Transcript ppt file - University of New Mexico

LAN Technologies
•
•
•
•
Ethernet
Token Ring
FDDI
Gigabit Ethernet
© Copyright 1997, The University of New Mexico
4-1
IEEE/ISO Protocol Set
• Deal with Physical and Data Link layers only
• 3 medium access control (MAC) standards
– IEEE 802.3 CSMA/CD; ISO 8802.3
– IEEE 802.4 Token bus; ISO 8802.4
– IEEE 802.5 Token ring; ISO 8802.5
Logical Link Control (LLC)
Layer 2
Medium Access Control (MAC)
Physical
© Copyright 1997, The University of New Mexico
Layer 1
4-2
Ethernet
•
•
•
•
•
•
•
Developed in early 1970’s Xerox PARC
Standardized as IEEE 802.3
Provides data communications for LAN systems
Baseband technology developed by Xerox, Intel, Dec
Broadband version developed by Mitre
Frame based transmission
Most commonly used medium access control
© Copyright 1997, The University of New Mexico
4-3
Ethernet
• CSMA/CD
– Carrier Sense Multiple Access with Collision Detection
•
•
•
Nodes listen before transmitting
Nodes may transmit at any time
Collisions result in back-off and retransmission
– random access contention based media access
• stations contend for access to the medium
© Copyright 1997, The University of New Mexico
4-4
CSMA/CD
B
A
C
Bus
If too many collisions or other errors
(CRC alignment, Jabber, Giant frames)
occur on a given port, hubs will auto
partition that port to prevent it from
repeating on the entire segment.
A
B
C
A
B
C
Collision
© Copyright 1997, The University of New Mexico
4-5
Ethernet
• MAC frame: 64 bytes min
IEEE 802.3
7
Preamble
1
SDF
2 or 6
DA
2 or 6
2
SA
Length
>/ 0
LLC data
>/ 0
Pad
bytes
4
FCS
Ethernet
7
Preamble
1
SDF
2 or 6
DA
2 or 6
2
SA
Type
© Copyright 1997, The University of New Mexico
>/ 0
LLC data
>/ 0
Pad
bytes
4
FCS
4-6
Ethernet
• IEEE 803.2 MAC frame
– Preamble: 7 byte pattern of alternating 0s and 1s used by
the receiver to establish bit synchronization
– Start frame delimiter (SDF): sequence 10101011 indicating
start of frame, enables receiver to establish first bit of
frame
– Destination address (DA): specifies the stations for which
the frame is intended ( 2 or 6 bytes, implementation
decision, usually 6 bytes)
• it may be one of the following
– a unique physical address
– aUniversity
group address
© Copyright 1997, The
of New Mexico
4-7
Ethernet
• Source address (SA): specifies the station that sent
the frame (2 or 6 bytes, usually 6 bytes)
• Length: length of the LLC data field
• LLC data: data unit supplied by LLC, less than 1518
bytes
• Pad: octets added to ensure the frame is long enough
for the proper collision detection operation (optional)
• Frame check sequence (FCS): 32-bit cyclic
redundancy check (CRC); except preamble, SDF and
FCS
4-8
© Copyright 1997, The University of New Mexico
Ethernet
• Ethernet v2
– Type: deals with protocols
• IP protocol, Type = 0800
• No two machines should have the same physical
address
• LSB of first octet
– if 0 then unique address
– if 1 then multicast address
© Copyright 1997, The University of New Mexico
4-9
Ethernet
• 10Mbps transmission
– Manchester Encoding
• Baseband 20MHz
• Network Interface Cards
– Each node has a NIC
– NIC is responsible for transmitting & receiving data
frames
– Each NIC has a different address
•
MAC address
© Copyright 1997, The University of New Mexico
4-10
Ethernet- Network Interface
Card
82556
PLX9032
or
PLX9036
BIG MAC
1 Meg
Flash
16 K
SRAM
FIFO
© Copyright 1997, The University of New Mexico
100 Mbps
Thunder
Module
Mags
82503
Mags
Relay
RJ45
4-11
MAC Address
• Media Access Control Address
• OSI Layer 2
• 6 Bytes
– 1st 3 Bytes- Manufacturer ID
– 2nd 3 Bytes- Random
• 00-00-81-10-a3-4b
– Vendor Specific 10-a3-4b
© Copyright 1997, The University of New Mexico
4-12
Ethernet
• Thick Coax (Thicknet)
–
–
–
–
–
–
–
IEEE 10Base-5
Maximum distance of 500m
Nodes “tap” into cable with transceivers
Supports up to 250 nodes per cable segment
Must be terminated at both ends
Bus topology
50 ohm coaxial cable
© Copyright 1997, The University of New Mexico
4-13
Ethernet- Thick Coax
AUI Cable
XCVR
XCVR
XCVR
XCVR
TAP
TAP
TAP
TAP
Thick Coax
© Copyright 1997, The University of New Mexico
4-14
Ethernet
• Thin Coax (Thinnet)
–
–
–
–
–
–
–
IEEE 10Base-2
Maximum distance 180m
Supports up to 30 nodes per cable segment
Nodes connect via BNC connectors
Must be terminated at both ends
Bus topology
50 ohm coaxial cable
© Copyright 1997, The University of New Mexico
4-15
Ethernet- Thinwire
BNC
BNC
BNC
BNC
BNC
BNC
BNC
BNC
Thin Coax
© Copyright 1997, The University of New Mexico
4-16
Ethernet
• Unshielded Twisted Pair (UTP)
–
–
–
–
–
–
IEEE 10Base-T
Star topology from hub
EIA/TIA category 3, 4, 5 supported
Uses RJ-45 connectors
Maximum distance 100m (500 m with fiber)
Supports 1 node per cable segment
© Copyright 1997, The University of New Mexico
4-17
Ethernet- UTP
UTP Cable
Hub
© Copyright 1997, The University of New Mexico
4-18
10BaseT Pinout
RJ45 Connector
Note that TX and RX are going
to be reversed on the hub side,
Thus when connecting hub to hub,
the need for MDI/MDIX.
MDIX looks for a NIC, MDI
looks for a repeater port
© Copyright 1997, The University of New Mexico
1
TX+
2
3
TX- RX+
4
5
6
7
8
RX-
4-19
Ethernet
• 10BROAD36
–
–
–
–
75 ohm coaxial cable, CTV
Bus/tree topology
Maximum segment length is 1800 m
Signaling technique is broadband (DPSK)
© Copyright 1997, The University of New Mexico
4-20
Ethernet
• Fiber Optic
–
–
–
–
–
–
IEEE 10Base-F, 10Base-FL, 10Base-FOIRL
Star topology from hub
Multi Mode Fiber supported
MMF supports up to 2km distance
Point-to-Point
Secure
© Copyright 1997, The University of New Mexico
4-21
Fast Ethernet
• Set of specifications developed by IEEE 802.3
committee to provide Ethernet compatible LAN
operating at 100Mbps
• Specifications are:
– 100Base-TX: 2 CAT 5 UTP
– 100Base-FX: 2 optical fiber
– 100Base-T4: 4 CAT 3 or CAT 5 UTP
© Copyright 1997, The University of New Mexico
4-22
Ethernet- Repeaters
Repeater
500m
500m
500 meters between repeaters for 10Base5 - Thicknet
2 repeater limit between any two stations
any more and you need to insert a bridge/switch or a router
© Copyright 1997, The University of New Mexico
4-23
Ethernet- Equipment
• Hubs
– Commonly referred to as concentrators or multiport
repeaters
– Used to connect multiple cable segments
– Retimes & Regenerates signal
– May support multiple cable types
© Copyright 1997, The University of New Mexico
4-24
Ethernet- Hubs
UTP Cable
Hub
If too many collisions or other errors
(CRC alignment, Jabber, Giant frames)
occur on a given port, many Bay Networks
hubs will auto partition that port to prevent
it from repeating on the entire segment.
© Copyright 1997, The University of New Mexico
4-25
Token Ring Logical Operation
• Control token concept by IBM and GM
• Deterministic access method, no contention
• Two main standards: IEEE 802.5 MAC protocol,
FDDI
• Where Ethernet uses CSMA/CD, Token Ring uses a
token passing mechanism where one station is
allowed to talk at a time
• Traffic moves in a logical circle, going from station
to station, until the appropriate station receives it and
holds on to it
© Copyright 1997, The University of New Mexico
4-26
Token Ring Physical Operation
• Ring is emulated through TR hub by passing frames
from station to station (1, 4 or 16Mbps).
• When a station has a problem or is pulled from the
ring, that port is “wrapped” to allow the ring to
continue
• An active hub retimes and regenerates the signal, a
passive hub connects the ring
• When a station inserts at the wrong speed or loses
track of its neighbor, a beacon condition is created on
the ring
4-27
© Copyright 1997, The University of New Mexico
Token Ring Pros and Cons
• Light load: TR is not very efficient. Stations must
wait for the token to arrive
• Heavy load: TR operates in a round robin fashion
• Priority and reservation schemes can be implemented
• Bandwidth can be guaranteed
• High overhead for token maintenance
• Loss of token prevents further use or TR
• Duplication of token disrupts operation
• A monitor station is need to manage the TR network
© Copyright 1997, The University of New Mexico
4-28
FDDI
• Defined by four separate specifications
Layer 2
Layer 1
Logical Link
Control
Media Access
Control (MAC) Station
Management
Physical Layer (SMT)
Control (PHY)
Fiber Distributed
Data Interface (FDDI)
Standards
Physical Layer
Medium (PMD)
© Copyright 1997, The University of New Mexico
4-29
FDDI
• MAC: how medium is accessed, frame format, token
handling, addressing, CRC check, error recovery
• PHY: defines data encoding/decoding, clocking
requirements, and framing
• PMD: defines characteristics of transmission
medium, power levels, bit error rates, optical
components, connectors
© Copyright 1997, The University of New Mexico
4-30
FDDI
• SMT: defines FDDI station configuration, ring
configuration, ring control features, station insertion
and removal, initialization, fault isolation and
recovery, scheduling, and collection of statistics
© Copyright 1997, The University of New Mexico
4-31
FDDI
•
•
•
•
•
Follows the token ring scheme
Similar to IEEE 802.5
Has no priority or reservation bits
Designed for LANs and MANs
Accommodates higher data rates of 100Mbps
© Copyright 1997, The University of New Mexico
4-32
FDDI
• FDDI operation
Dual Attach Station
Single Attach Station
© Copyright 1997, The University of New Mexico
Primary Ring
Secondary Ring
4-33
Primary and Secondary Rings
• Primary and Secondary Rings
– normal traffic is on primary ring
– secondary “counter-rotating” ring is redundant for primary failure
• Three Paths
– primary - normal conditions
– secondary - backup
– local - intelligent insertion and removal
© Copyright 1997, The University of New Mexico
4-34
FDDI Stations
• Dual Attached Stations (DAS)
– connected to both rings
• Single Attached Stations (SAS)
– attached only to the primary ring through
a hub/concentrator port
DAS
Primary Ring
Secondary Ring
Concentrator
SAS
SAS
© Copyright 1997, The University of New Mexico
SAS
4-35
FDDI Token Frame
• Preamble: 64 bits
• Starting Delimiter (SD): 8 bits
• Frame Control (FC): 8 bits
– bit format of 10000000 or 11000000 to indicate that it is a
token
• Ending Delimiter (ED): 4 bits
– contains a pair of non-data symbols (T) that terminate
Preamble
SD
FC
ED
token frame
© Copyright 1997, The University of New Mexico
4-36
FDDI Frame
• Preamble: 64 bits
– synchronize frame with each station’s clock
– repeating stations may change length of the field as per
clocking requirements
• Starting Delimiter (SD): 8 bits
– indicates beginning of a frame;
– coded as JK where J and K are non-data symbols; J same
Preamble
SD as
FCpreceding
DA SA symbol,
DATA KFCS
EDpolarity
FS
polarity
opposite
© Copyright 1997, The University of New Mexico
4-37
FDDI Frame
• Frame Control (FC): 8 bits
–
–
–
–
has bit format CLFFZZZZ
C indicates synchronous or asynchronous frame
L indicates use of 16 or 48 bit addresses
FF indicates whether it is a LLC, MAC control or reserved
frame
– in a control frame ZZZZ indicates the type of control
• Destination Address (DA): 16 or 48 bits
– specifies station for which the frame is intended
• Source Address (SA): 16 or 48 bits
– specifies station that sent the frame
© Copyright 1997, The University of New Mexico
4-38
FDDI Frame
• Data: 0 - 5000 bytes
– contains an LLC data unit or information related to a
control operation
• Frame Check Sequence (FCS): 32 bits
– cyclic redundancy check based on FC, DA, SA and Data
fields
• Ending Delimiter (ED): 4 bits
– contains a non-data symbol T; T=0 token frame, T=1
normal frame
• Frame Status (FS): 1 bit
– contains one of the following:
© Copyright 1997, The University of New Mexico
4-39
FDDI Limitations
• STP, UTP, Single and Multimode Fiber
• implementation over copper is known as CDDI
• 500 stations per ring
• 200 km or 124 miles total length
• 10 km between station and concentrator single mode
• 2 km between station and concentrator multimode
© Copyright 1997, The University of New Mexico
4-40
FDDI Ports
• DAS A - Primary In, Secondary Out
• DAS B - Primary Out, Secondary In
• Master - concentrator port only
• Slave - SAS only primary in and out
FDDI DAS
Primary
Port
A
© Copyright 1997, The University of New Mexico
Port
B
Secondary
4-41
Optical Bypass Switch
• When a station is off this goes into by-pass mode
• Prevents Ring-Wrap
Station
A
Ring Wrap
MAC
B
© Copyright 1997, The University of New Mexico
4-42
FDDI Architecture
• Use only concentrators on the FDDI dual ring
– Minimize chance of ring segmentation by reducing number
of components on dual ring
– Enhance scalability by allowing stations to be added and
removed without disrupting the dual ring
– Do not use ring wiring unless forced to do so by existing
cable plant
• Place all equipment attached to the FDDI dual-ring in
one or two physically secure locations
© Copyright 1997, The University of New Mexico
4-43
FDDI Architecture
• Dual home critical bridges, routers and servers –
provide physical layer redundancy
– Connect remote locations to the FDDI backbone via dualhomed routers and/or concentrators
• Single attach non-critical desktop connections
• Use star wiring wherever possible
© Copyright 1997, The University of New Mexico
4-44
What is Gigabit Ethernet?
Ethernet
Speed....................................... 10 Mbps
Cost......................................... X
Fast Ethernet
100 Mbps
2X 10BT
Gigabit Ethernet
1000 Mbps
2-4X 100BT
IEEE Standard...................... 802.3
Media Access Protocol.......... CSMA/CD
Frame Format........................ IEEE 802.3
802.3u
CSMA/CD
IEEE 802.3
802.3z
CSMA/CD
IEEE 802.3
Topology.................................
Cable support.........................
Network diameter (max).......
UTP link distance (max)........
Media independent interface
Star
UTP, fiber
210 meters
100 meters
Yes (MII)
Star
UTP, fiber
200 meters
100 meters
Yes (G-MII)
Full duplex capable?............. Yes
Yes
Yes
Broad multivendor support.. Yes
Multivendor Availability....... Now
Yes
Now
Yes
1998
© Copyright 1997, The University of New Mexico
Bus or star
Coax, UTP, fiber
2,500 meters
100 meters
Yes (AUI)
4-45
Gigabit Ethernet Technology
10 Mbps Ethernet
CSMA/CD MAC
100 Mbps Ethernet
CSMA/CD MAC
No change
AUI
Thick Coax
(10Base5)
Fiber
(10BaseF)
Thin Coax
(10Base2)
Twisted
Pair
(10Base-T)
Minor
change
CSMA/CD MAC
MII
G-MII
Four Pair UTP
(100Base-T4)
Four Pair UTP
(TBD)
Fiber
(100Base-FX)
© Copyright 1997, The University of New Mexico
1000 Mbps Ethernet
Two Pair
UTP, STP
(100Base-TX)
MM Fiber
CD Laser
1300nm Laser
SM Fiber
1300nm
Laser
4-46
Gigabit Ethernet Topology
• CSMA/CD Topology
– Supports 100 Meter distance for CSMA/CD protocol
over all media options.
• Full Duplex Topology
– Supports variable distance dependent on media
capabilities
– Supports server, switch and router DTE applications.
© Copyright 1997, The University of New Mexico
4-47
Gigabit Ethernet vs. ATM
Link Speed.............................
Ethernet(Gigabit)
10, 100, 1000 Mb/s
ATM
25, 155, 622, 1.2 Gb/s
Standards.. .............. ..............
Shared Access Protocol.........
Switched Access Protocol.....
Frame Format........................
Transmission Control............
Redundancy............................
VLAN support........................
Network diameter (max).......
UTP link distance (max)........
IEEE 802.3
CSMA/CD
N/A
IEEE 802.3 Frame
Connectionless
802.1 Spanning Tree
802.1Q (?)
200 meters
100 meters
ATM Forum
N/A
SVC/CMS
ATM Cell
Connection Oriented
ATM Mesh
LANE
2,500 meters
100 meters
Priority...................................
Full duplex switching............
802.1P (?)
Yes
ATM QoS
Yes
Broad multivendor support..
Multivendor Availability.......
Yes
1998
Yes
Now
© Copyright 1997, The University of New Mexico
4-48
Three generations of LANs
• Apple, Bellcore, Sun and Xerox identified the three
generations of LANs in their document [ABSX92].
© Copyright 1997, The University of New Mexico
4-49
First Generation LANs
• Typified by CSMA/CD and Token ring LANs.
• Provided terminal to host connectivity at moderate
data rates.
• Supported client server architecture at moderate data
rates.
© Copyright 1997, The University of New Mexico
4-50
Second Generation LANs
• Typified by FDDI.
• Supported backbone LANs.
• Supported high performance work stations.
© Copyright 1997, The University of New Mexico
4-51
Third Generation LANs
• Typified by ATM LANs.
• Provide aggregate throughputs and real time
transport guarantees for multimedia applications.
© Copyright 1997, The University of New Mexico
4-52
Further Criteria For Third
Generation LANs
• Support multiple, guaranteed classes of service.
– examples:
• a live video application requires 2 Mbps guaranteed.
• a file transfer uses background class of service.
• Provide scaleable throughput in terms of per host capacity.
– enabling applications requiring large volumes of data in and out of a
single host.
• Provide scaleable throughput in terms of aggregate capacity.
– enabling installations to grow from a few to several hundred high
performance hosts.
• Internetwork between LAN and WAN technology.
© Copyright 1997, The University of New Mexico
4-53
ATM - A Third Generation LAN
• Multiple classes of service provided by virtual paths and
virtual channels.
• Permanent connections provide pre-configured service.
• Switched connections provide on demand service.
• ATM is scaleable
– more switching nodes can be added.
– higher data rates can be used for attached devices.
• Cell based data transport at high rates allow wide area
networking and local area networking, with seamless
connectivity between the two.
© Copyright 1997, The University of New Mexico
4-54
Types of ATM LANs
• Gateway to ATM WAN
– ATM switch acts as a router and traffic concentrator,
linking premises networks to an ATM WAN.
• Backbone ATM switch
– Single ATM switch interconnects LANs.
– Network of ATM switches interconnects LANs.
• Workgroup ATM
– End systems(high performance multimedia workstations)
connect directly to an ATM switch.
© Copyright 1997, The University of New Mexico
4-55
Data Communications With
Processors
• Two types
– Input/Output channel
– Network communications
© Copyright 1997, The University of New Mexico
4-56
I/O Channel Features
• Direct point to point or multipoint high speed communications
link over very short distances.
• Hardware based link.
• Data is transferred between a buffer at source device and a
buffer at destination device.
• No special formatting is used, data is transferred in raw form.
• Transfer control and error detection is implemented in
hardware.
• I/O channel links processors with peripheral devices.
© Copyright 1997, The University of New Mexico
4-57
Network Communications
Features
•
•
•
•
Collection of interconnected access points.
Software protocol structure enables communication.
Many different types of data transfer are allowed.
Network protocols that provide flow control, error
detection and error recovery are implemented in
software.
• Data transfer, between end systems in the local,
metropolitan and wide area distances, is managed.
© Copyright 1997, The University of New Mexico
4-58
Fiber Channel Features
• Has simplicity and speed of channel communications.
• Implements communications oriented protocols as in network
communications.
• Provides interconnectivity between end systems like in the
network communications and also short distance connectivity
as in the case of channel communications.
• One multiprotocol interface is used to provide:
–
–
–
–
Peripheral connectivity.
Host to host internetworking.
Connected processor clusters.
Multimedia capabilities.
© Copyright 1997, The University of New Mexico
4-59
Channel Oriented Features of
Fiber Channel Architecture
• Data-type qualifiers for routing frame payload into
particular interface buffers.
• Link-level constructs associated with individual I/O
operations.
• Protocol interface specifications to allow support of
existing I/O channel architectures, such as the small
computer interface (SCSI).
© Copyright 1997, The University of New Mexico
4-60
Network Oriented Features of
Fiber Channel Architecture
• Full multiplexing of traffic between multiple
destinations.
• Peer-to-peer connectivity between any pair of ports
on a Fiber channel network.
• Capabilities of internetworking with other
technologies.
© Copyright 1997, The University of New Mexico
4-61
Fiber Channel Requirements
as per Fiber Channel
Association
• Full duplex links with two fibers per link.
• Performance from 100 Mbps to 800 Mbps on a
single link(200 Mbps to 1600 Mbps per link).
• Support for distances upto 10 kms.
• Small concentrators.
• High capacity utilization with distance insensitivity.
© Copyright 1997, The University of New Mexico
4-62
Fiber Channel Requirements
as per Fiber Channel
Association
• Greater connectivity than existing multidrop
channels.
• Broad availability.
• Support for multiple cost/performance levels, from
small systems to super-computers.
• Ability to carry multiple existing interface command
sets for existing channel and network protocols.
© Copyright 1997, The University of New Mexico
4-63
Fiber Channel Elements (Slide
1)
• Nodes
– These are the end systems
• Fabric
– Collection of switching elements
– Fabric connects nodes and switches by point to point links
– Routes frames between nodes connected to the fabric
– May buffer frames to enable connections at different data
rates
– Can be a single fabric element or can be a collection of
fabric elements networked together
© Copyright 1997, The University of New Mexico
4-64
Fiber Channel Elements (Slide
2)
• N_port
– Port on a node which participates in a point to point link.
• Frames
– Data units used for communications on the point to point
links.
• F_port
– Port on a switching fabric which participates in a point to
point link.
© Copyright 1997, The University of New Mexico
4-65
Fiber Channel Vs. Other LANs
• Circuit switched or packet switched as opposed to shared
medium LAN.
• No medium access control issues.
• Scales easily in :
– Number of N_ports
– Data rate
– Distance covered
• New transmission media, technology and data rates can be
accommodated by adding new switches and F_ports to
existing fabrics.
© Copyright 1997, The University of New Mexico
4-66
Fiber Channel Protocol
Architecture (Slide 1)
• FC-O physical media
– Optical cable with laser or LED transmitters for long
distance transmissions.
– Copper coaxial cable for highest speeds over short
distances.
– Shielded twisted pair for low speeds over short distances.
– Rates of 100 Mbps, 200Mbps, 400Mbps, 800Mbps
© Copyright 1997, The University of New Mexico
4-67
Fiber Channel Protocol
Architecture (Slide 2)
• FC-1: Transmission protocol
– Byte synchronization and encoding
– 8B/10B encoding/decoding scheme
• Balanced scheme
• Simple to implement
• Has useful error detection capability
– Special code character maintains byte and word alignment
© Copyright 1997, The University of New Mexico
4-68
Fiber Channel Protocol
Architecture (slide 3)
• FC-2: Signaling protocol
– Transport mechanism
– Framing protocol and flow control between N_ports
– Three classes of service between ports
• FC-3: Common Services
– Port related services
– Services that appear across two or more ports in a node
© Copyright 1997, The University of New Mexico
4-69
Fiber Channel Protocol
Architecture (slide 4)
• FC-4: Upper layer protocols
– Supports a variety of channel and network protocols.
– Channel:
• SCSI, HIPPI,IPI-3, SBCS
© Copyright 1997, The University of New Mexico
4-70
Fiber Channel Physical
Interface and Media
• Specified by fiber channel FC-0
• Several data rate allowed :
– 100 Mbps to 800 Mbps
• Several mediums allowed :
– Optical fiber
– Coaxial cable
– Shielded twisted pair
• Maximum distances for point-to-point links range
from 50 meters to 10 Km.
© Copyright 1997, The University of New Mexico
4-71
Fiber Channel Transmission
Protocol
• Specified by fiber channel FC-1
• Defines:
– Signal encoding technique for transmission
• 8B/10B encoding scheme is used.
• 8 bits of data from level FC-2 is converted into 10 bits for
transmission.
– Signal encoding technique for synchronization
© Copyright 1997, The University of New Mexico
4-72
Fiber Channel Common
Services (Slide 1)
• Specified by fiber channel FC-3.
• Striping
– Transmit a single information unit across multiple links
simultaneously.
– Good for transferring large data sets in real time.
– Good for video-imaging applications.
© Copyright 1997, The University of New Mexico
4-73
Fiber Channel Common
Services (Slide 2)
• Hunt groups
– Set of associated N_ports at a single node identified by an
alias.
– Any frame sent to the alias is processed by an available
N_port within the set.
– Good for decreasing latency.
© Copyright 1997, The University of New Mexico
4-74
Fiber Channel Common
Services (Slide 3)
• Multicast
– Sends to all N_ports on a fabric.
– Sends to a subset of the N_ports on a fabric.
© Copyright 1997, The University of New Mexico
4-75
Fiber Channel Mapping
• Specified by FC-4
• I/O channels mapped by FC-PH
– Small Computer System Interfaces (SCSI) supports.
• High capacity devices like disks.
• High data rate devices like graphics and video equipment.
– High Performance Parallel Interface (HIPPI).
• Used for mainframe/supercomputer environments.
– Network protocols mapped by FC-PH.
• IEEE802 Mac frames to FC frames.
• ATM
• IP
© Copyright 1997, The University of New Mexico
4-76
Fiber Channel Transmission
Media (Slide 1)
• Single mode fiber
– 10Km at 800Mbps, 10Km at 400Mbps, 10Km at 200Mbps
• 50 um multimode fiber
– 0.5Km at 800Mbps, 1Km at 400Mbps, 1500m at
200Mbps, 1500m at 100Mbps.
• 62.5 um multimode fiber
– 175m at 800Mbps, 350m at 400Mbps, 1500m at 200Mbps,
1500m at 100Mbps.
© Copyright 1997, The University of New Mexico
4-77
Fiber Channel Transmission
Media (Slide 2)
• Video coaxial cable
– 25m at 800Mbps, 50m at 400Mbps, 75m at 200Mbps,
100m at 100Mbps.
• Miniature coaxial cable
– 10m at 800Mbps, 15m at 400Mbps, 25m at 200Mbps, 35m
at 100Mbps.
• Shielded twisted pair
– 50m at 200Mbps, 100m at 100Mbps.
© Copyright 1997, The University of New Mexico
4-78
Fiber Channel
Topologies(Slide 1)
• Fabric or switched topology
– An arbitrary topology requiring at least one switch to
interconnect a number of N_ports.
– The fabric may be a switched network of a number of
switches, with or without end nodes.
– Each port in fabric has a unique address.
– Destination address of data from a node is examined by an
edge switch and is appropriately forwarded to a node or
another switch enroute to the node.
© Copyright 1997, The University of New Mexico
4-79
Fiber Channel
Topologies(Slide 2)
• Point to point topology
– Only two N_ports directly connected.
• Arbitrated loop topology
– Connects up to 126 nodes in a loop.
– Ports on loop must support N_port and F_port
functionalities.
– Ports are called NL_ports
– Operation is roughly equal to token ring protocols.
– Each port sees all frames and passes them along, keeping
the ones it needs and ignoring the ones not addressed to it.
– A token access protocol is used to get loop access by a port
© Copyright 1997, The University of New Mexico
4-80
Wireless LANs Terminology
• Standards defined by IEEE 802.11 committee.
• Basic services set:
– Contains stations executing the same MAC protocol.
– Stations compete to use the shared wireless medium.
• Access point
– Allows a basic set to connect to a backbone distribution
system.
– Access point acts as a bridge.
© Copyright 1997, The University of New Mexico
4-81
Wireless LANs Terminology
(Cont.)
• Distribution system
– Connects two or more basic sets together.
– Typically it is a wired backbone LAN.
• Extended service set
– Collection of BSSs connected by a distribution system.
© Copyright 1997, The University of New Mexico
4-82
Types of Wireless Stations
• Definitions specified in IEEE 802.11, based on
mobility.
• Three types of stations are specified:
– No transition
• Usually stationary.
• May be mobile in direct communication range of communicating
stations within the BSS.
– BSS transition
• Station moves from one BSS to another within the same ESS
– ESS transition
• Station moves from one BSS in one ESS to a BSS in another ESS.
© Copyright 1997, The University of New Mexico
4-83
Wireless Physical Medium
Specification
• 802.11 defines 3 physical media
– Infrared
• 1Mbps using range 850nm to 950nm.
• 2Mbps using range 850nm to 950nm.
– Direct sequence spread spectrum
• 7 channels, each with a data rate of 1Mbps or 2 Mbps can be used.
• Operates in 2.4GHz ISM band.
– Frequency hopping spread spectrum
• Operates in 2.4GHz ISM band.
• Details are still being worked on.
© Copyright 1997, The University of New Mexico
4-84
Wireless Media Access
Control (Slide 1)
• Two types of MAC algorithms were considered:
– Distributed access protocols
• Decision to transmit was distributed over all nodes using a carrier
sense mechanism.
• Makes more sense for the adhoc peer station networks.
• Good for networks whose traffic is busy.
– Centralized access protocols
• Access is managed by a central management station.
• Makes more sense in wireless networks including base stations
attached to backbone networks.
• Good for networks whose traffic is time sensitive and high priority.
© Copyright 1997, The University of New Mexico
4-85
Wireless Media Access
Control (Slide 2)
• Distributed Foundation Wireless MAC
(DAFWMAC) was the final product of 802.11
committee’s efforts to come up with a wireless MAC
protocol.
• Allows distributed access control mechanisms.
• Allows central control of access as an optimal feature
• Contains 3 layers:
– Physical layer
– Distrbuted coordination function (DCF)
– Point coordination function (PCF)
© Copyright 1997, The University of New Mexico
4-86
Wireless Media Access
Control (Slide 3)
• Distributed coordination function
–
–
–
–
Uses CSMA algorithm.
No collision detection function is used.
A priority scheme based on a set of delays is used.
A single delay is known as an interframe space (IFS)
© Copyright 1997, The University of New Mexico
4-87
Wireless Media Access
Control (Slide 4)
• The rules of CSMA algorithm
– Transmitting station senses the medium and may transmit
if it finds the medium to be idle for one IFS.
– In case the medium is found busy, transmitting station
defers transmission and continues to sense the medium
until current transmission ends.
– If the medium is found busy over another IFS, transmitting
station backs of for a random period of time and tries again
later using a binary exponential backoff algorithm.
© Copyright 1997, The University of New Mexico
4-88
Wireless Media Access
Control (Slide 5)
• Three values of IFS are used:
– SIFS
• Short IFS is used for all immediate response actions.
• Upon reception of a frame ack frames are sent after a SIFS delay.
– PIFS
• Point coordination function IFS is used for central access control.
– DIFS
• Distributed coordination function IFS is used as a minimum delay
for asynchronous frames contending for access.
© Copyright 1997, The University of New Mexico
4-89
Wireless Media Access
Control (Slide 6)
• Point coordination function
– Built on top of DCF.
– A central point coordinator issues polls in a round robin
fashion to stations and offers each one the medium as
required by it.
– The central point coordinator gains control of the channel
through the CSMA mechanism described above.
– The central point coordinator has to relinquish control of
the medium after it has controlled it to some pre set time.
© Copyright 1997, The University of New Mexico
4-90