffic is not prioritized; synchronous - allows for prioritization of time-sensitive traffic; circuit-based (in FDDI II only) -allows for dedicated communication line with guaranteed bandwidth |
FDDI-II is designed for networks that need to transport real-time video or traffic that cannot tolerate delays. Bandwidth is divided into up to 16 separate circuits that operate at from 6.144 Mbps each to a maximum 99.072 Mbps. Each of these channels can be subdivided further to produce a total of 96 64- Kbps circuits. |
FDDI-II is incompatible with existing FDDI design. |
CDDI - Copper Distributed Data Interface (officially named: Twisted-Pair Physical Medium Dependent standard)
|
Fiber optic cable
|
FDDI Specifications
|
FDDI fault tolerance
|
standard: IEEE 802.11 |
types of implementation:
|
OSI Physical layer (Access Point, antennas, LAN interfaces)
|
OSI Data Link layer:
|
typical parameters:
|
topologies: point-to-point, point-to-mulitipoint |
Standards to transmit TCP/IP over cellular connection
|
Wireless services:
|
Fundamentals: ATM stands for Asynchronous Transfer Mode.ATM technology is viable for both local (LANs) and wide (WANs) area networks. It supports many types of traffic including voice, data, facsimile, real-time video, CD-quality audio, and imaging. ATM offers bandwidth of up to 155Mbps (and now 622 Mbps or 2.4 Gbps between switches) on the backbone and delivery pipes of 25 Mbps or more to the desktop. ATM is a transport protocol that operates roughly at the MAC sub-layer of the Data Link Layer (OSI) |
Characteristics: |
fixed size packets (cells) - 53 bytes |
switching technology |
high, scalable bandwidth |
ability to dedicate a certain bandwidth to an application (QoS - Quality of Service) |
can handle real-time applications |
broadband technology (for voice, video, and data) |
viable for LANs and WANs |
still the most expensive service |
ATM network consists of ATM switches |
wireless ATM (28 GHz range), SMDS over ATM, Frame Relay over ATM |
ATM switch simply relays cells. It looks at the header and immediately begins forwarding the cell. Because all the cells are the same size, the delay is minimal. ATM switch performs its switching in hardware.No error checking is done. ATM assumes the use of high-quality, error free transmission facilities. |
ATM Architecture: ATM was originally defined as part of B-ISDN, developed by CCITT in 1988. ATM Forum, created in 1991, has been working on development of ATM standards. |
Physical Interface methods
|
Logical connection between end-stations
|
VPI |
VCI |
VCI |
Permanent Virtual Channels and Switched Virtual Channels
|
ATM virtual path switching |
ATM architecture model |
ATM Physical Layer: |
medium:
|
physical interface defined by ATM Forum:
|
SONET / SDH - Synchronous Optical Network / Synchronous Digital Hierarchy |
physical layer standard for fiber optic transmissions standardized by ANSI and recommended by CCITT |
SONET is the carrier format used in the USA in which OC3c corresponds to a 155 Mbps |
SDH is a European carrier system in which STM-1 corresponds to 155 Mbps |
ATM and SMDS operates on top of SONET |
interface rates: OC1, OC3, OC9, OC12, OC18, OC24, OC36, OC48
|
fiber based or microwave based (OC3, 6 GHz) |
framing: STS-1 frame = 810 bytes; STS-3c frame = 1890 bytes |
ATM Layer defines:
|
ATM Adaptation Layer:
|
ATM multilayer architecture |
Quality of Service (QoS) classes and Traffic Classification
|
ATM addressing and signaling
|
Classical IP-Over-ATM (IPOA RFC 1577) |
ATM is a connection-oriented protocol, which means that connection must be established between two communicating entities before data transfer can begin. IP is inherently connectionless. There are 2 ways to run transparently IP traffic over ATM: Classical IP and LAN emulation |
The term "classical" indicates that the ATM network has the same properties as existing legacy LANs. |
MTU (Maximum Transmit Unit) = 9180 bytes + 8 bytes LLC/SNAP header |
SVC management is performed via UNI specification, which defines signaling: Q.93B (UNI3.0), Q.2931 (UNI3.1). |
Once a Classical IP connection has been established, IP datagrams are encapsulated using IEEE 802.2 LLC/SNAP and are segmented into ATM cells using AAL5 (RFC 1583) |
Classical IP-Over-ATM (IPOA RFC 1577) |
There is no support for broadcast and multicast in a Classical IP environment. The Classical network is divided into LISs (Logical IP Subnets). |
Each LIS must have only one ATM ARP server configured on a switch or a host. ARP server is responsible for resolving an IP addresses into ATM addresses. |
MARS - Multicast address resolution server (central registry for IP multicast membership) |
ATM |
switch |
ATMARP Server |
IP to ATM mapping |
IP address |
ARM address |
IP address |
ARM address |
Client 1 |
Client 2 |
LAN Emulation: IP-Over-ATM (IPOA LANE 1.0) |
LANE over ATM is the natural path for migrating ethernet and token ring LANs to fault tolerant ATM network. LANE ver1.0 emulates MAC layer ; standard 1996 |
Applications can run unmodified over an ATM network. LANE works by allowing the OS and all protocols at and above Layer 2 to seamlessly operate with ATM |
components: LES, LEC, LECS, BUS |
version 1.0 permits only one LES / BUS on an emulated network creating single point of failure |
LECS - LAN Emulation Configuration Server contains the configuration of the emulated networks |
LES - LAN Emulation Server mapping between MAC addresses and ATM addresses |
BUS - Broadcast and Unknown Server handles broadcast and multicast as well as unknown addresses |
LEC - LAN Emulation Client runs on every workstation |
Emulated LAN operates in client/server environment applying Virtual LAN concept |
LAN Emulation protocol stack |
LANE vs. Classical |
Voice over ATM (VTOA - Voice and Telephone over ATM) |
QoS parameters: delay, jitter in delay, loss rate |
CBR for voice current ATM Forum specification |
silence detection (voice has 60% of silence) |
unstructured circuit emulation: maps entire T1 circuit to a single ATM VC |
structured circuit emulation: maps individual 64 Kbps circuits to ATM VCs |
Compression: G.726 ADPCM, G.728 LD-CELP, G.729 CS-ACELP |
average national network delay for voice traffic specified by G.114, G.131, G.126 |
current VTOA: lack of compression, lack of silence suppression, limitations of AAL1 |
Other ATM Protocols |
MPOA - Multi Protocol Over ATM : layer 3 protocols like IP or IPX can be carried transparently over an ATM network
|
NHRP - Next Hop Reservation Protocol is used for forwarding and ATM address resolution based on classical IP |
CIF - Cells in Frames; solution for delivering ATM to the desktop |
IISP - Interim Inter-switch Signaling Protocol - uses UNI-based signaling for switch to switch communication |
Other ATM Protocols |
PNNI - Private Network to Network Interface: the NNI used in private networks for efficient, dynamic, and scalable routing of SVC requests in a multivendor private ATM environment |
I-PNNI- Integrated PNNI: is a single routing protocol that is used between IP routers and ATM switches |
Hardware: ATM switch, ATM analyzer, ATM adapter |
ATM Carrier Services: MCI, WorldCom, Sprint, AT&T, GTE, Pacific Bell, US West, Ameritech, Southwestern Bell, Bell Atlantic. The services are provided by hundreds POPs (Point of Presence) |