Tööstuslik andmeside kontrolltöö 2 abimaterjal - vastused (0)

oData transparency : In bit and byte oriented protocols, there is a problem if a control character (for byte-oriented protocols) or the start-of- frame flag (for bit-oriented protocols) appears in the actual data. This was not likely to happen in ASCII text, but is very likely with binary data. This is known as a data transparency problem an can be rectified with byte stuffing (for byte-oriented protocols) and bit stuffing (for bit-oriented protocols).
> Byte stuffing (STX, ETX, DLE) Transmit side: insert DLE before control characters 1. transmit start of frame – send DLE STX 2. prior to transmission insert DLE for any inadvertent DLE 3. transmit end of file – send DLE ETX Receive side: 1. receive DLE STX -> indicates start of frame sequence 2. inspect each byte in the payload until EOF sequence. If byte(i) = DLE: – if byte(i+1) = DLE -> indicates inadvertent DLE discard byte – if byte(i+1) =ETX -> indicates end of frame
> Bit stuffing 1. scheme: flags (often used on point-point links ) – flag = 01111110, unique byte pattern for SOF & EOF – frame read on 8 bit boundaries until EOF detected – reception
terminated. - ensure flag pattern isn’t in frame data -> bit stuffing • transmitter inserts stuff bits : if ‘11111’ detected -> insert ‘0’ – flag pattern can never be present in frame data – performed by transmit circuit at PISO output – disabled during transmission of SOF, EOF • receive circuit: if ‘111110’ detected -> remove ‘0’ – prior to input into SIPO – normally FCS data (prior EOF) subjected to same bit stuffing
O Asynchronous Transmission Independent transmit & receive clocks . – receiver explicitly resynchronizes on 1st bit of each byte – remaining bits recovered by estimating bit boundaries – explicit byte synchronization become critical for correct bit synchronization.
is useful for: • data with irregular arrival times (human input/output) • transmission line with long idle states • used when relatively coarse synchronization is ok e.g. terminal & keyboard I/O Async. transmission is characterized by: • no direct clock information between receiver and transmitter • receiver has to re-synchronize at each byte: uses additional bits – bytes are encapsulated in between start bit & stop bit. – bit synchronization uses start bit & stop bit for each byte. Internally , nodes store , process , & transfer data in parallel Transmission control circuit - interface between node & link : • PISO shift register : parallel-> serial conv. for transmission • SIPO shift register: serial->parallel conv. for receiveing • bit, character, frame synchronization • set data rate , data format • error control: generate /evaluate check digits for error detection
> Bit synchronization Independent transmit & receive clocks.– receiver explicitly resynchronizes on 1st bit of each byte – remaining bits recovered by estimating bit boundaries – explicit byte synchronization become critical for correct bit synchronization. Receiver clock:
– used to ‘clock’ incoming signal into SIPO register – runs asynchronously with respect to incoming signal – local clock used to sample signal near center -> recover & shift bit into SIPO – local clock = N·R, N = 16 is common 1. 1st signal transition from 1-> 0 = start of byte (SOB) 2. Each signal sampled at ½ bit time – 1st signal sampled after N/2 cycles – ith signal sampled after N/2 + N(i-1) cycles 3. Each sample represents bit that is shifted into SIPO register
• Effect of clock drift on bit synchronization Bit synchronization is unreliable at higher data-rates – detection of bits are only an approximation – clock drift can cause incorrect sampling instant .
> Byte synchronization Bit counting used to determine byte boundary e.g. SOB, 8 bits, EOB – relatively simple once bit correctly recovered. After received , character transferred into local buffer – cpu notified that new character received (interrupt or polling) – await next SOB (1 -> 0 transition) Receiver & Transmitter must operate with same parameters – bits per character – number of stop-start bits – signal rate
> Frame synchronization Receiver must identify SOF & EOF 1. frames consist of text characters only -> use non-text characters – STX & ETX: ASCII characters for ‘start of text’ & ‘end of text’ 2. frames are pure binary data (no restrictions) – inadvertent STX or ETX in payload can cause incorrect behavior .
> Xmodem, – One of the oldest async file transfer protocols– Uses stop-and- wait ARQ. – Advantages • Universally available • Copes well with noisy lines – Disadvantages • Small packet size = high overheads = low efficiency • No batch facility
Ymodem, Zmodem, Zmodem: – Newer protocol than Xmodem – Incorporates features of several protocols – Uses CRC-32 with continuous ARQ (sliding window) – Dynamically adjusts packet size according to communication circuit conditions – Usually Zmodem is preferred to Xmodem. – Advantages • Large packet size = low overheads = high efficiency • Batch Transfers • Auto receive – Disadvantages • Relatively susceptible to noisy lines • Not available in all comms packages Kermit: – Supports different packet data sizes and error detection methods – Typically uses 1 KByte packets with CRC-24 • size adjusted during transmission to optimize efficiency
o Synchronous Transmission 2 general types character oriented & bit oriented – both use same bit synchronization – major difference : byte & frame synchronization – bit oriented synchronous transmissions is most prevalent Distinguishing synchronous & asynchronous transmission: – asynchronous has ‘start bit’ & ‘stop bit ’ – asynchronous: receiver clock not synchronized with incoming signal – synchronous: receiver clock synchronized to incoming
signal. Global clock or synchronized transmit & received clocks – bits are explicitly recovered – frame is transmitted as contiguous bit stream – receiver synchronizes each bit for duration of frame. In synchronous transmission, greater efficiency is achieved by grouping characters together, and doing away without the start and stop bits for each character. We still envelop the information in a similar way as before, but this time we send more characters between the start and end sequences. In addition , the start and stop bits are replaced with a new format that permits greater flexibility. A start type sequence, called a header, prefixes each block of characters, and a stop type sequence, called a tail , suffixes each block of characters. Synchronous tranmission uses explicit bit synchronization: •higher bit rates, more efficient • Data is transmitted in large contigous blocks of bits or characters. •frame = unit of transmission. •frame contains payload (data) and overhead (control bits)
> Bit synchronization bits are encoded in signal transitions. • each frame transmitted as contiguous bit stream • 2 ways receiver obtains & maintains bit synchronization 1. Clock Encoding: clock timing information embedded in transmitted signal extracted by receiver 2. DPLL ( digital phase locked loop ): uses forced signal bit transitions in signal – receiver clock synchronized to signal by DPLL – requires forced signal transitions in signal (e.g. bit stuffing)
• Clock Encoding and extraction
 Bipolar encoding: uses RZ ( Return to Zero ) signal • uses 3 signal levels: +V, 0, -V • ‘1’ encoded by ‘+V’ signal • ‘0’ encoded by ‘-V‘ signal • return to 0 after each bit
 Bit centre vs bit edge transition
 Manchester encoding or phase encoding: uses NRZ (Non Return to Zero) signal • always center bit transition, edge transition if needed – binary ‘1’ -> lo-hi center transition – binary ‘0’ -> hi-lo center transition – center bit transition used by receiver’s clock extraction circuit to produce clock pulse in center of 2nd half of bit cell. • sampled signal is either high (1) or low (0) -> shifted into SIPO
 Differential Manchester encoding •always center bit transition for clock synchronization •edge transition at start of next bit only if next bit = ‘0’ •encoded output takes on 1 of 2 forms , depending on initial state – either is an inverted form of the other – clock generated at end of each bit cell – edge transition determines if bit cell is ‘0’ or ‘1’
Manchester Schemes are balanced codes:• no mean DC value • 000… or 111… always has transitions (no constant DC signal) • important for AC coupling to receiver using transformer •isolates receive electronics power supply from transmitted signal
• DPLL (digital phase locked loop) • Alternative to clock encoding for bit synchronization •Requires a sufficient edge transitions in bit stream – receiver resynchronize’s clock with edge transition – scramble input data or NRZI with bit stuffing •Estimates bit timing in between edge transitions Pass data through scrambler before transmission – randomize bit stream -> remove contiguous 1’s & 0’s – unscramble at receiver using inverse operation . DPLL circuit: • used to maintain bit synchronism • DPLL clock frequency ≈ 32 · bit rate • must hold frequency sufficiently constant – require only small adjustments at irregular intervals. DPLL uses clock to derive timing interval to sample incoming bits – assumes bit stream & local clock are synchronized – input signal state (‘1’ or ‘0’) sampled at center of each bit – signal clocked onto SIPO register – exactly 32 clock periods between each sample.
 NRZI encoding (non-return to zero, inverted) • signal level changes for ‘0’ doesn’t change for ‘1’ – ‘111…’ results in no transitions – ‘000…’ results in n transitions - differential encoding
‘0‘ insertion bit used for every 5th ‘1’ bit -guaranteed number of transitions enables receiver to adjust clock to synchronize with bit stream
 Bit synchronization(sampling instant adjustment)
Assume only small clock variations • cause incoming bit stream & clock to drift out of synchronization • adjust sampling instants in discrete increments 1. no bit transition -> DPLL generates sampling pulse at 32 ticks 2. if bit transition detected (at least every 5th bit): – determine drift amount = actual signal transition relative to last – DPLL sample: drift = tactual_transition - (tsample + N/2) – adjust time to next sampling pulse depending on amount of drift.
• Baud rate vs bit rate
> Byte synchronization: hunt mode
> Character oriented synchronous transmission• Differences with asynchronous transmission • used primarily for transmission of blocks of characters (ASCII files) • no ‘start’ & ‘stop’ bits -> uses transmission controlcharacters before each block • known as SYN (synchronous idle) characters – 1st SYN allow receiver to obtain & maintain bit synchronization – 2nd SYN provides byte synchronization – Frame Synchronization ~ Asynchronous Transmission using STX & ETX 1. obtain bit synchronization with 1st SYN – starts receiving bits 2. enter hunt mode for byte synchronization: after each bit -> check last 8 bits 3. frame synchronization: process each byte looking for STX - after STX, read frame contents look for ETX – after ETX, transmitter can maintain synch by sending SYNs, otherwise receiver has to repeat hunt mode with new frame – data transparency: same as with asynchronous (DLE)
– inefficiency: additional control characters: STX,ETX, DLE, SYN
• Binary synchronous protocol: PAD, SYN, STX, ETX, … Every package that is transmitted is packed between an STX and an ETX character, followed by one (LRC-check) or two (CRC-check) BCCs (Block Check Character). After this (and the check on the receiving side) the receiver will send an ACK.SYN (Synchronous Idle) •Provides the hardware recognizable bit pattern rquired to establish character synchronization at the receiving adapter . ENQ ( Enquiry ) •Recognized as a request for a response, or a bid for line control. In some cases it may be used to signify an abnormal end of text or message ' abort '. SOH (Start of Header) •Indicates the inclusion of auxiliary data preceding the message text. STX (Start of Text) •Indicates the beginning of data in a block. STX may be preceded by a header. Directly behind the STX is the first character of the data field . NAK ( Negative Acknowledgement) •Indicates that there was an error in a data block. Also used as a response to a bid for line control to indicate a 'Not Ready ' condition . DLE (Data Link Escape ) •Multiple usage as a control character modifier. ETB (End of Transmission Block) •Indicates an end of data block, but more will follow . Is used to instruct the receiving unit to perform error checking and acknowledge.
ETX (End of Text) •Same as ETB, only no more blocks will follow. ITB (End of Intermediate Transmission Block) •Same as ETB, except that the receiving statio will not acknowledge after the error checking. EOT (End of Transmission) •Indicates that a station has no data to transmit.
>Bit oriented synchronous transmission: preferred scheme • transparent to either printable characters or binary data • 3 general schemes that vary on how SOF, EOF is signaled •received stream searched on bit basis for SOF & EOF • frame divided into fields that are often multiples of 8 bits 1. scheme: flags (often used on point-point links) – flag = 01111110, unique byte pattern for SOF & EOF – frame read on 8 bit boundaries until EOF detected – reception terminated.
2. scheme: SOF delimiter & length indicator i) receiver searches for SOF delimiter ‘01111110’ ii) fixed header follows with address & length of data iii) receiver counts number of bytes to determine EOF • used in LANs ( broadcast transmission media) – node’s destination address precedes user data (payload) – clock synchronization achieved when sender transmits preamble – allows other nodes to obtain bit synchronization. 3. scheme: bit-encoding violations ( also used w/ LANs) • 1.5 bit manchester (3 signal pulses) -> no transition in 1st bit • no signal transition -> illegal manchester code – J: signal level remains same as previous for 1 bit period – K: signal level remains opposite as previous for 1 bit period – frame payload can’t contain J,K symbols
• Frame synchronization schemes:
•Data Link Control ProtocolsoHDLC Within HDLC, there are three types of stations defined:1. Primary Station - this completely controls all data link operations issuing commands and receiving responses from secondary stations. It has the ability to hold separate sessions with different stations(on multipoint line) 2. Secondary Station - can only communicate with the
primary station. Secondary stations only talk to each other via a Primary station: receive command frames from Primary, transmit response frames to Primary. 3. Combined Station – has the combined functionalities/ responsibilities of both primary and secondary station.
> Operation of combined station • Contains protocol components of primary and secondary in one physical station. • Transmits both commands and responses.
• Receives both commands and responses. • Uses line addresses to distinguish between command and response frames: – Frame received with own address -> command – Frame received with partner address -> response – Frame transmitted with own address -> response – Frame transmitted with partner address -> command> Balanced vs Unbalanced configuration Configuring a channel for use by a 1. Unbalanced configuration - consists of a primary station
and one or more secondary stations. Primary is responsible for controlling all secondary stations In unbalanced configurations, any of the following can be
used: full duplex or half duplex operation, point to point or
multi-point links.2. Balanced configuration - consists of two combined stations(peers). Stations have equal responsibility for error recovery and line management . Balanced configuration is used only with point to point
links, which can be half duplex or full duplex.
> Operation modes :• Normal Response Mode (NRM):
– the primary station always initiates transfers to the
secondary station.– the secondary station must receive explicit permissionfrom the primary station to transfer a response.– after receiving permission from the primary station, thesecondary station initiates it's transmission.
– this transmission from the secondary to the primarymay be much more than just an ack of a frame.– it may in fact be more than one information frame.
– once the last frame is transmitted by the secondarystation, it must wait once again for explicit permissionfrom the primary station, to transfer anything.
• Asynchronous Response Mode (ARM):– the secondary does not have to wait to receive explicitpermission from the primary to transfer any frames.
– the frames may be more than just acknowledgmentframes. They may contain data, or control informationregarding the status of the secondary station.
– this mode can reduce overhead on the link, as noframes need to be transferred in order to give thesecondary station permission to initiate a transfer.
– on multipoint lines, only one secondary can be in ARMmode.
– the primary station still retains responsibility for errorrecovery, link setup, and link disconnection.• Asynchronous Balanced Mode (ABM):
– this mode uses combined stations.– there is no need for permission on the part of either station in this mode.– this is because combined stations do not require any
sort of instructions to perform any task on the link.– both stations are equally responsible for error recovery and can establish or clear a connection .
– this is the most common mode used on P-to-P links. > Non operational Modes HDLC also defines three non-operational modes. These are:
1. Normal Disconnected Mode(NDM) – for unbalanced mode
– secondary not ready to receive any I or S frame
2. Asynchronous Disconnected Mode(ADM) – for unbalanced mode
– combined station not ready to receive any I or S frame 3. Initialization Mode(IM)
– used for initialization of stations (download of software) or exchange of parameters between stations. The two disconnected modes (NDM and ADM) differ from the operational modes in that the secondary station is logically
disconnected from the link ( note the secondary station is not physically disconnected from the link). The IM mode is different from the operations modes in that the secondary station's data link control program is in need of
regeneration or it is in need of an exchange of parameters to be used in an operational mode. > Principle Frame types: Information vs Supervisory vs Unnumbered There are three control field formats and hence three principle
frame types: 1. Information format - I frame “I” frames are used for the data transfer between stations. The send sequence, or next send N(S), and the receive sequence, or next receive N(R), hold the frame sequence numbers .
The Poll / Final bit - for unbalanced link access control: – when the primary sends something with the P/F bit turned on, it's a poll to the secondary (gives that station access to the line) – secondary transmits with P/F bit off. – when secondary station done transmitting, it turns the P/F bit back on, indicating that it is finished (Final bit).
2. Supervisory format - S frame
”S” frames are used for error control (to acknowledge frames or
request for retransmissions) or flow control (to ask for
suspension of transmission).
The Supervisory code denotes the type of supervisory frame
being sent . Supervisory codes(used for both Command and Response):
– 00: Receiver Ready (RR)– 01: Reject (REJ)
– 10: Receiver not Ready (RNR)– 11: Selective Reject (SREJ)
3. Unnumbered format - U frame
Unnumbered Commands(uncomplete):
• Set Normal Response Mode (SNRM)• Set Async Response Mode (SARM)
• Set Async Balanced Mode (SABM)• Unnumbered Information (UI)
• Exchange Identification (XID)• Disconnect ( DISC )
”U” Frames are used for link initialisation or link disconnection.
The Unnumbered bits indicate the type of Unnumbered frame
being used.Unnumbered Responses(uncomplete)
• Unnumbered Acknowledge(UA)• Frame Reject (FRMR)
• Request Disconnect (RD)• Unnumbered Information (UI)
• Exchange Identification (XID)• Disconnect Mode (DM)
> Special bit sequences– Flag field: 01111110 (7E hex)– Abort sequence: at least 7, but fewer than 15 Ones • when Abort received, the stations on the link know
there is a problem on the link– Idle sequence: 15 or more OnesMost synchronous links constantly transmit data:– during the inter -frame period, these links can transmitall 1s (mark idle), or all flag characters (flag idle).HDLC is code transparent:– uses bit stuffing (zero insertion) if flag sequence would
appear within frame
Information transfer format command and response
• The functions of the Information command and response(C/R) frames are to transfer sequentially numbered frames(each containing an information field) across the data link.– Requires connection setup prior to data transfer
Supervisory format commands and responses• Supervisory(S) commands and responses are used toperform numbered supervisory functions such as
acknowledgment, polling, temporary suspension ofinformation transfer, or error recovery.• Frames with the S format control field cannot contain an
information field. Supervisory format commands and responses• A primary station may use the S format command framewith the P bit set to 1 to request a response from asecondary station regarding its status.• Supervisory format commands (C) and responses (R) are asfollows:– Receive Ready(RR) (C/R) is used by the primary orsecondary station to indicate that it is ready to receive
an information frame and/or acknowledge previously received frames.
– Receive Not Ready(RNR) (C/R) is used to indicate that the primary or secondary station is not ready to receive any information frames or acknowledgments. – Reject(REJ) (C/R) is used to request the retransmission of frames.– Selective Reject(SREJ) (C/R) is used by a station torequest retransmission of specific frames.
Unnumbered Format commands and responses
• The U format commands and responses are used to extendthe number of data link control functions.• The U format frames have 5 modifier bits which allow for
up to 32 additional commands and 32 additional responsefunctions.• Below , 13command (C) functions, and 8 response (R)functions are described .– Set NormalResponse Mode(SNRM) (C) places thesecondary station into NRM.
NRM does not allow the secondary station to send anyunsolicited frames.Hence the primary station has control of the link.– Set Asynchronous Response Mode(SARM) (C) allows asecondary station to transmit frames without a pollfrom the primary station.– Set Asynchronous Balanced Mode(SABM) (C) sets theoperational mode of the link to ABM.– Disconnect(DISC) (C) places the secondary station into a disconnected mode.– Set Normal Response Mode Extended (SNRME) (C)
increases the size of the control field to 2 octets instead
of one in NRM. This is used for extended sequencing. – Disconnected Mode(DM)(R) is transmitted from asecondary station to indicate it is in disconnectedmode(non-operational mode.)– Request Initialization Mode(RIM) (R) is a request froma secondary station for initialization to a primarystation. Once the secondary station sends RIM, it canonly respond to SIM, DISC, TEST or XID commands.– Request Disconnect(RD) (R) is sent by the secondarystation to inform the primary station that it wishes todisconnect from the link and go into a non-operationalmode(NDM or ADM).– Frame Reject(FRMR) (R) is used by the secondarystation in an operation mode to report that a conditionhas occurred in transmission of a frame andretransmission of the frame will not correct the
condition.> NRM operation and error recovery> ABM operation and error recoveryoDLC service modes:> Connection–oriented vs connectionless
1. Reliable/ virtual -circuit (connection-oriented mode) –– users establish connection before sending information packets– errors dealt with using ACKs, NACKS2. Best -try/datagram (connectionless mode)– frames in error are just discarded (unacknowledgedservice)
> HDLC service modes implementation HDLC can provide connection-oriented service:• Setup of connection done by U-frames– SNRM, SARM, SABM, UA• I-frames and S-frames can be used only after connection setup– I, RR, RNR, REJ, SREJ• Clearing of a connection done by U-frames– DISC, UA HDLC can provide connectionless service:• Only U-frames can be used– UI for data transporto LLC• Used with Local Area Networks (LANs)• Defined in IEEE 802.2 specification• LANs have two sublayers in Data Link Layer (LLC and
MAC),– LLC is at higher level than MAC
– will discuss MAC later . (MAC = Medium Access Control)
• LLC is required by many IEEE 802 MAC protocols, since
they don’t provide a multiplexing identifier to identify the
higher layer protocols.
• LLC is a peer -to-peer protocol (since MAC takes care of
access to network we do not need primary/secondary
relationship)
> Differences with HDLC
Differs from HDLC because of multiaccess MAC that provides
framing/error detection:
– Has 2 address fields (source & destination) for multiaccess
– Lacks framing delimiters and CRC
• Byte oriented (bit orientation is rarely useful, and MAC
layers provide byte oriented service)
– Sequence numbers grow from 3b to 7b
• Control field for unnumbered frames (lacking sequence
numbers) is shorter than for information/supervisory
frames
> Functionality of SAP and SNAP • MAC address identifies a hardware interface (“station”). • LLC Service Access Point identifies a protocol
within the device having that hardware interface.
SAPs are 7b: shortage -> SNAP above LLC
– Typical network protocol evolution : • extend one protocol by tacking on parts (in this caseanother protocol)
The DSAP, or Destination Service Access Point, is a 1 byte
field that simply acts as a pointer to a memory buffer in the
receiving station. It tells the receiving NIC in which buffer
to put this information.
The SSAP, or Source Service Access Point is analogous to
the DSAP, and specifies the Service Access Point (SAP) of
the sending process.
In order to specify that this is a SNAP frame, the DSAP/SSAP
is set to AA hex.
The Sub-Network Access Protocol (SNAP) field: 5 bytes
LLC Service types compared to HDLC service modes
What's new in the frame, compared to HDLC:
• Two octet address (first is destination, second is source):
– these are LLC addresses and NOT addresses that appear on
the network medium!
• No FCS (assumes MAC layer will handle it, so why bother!)
• For connection-oriented service, unlike HDLC:
– must operate in SABME mode; no SNRM, SNRME, or
SABM modes.
– No selective reject
• For acknowledged connectionless service:
– AC (C/R): Acknowledged Connectionless: permits
unnumbered stop-and-wait transmission (in addition to
HDLC’s sliding window) (C=data,R=ack)
LLC Can provide 3 types of service to higher layers:
• Connectionless, either
– Unacknowledged (“Type 1”), or
– Acknowledged (“Type 3”). 2 forms:
• Push : source initiates transfer and receives ack
• Pull : destination initiates transfer and receives data
(Service primitives at interface to higher layer:
– REPLY .request to poll– REPLY.indication when response comes– REPLY-UPDATE for higher layer to feed data into
LLC for responding to polls)
Unacknowledged connectionless (“Type 1”) is similar to UDP.
Acknowledged connectionless (“Type 3”) provides acks, but
doesn’t provide error recovery, sequencing or flow control.
• Connection-oriented (“Type 2”) mode is similar to TCP.
– Only supports unicast connections
– Primitives to manage established connections:
• DISCONNECT a connection (i.e. terminate it)
• RESET a connection (i.e. reinitialise, possibly losing
data – up to higher layers to recover)
logical connection must be established using L_CONNECT
prior to data transfer
• after all data transfer, connection must be cleared using
L_DISCONNECT
• during data transfer all error free data units are acknowledged
• L_RESET abort, any unacknowledged frames are discarded
• L_FLOWCONTROL specifies the amount of data the user is
prepared to accept
Priority can be specified for all services .
– This is always passed to the MAC layer to prioritise access
(if supported).
– For connection-oriented service, it also prioritises use of
LLC resources
• All nodes must support the unacknowledged connectionless
service (common in Ethernet 802.3), whereas support of the
other services is optional.
> Operation of LLCLLC uses underlying MAC layer to communicate with endpointby passing parameters specifying network address and all thedata.a) local confirm: whether request has been successfully transmittedb) remote confirm: whether request has been successfully received
oPPP> Importance , Application areas , Protocol stacks
One sender, one receiver, one link - easier than broadcast link:
– no Medium Access Control(MAC)
– no need for explicit MAC addressing
– e.g., dialup link, ISDN line, DSL line
Hence, it’s called the Point-to-Point Protocol (PPP)
Why is PPP important:
• multiplex multiple protocols over a single serial connection
• handle compression and encryption at lowest possible layer
• easy authentication at other end of connection
Two precise definitions of "point-to-point" in the context of
data communications follow:
1. A network configuration in which a connection is
established between two, and only two points .
The connection may include switching facilities .
2. A circuit connecting two points without the use of any
intermediate terminal or computer.
These definitions explain the point-to-point aspect of PPP.
The protocol aspect lies in the fact that PPP is the intermediate
packet structure which facilitates transmission of higher
level protocols, such as TCP/IP, across diverse
communication links.
Application areas
• Designed to be used when the carrier provides bit transfer,
not frame transfer, e.g.:
– Routers connected by leased lines
– Home user to ISP over phone line
• Includes specs for how to test the physical channel
-> PPP over Sonet [RFC(Request for Comments) 2615], etc
• So popular , than often layered on top of protocols that
provide frame transfer so that these protocols can be
accessed by higher layers built to use PPP!:
– PPP over AAL5 (ATM Adaption Layer) [RFC 2364]
– PPP over Ethernet (PPPoE) [RFC 2516]: discussed later
> Design requirements and non requirements
• packet framing: encapsulation of network layer datagram
in data link frame
– carry network layer data of any network layer protocol
(not just IP) at same time
– ability to demultiplex upwards(multiple protocol support)
• bit transparency: must carry any bit pattern in the data
field
• error detection (no correction )
• connection liveness: detect link failure and signal status
change to network layer
• network layer address negotiation: endpoints can
learn/configure each other’s network address
• no prevention of long runs of certain bit strings
• no error correction/recovery
• no flow control
• no need to support multipoint links (e.g., polling)
Error recovery, flow control, data re-ordering(sequencing):
– all delegated to higher layers!
> PPP Frame format compared to HDLC
PPP is character-oriented version of HDLC.
It uses similar frame structure as HDLC, except:
• Flag: same flag as HDLC, but uses byte stuffing.
• Address & Control fields: legacy of the HDLC frame
format.
– Address field value “FF” = broadcast,
– Control field value “11” = UI (Unnumbered Information)
An option allows the omission of the A and C fields.
• Payload(info) field contains an integer number of bytes.
New field in PPP, compared to HDLC:
• Protocol type field: indicates, which higher layer protocol
should this frame be delivered to.
e.g. IP, IPX, PPP control protocol(LCP or NCP)
– Link Control Protocol (LCP): bringing up, testing,
bringing down lines; negotiating options.
– Network Control Protocol (NCP): to configure upper
layer protocol, e.g. IP Control Protocol( IPCP )
> PPP version of byte stuffing
Due to the nature of PPP byte stuffing, a problem can appear:
– Size of PPP frame can grow unpredictably due to byte
insertion.
Problem appears only if the maximum size of frame is limited.
Example of PPP byte stuffing:
Flag=0x7E (01111110), Control escape=0x7D (01111101)
Any occurrence of flag or control escape inside of frame is
replaced with 0x7D, followed by original octet XORed
with 0x20 (00100000):
The three major components of PPP:
1. A method for encapsulating datagrams over serial links –
framing and transmitting frames.
2. An extensible Link Control Protocol(LCP), to establish,
test, configure, and tear down a link.
3. A family of Network Control Protocols(NCPs) for
establishing and configuring different network layer
protocols.
• Framing and transmission
• Link Control Protocol(LCP)
 Link configuration,
Configuration options negotiated using LCP:
• Header compression:
– omit fields that aren ’t used (to save bandwidth over
slow lines), e.g. omit Address & Control fields.
– length of protocol field (1B or 2B)
• Maximum payload length (default: 1500)
• Type of CRC (2B or 4B)
• Disable or select authentication stage (PAP, CHAP , EAP)
• Line quality monitoring during normal operation (the
ability to verify whether the line has a good enough quality
to reliably support the connection).
link termination Using LCP, PPP can terminate the link at anytime.
This might happen because of:
– the loss of carrier,
– authentication failure,
– link quality failure,
– the expiration of an idle-period timer ,
– or the administrative closing of the link.
 LCP Frame types
• Network Control Protocols(NCPs)
 Functionalities, message types Authentication phase: Link establishment (LCP) can be followed by optionalauthentication phase before proceeding to network layerprotocol (NCP) phase.
• Auth . protocols
Various protocols for authentication:
– PAP(Password Authentication Protocol): password sent in clear;
no playback protection -> PAP should be avoided!
– CHAP( Challenge Handshake Authentication Protocol): encrypted password
but the pwd must be stored as cleartext on the server .
EAP(Extensible Authentication Protocol) – most flexible, best choise.
• Why auth. at link layer?
• Client doesn’t need network access to authenticate
– No need to resolve names , obtain an IP address prior to
auth• In a multi-protocol world, doing auth at link layer enablesauthorizing all protocols at the same time– Doing it at the network layer would mean adding
authentication within IPv4 , IPv6 , AppleTalk, IPX,
SNA, NetBEUI• Would also mean authorizing within multiple layers• Result : more delay
> PPP operation
After the LCP component of PPP establishes the link, then the
NCP component will negotiate the network layer protocol
that PPP actually encapsulates and transports.
For example, if the PPP link is configured to connect with IP,
then the Internet Protocol Control Protocol (IPCP) will
negotiate and configure the link to carry IP.
An NCP, such as IPCP, may close the link over the network
layer protocol, and yet the PPP connection can remain
open.
But if LCP closes the link, then all communication layers
terminate, including the network layer, NCP fraffic, and the
PPP connection itself.
> PPPoE PPPoE (defined in [RFC2516]) is a method to transport PPP packets over Ethernet segments.
Since PPP was designed to do things that were either impossible or unnecessary with Ethernet, users are often
confused as to why one would want to use PPP over Ethernet at all.
• When is it useful?
Compare TCP/IP traffic to vehicle traffic:
– Basic TCP/IP protocol would be comparable to a
network of city streets.
– Streets can serve many access points.
– It is easy to get on to and off of the street .
– Additional access points can be added with little
disruption.
– It is hard to tell how many cars are actually using each street.
PPP, on the other hand , would be comparable to a railway
– Travel is generally between two well defined points.
– You can't get on and off anywhere.
– It is relatively easy to count and monitor passengers.

• You need a ticket to board .
  

If the previous is true, then isn't PPPoE like running railway
tracks down main street?
In fact, yes, it is.
– That is what tramways do.
– Without disturbing main street traffic, they bring the
advantages of railways.
– They offer speedy access between two well defined
points and allow you to count passengers.

• And you need a ticket to board.
  

PPP over Ethernet brings this sort of functionality to ISPs that
do not use serial links to connect their users.
Serial ISPs use PPP over modem communications.
DSL providers on the other hand use Ethernet, not serial
communications.
Because of this, many require the added functionality of PPP
over Ethernet, which allows them to:
– secure communications through the use of user logins
– have the ability to measure the volume of traffic each

• user generates.
  

• Frame format
An important restriction: 1492 Octets is the maximum number
of octets in the PPP payload and Padding fields:
1492 Octets PPP payload & PPP Padding + 2 octets PPP–ID +
6 octets PPPoE header yield a maximum Ethernet V2.0
payload of 1500 octets.
• Protocol architecture

• DLC protocols comparison
  

• Multiple Access
Multiple nodes may share a transmission channel.
• This mode of operation is used for wireless networks, and
was common for wired LANs before advent of switches.
• Nodes are often called “stations” in this context
Sometimes “ terminals ” (particularly for WANs) or “ devices ”
• Channel called broadcast or multiple access
• Two or more simultaneous transmissions by nodes ->
interference, from “contention”/“ collisions ”.
– only one station can send successfully at a time
o Types of links
Need to coordinate transmissions to limit interference.
Analogy with human meetigs: Audio channel is shared.
• Face -to-face: Use visual channel for control (e.g. raise hand,
eye contact, rank, etc).
• Audio conference: Coordination is less effective.
CB radio “Over” & “Over and out” Multiple Access protocol(=Medium Access Control protocol):
– algorithm that determines how stations share channel,i.e., determine when station can transmit.– note that communication about channel sharing must usethe channel itself!What to look for in multiple access protocols:– synchronous or asynchronous– information needed about other stations– robustness (e.g., to channel errors)– performanceClaim: humans use multiple access protocols all the time!Design goals for MAC protocol :efficient, fair , decentralized, simple, scalable,priority-dependent, deterministic
oMultiple access protocol: main propertiesoIdeal MAC protocol properties
Broadcast channel of rate R bps:• Efficient: when one node wants to transmit, it can send at
rate R.• Fair: when M nodes want to transmit, each can send ataverage rate R/M
• Fully decentralized:– no special node to coordinate transmissions– no synchronization of clocks, slots• Simple• Scalable: large number of nodes; large bandwidth-delayproduct (geographical scope)
• Able to prioritize traffic, and provide delay bounds
oMessage priority: Local vs Global priorityLocal priority:• each node can transmit its highest priority message when itgets a turn on the bus
Global priority: which node gets the next turn on the bus?• could be a organized as round -robin selection of nodes• could be a function of the node’s inherent(static) priority• could be a function of the highest priority message on eachnode – a “global message priority” scheme
Fundamental tension:• reducing latency for high-priority nodes/messages, vs.• ensuring global fairness: no starvation for low-prioritynodes/message
o3 broad classes of MAC protocols:MAC protocols can be divided into three broad classes:1. Channel Partitioning (fixed assignment, multiplexing)– divide channel into smaller “pieces” (time slots,
frequency, code)– allocate piece to node for exclusive use – no collisions2. Random Access (contention, “transmit and hope ”)– allow collisions,– “recover” from collisions3. Handshaking (scheduling, reservation- based )– tightly coordinate shared access to avoid collisions– for coordination, uses polling or token passing
> Channel Partitioning MAC protocols
• Partitioning techniquesTechniques:• “pieces” (x) = Time slots, Frequency(Wavelength), Code
• “x Division Multiplexing (xDM)”: Different flows ofinformation use different pieces of the channel, e.g. TDM,FDM, Orthogonal FDM, etc• ´“x Division Duplexing (xDD)”: Duplex communication is
achieved by 2 ends using different pieces of the channel(typically time (TDD) or frequency (FDD))
• “x Division Multiple Access (xDMA)”: When multiplenodes are directly transmitting on a channel, e.g. TDMA ,FDMA, CDMA
• Issue : Contention at the receiverIssues:• Contention at the receiver:Division may resolve contention for the medium, but whatabout contention for the receiver?e.g. multiple FDM sources transmitting simultaneously toone receiver that can only tune into one frequency at a time.• TDMA: cycle TDM (Time Division Multiplexing):channel divided into N time slots, one per user;
inefficient with low duty cycle users and at light load .TDMA: time division multiple access
• access to channel in "rounds"• each station gets fixed length slot (length = pkt trans time)
in each round• unused slots go idle• example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6 idleBetter to think of this as a “cycle”. While it is often called a “frame”, it is different from the
“frames” that we have considered , which are essentially linklayer versions of packets.Only one station transmits any particular “frame” (as we’veconsidered), whereas multiple stations transmit in one of
these “frames” (cycles).
• FDMA: frequency bandFDM (Frequency Division Multiplexing): frequency subdivided.FDMA: frequency division multiple access• Channel spectrum divided into frequency bands• Each station assigned fixed frequency band• Unused transmission time in frequency bands go idle
• Combined TDMA & FDMAOften combined in practice, for example, in cellular phonenetworks:• TDMA cellular phones :– use 30 KHz channels , with each channel divided intothree time slots.– a single handset uses one timeslot for sending and theother for receiving.• GSM uses 200 KHz channelsdivided into eight time slots.A single handset uses one slotin two channels for sending
and receiving.
• CDMA: encoding and decodingCDMA: code division multiple access• unique “code” assigned to each user; ie, code set partitioning• used mostly in wireless broadcast channels (cellular,
satellite, etc)• all users share same frequency, but each user has own“chipping” sequence (ie, code) to encode data– encoded signal = (original data) X (chipping sequence)– decoding: inner-product (summation of bit-by-bitproduct) of encoded signal and chipping sequence.• allows multiple users to “coexist” and transmitsimultaneously with minimal interference (if codes are“orthogonal”)
• TradeoffsAdvantages of Fixed assignment protocols:• Simple protocol to implement • Deterministic response time• No wasted time for coordination messagesFixed assignment(fixedmultiplexing) protocols are ideal for continuous streams such as video or audio.
What about for packet switched data?A “perfect” multiple access scheme would always use the
channel when there are packets waiting (i.e. statistical
multiplexing)Fixed multiplexing protocols divide the channel into Nseparate independent, identical subchannels.So, if we use fixed assignment protocols for packet switcheddata, mean delay goes up by a factor of N!Fixed assignment protocols are not appropriate for multipleaccess in a packet switched network with a large numberof users.Packet arrivals are fairly random, so there will be many times
when packets are waiting at one user while other users areidle.The idle resources (time slots or bandwidth or both are wastedin this case ). Disadvantages of Fixed assignment protocols:
• Wasted bandwidth when some nodes are idle• Network size fixed during installation – adding new
stations is not “plug-n-play”• Prioritization is local to each node• If TDM used, depends on synchronized clocks:requires stable clocks
> Random Access MAC protocolsRandom access (or contention or “transmit and hope”)MAC protocol specifies:• how to sense a carrier• how to detect collisions• how to recover from collisions (e.g., via delayedretransmissions)Examples of random access MAC protocols:ALOHA, CSMA ,CSMA/CD• ALOHA: vulnerable periodALOHA was the first random access protocol, with no carriersense or collision detection.Operation:• When node has packet to send:– transmit at full channel data rate R.– no a priori coordination among nodes• Two or more transmitting nodes: collision, packet corrupted.With ALOHA, packet vulnerability depends on time of flight .Note that if the first bit of a new packet overlaps with the last bitof a packet almost finished, both packets are totally
destroyed.
• CSMA:
In some shorter distance networks, it is possible to listen to thechannel before transmitting.
In radio networks, this is called “ sensing the carrier”The CSMA protocol works just like Aloha except:If the channel is sensed busy , then the user waits to transmitits packet, and a collision is avoided.Human analogy: Don’t interrupt others !This really improves the performance in short distance networks:
– As soon as all network nodes have sensed a packet, itbecomes invulnerable.Vulnerability window is ~End-to-End propagation delay.
 Vulnerable period Collisions can still occur :propagation delay meanstwo nodes may not yearhear each other’s transmission Collision: entire packettransmission time wasted. Note: role of distance and
propagation delay in determining collision probability .Using Carrier sense, multiple stations politely wait until anotherstation finishes , then transmit simultaneously -> collision.If the channel is sensed idle, how long (if at all) should a stationwait before starting transmission?
If the channel is sensed busy, how long should a station persistin sensing carrier?
 Persistence
Waiting a random period may involve waiting longer than
necessary. Note also that this random wait is before a tx attempt,not after a collision‡ p- persistent is time-slotted:Transmissions will only start atdiscrete instants.
• CSMA/CD Carrier sense in CSMA improves performance, but still itwastes the channel during collisions.CSMA/CD: carrier sensing & deferral as in CSMA• Sense during transmission, as well as before.• Abort the transmission as soon as a collision is detected(reducing interval over which channel is wasted)Human analogy: the polite conversationalistCollision detection:• Easy in wired LANs: measure signal strengths, comparetransmitted, received signals • But in wireless LANs, difficult to receive while
transmitting -> use channel reservation instead.
 Advantage of CDContinuous time:– Frame transmission can begin at any instant.– No master clock.Slotted time– Time is divided into discrete intervals (slots).– Frame transmission begins at the end of a previous slot• idle slots, frame slots, collided slots.
 Contention slots CSMA/CD passes through a “contention period” (in whichstations attempt transmission but collide & abort), followedby successful transmission.But how long must the contention period(slot) last before astation can be sure there will be no collisions? Why need to limit the minimum frame length? Why need to limit the maximum frame length? If frames are too large, one station can monopolise the medium;but if a frame is too small, a collision may not be detected– Collision detection requires a minimum size frame so astation can detect a collision before it finishes sending its
frame.– If it detects a collision after the frame is sent, it does notknow if its frame was involved.
How small can a frame be?Example: Assume coax cable with a rate B = 10 Mbps; longest distance L = 2 km; propagation rate P = 200 m/μsec.Then the minimum frame size required:MF = ((2*L)/P)*B = 200bits = 25 bytesDerivation of minimum frame length. Consider the first IEEE standard: 10base5 Ethernet:• Calculate geographical spanMaximum segment length = 500 mConcatenate segments by using repeaters.Maximum span :5 segments (passing through 4 repeaters)-> Maximum LAN span = 2500m• Translate into round-trip time:Signal propagates at 2/3 C = 2E8 m/sRound-trip time = 5 km / 2E8 = 25 μsFudge factor: Double to account for repeater delays: 50 μs• Round-trip time: 50 μs• Translate into frame length:50 μs = time to transmit 500 b for 10 Mb/s
conservatively round up to 512 b-> minimum frame length = 64 BAssuming preamble is part of physical layer,DA+SA+Type+Null Data+CRC=18 B-> pad null data with 46B
 Operation of Exponential BackoffWhen collisions occur, Ethernet uses a random retransmission
scheme, called binary Exponential Backoff:• Goal : adapt retransmission attempts to estimated current load - heavy load: random wait will be longer
Exp. backoff operation:• first collision: choosefixes ” random access stability problem by passing it to the layer above!
 Collision detection time, abort time Need to distinguish collision from noise, since no need for
source to back off exponentially in response to noise-> non-zero detection time.Need to ensure that all stations are aware of the collision-> transmit a “jam signal” (32 or 48 arbitrary bits) afterdetecting the collision to ensure that other nodes also detectit and back off.-> non-zero abort time.
 Inter Frame Gap: what for?Don’t allow a station to transmit continuously, since other
stations would defer indefinitely-> inter Frame Gap between transmitting consecutiveframes – 96 bit times. (9.6 ms for 10 Mb/s Ethernet)Other stations only required to sense for carrier for 2/3 of
inter Frame Gap after end of transmission-> can contend with the next transmission.
 Frame bursting
If:• spending lots of time in contention, or• frames are short compared to time needed to resolve
contention (e.g. Gigabit transmission)then:• may permit stations to transmit multiple frames in atransmission opportunity after they win the contentionprocess.In this case, only the first frame needs to be long enough(padded up to minimum length) for collision detection. Burst transmission must be limited:burst Limit = 8 KB for Gigabit Ethernet
Advantages of contention protocols:• Small latency for low traffic load• Networkinitialization/configuration is not required• Node can enter or leave the network without anyinterruption• Probabilistic global prioritization is possibleDisadvantages of contention protocols:• Designed for aperiodic traffic - not ideal for synchronizedcontrol loops
• Unbounded individual message latency• Poor efficiency under heavy loads• Detecting collisions may require analog circuitry that adds tothe system expense. In fact, if the network environment is very noisy orthe wiring runs are long and poor quality, collision detection may not work at all!
Reservation-based MAC protocolsChannel partitioning (fixed assignment) MAC protocols:
• share channel efficiently and fairly at high load (high # ofactive stations)• inefficient at low load: delay in channel access,1/N bandwidth allocated even if only 1 active node!Random access (contention-based) MAC protocols:• efficient at low load: single node can fully utilize channel• high load: high collision overheadHandshaking (dynamic assignment, contention-free) protocols:• look for best of both worlds!• requires ordering of nodes(token passing) or central failure
point(polling).
• Polling: operation, tradeoffs
Advantages:• Simple protocol to implement; historically very popular(HDLC)• Bounded latency for real -time applicationsDisadvantages:• Single point of failure - centralized master• Polling consumes bandwidth(= polling overhead)• Network size fixed during installation (not robust )– Or, master must discover nodes during reconfiguration• Prioritization is local to each node– But, can use centralized load balancing:Polling doesn’t need to be in strict order -> it could be,
for example: 1, 2, 1, 3, 4, 1, 5, 1, 3, 1, 6, …(repeats)
• Token passing• Token holder = OWNER ; only the owner may transmit• Polling is a centralized form of token passing, where tokenis passed by master and returned to master by slave .• In the decentralized(distributed) token passing, each nodeknows its successor in the polling sequence and sends thepoll directly to that node.• Token is passed as node number or other similar value– may be tacked on to end of data-bearing message Token bus: operation, tradeoffs
Operation:• A token signal is passed from a node to node on a bus(virtual ring, according to station numbers)• Adding and deleting stations to/from the virtual ring
requires additional work.• Examples: IEEE 802.4, Arcnet, AN192, MAP, Profibus
Advantages:• Bounded latency for real-time control applications• High throughput during heavy traffic• On-the-fly reconfiguration
Disadvantages:• Explicit token passing latencies under light traffic conditions
• Prioritization local to each node• Lengthy reconfiguration process
• Token initialization, loss, and duplication recovery overhead
• Collisions may occur during initialization and
reconfiguration• Complex protocol (especially a MAC sublayer)
 Token ring: operation, tradeoffs
• Nodes are connected in a ring using point-to-point links– This is not a circular bus: every wire is independent andoperates concurrently• A token signal is passed from one node to another in acircular fashion • This is a bit-at-a-time transfer protocol– Bits are shifted around the ring– All other MAC protocolsdiscussed deal withwhole messagesExamples:• IEEE 802.5, TRON, FDDI
Advantages:• Bounded latency for real-time control applications• High throughput during heavy traffic• Global and local priority mechanism available• On-the-fly reconfiguration with node bypass hardware• Well suited for fiber optic media• Each station acts as a repeater - strengthens the signal
Disadvantages• Moderate latency for light traffic (explicit token passingoverhead)– (Not as bad as token bus)• Centralized monitor node (designated at initialization)• Token initialization, loss, and duplication recovery overhead
• Propagation delay is based on the number of nodes• Cut in the wire disables the entire network ( unless redundantring is available)
oHybrid MAC protocolsA hybrid of TDMA and Handshaking protocols: Implicit token.• Length of waiting period is used as a time- domain implicit“token”:
– Owner of bus determined by what time it is, instead ofexplicit token message -> Variable Length TDMA!TDMA: uses time slices - waiting period is a whole message long
> Implicit token protocolImplicit token protocol uses time slots -> waiting period is asshort as possible: ~ End-to-End propagation delay
• Operation• IDLE: Active station transmits immediately– After each message, reserve S slots for N nodes• BUSY: Transmit during your assigned slot– If S=N, no collisions - known as Reservation CSMA– If S prioritized messages• Slot1 ..SlotN: rotating slots –> non-prioritized messagesAdvantages:• Small latency for light traffic• Good throughput under heavy traffic• Global prioritization through fixed slots –> prioritizedimplicit token passes• Bounded latency through rotating slots –> non-prioritizedimplicit token passes
Disadvantages:• Restarting time slots from an idle bus can be difficult– Send dummy messages to avoid idle state• Collisions can occur ( if #slots(S) mapping Sth slot to Nth node
> Binary Countdown(CAN) A hybrid of Random access and Handshaking protocols:Binary countdown / Bit dominance protocol.
Operation:• Each node is assigned a unique identification number• All nodes wishing to transmit compete for the channel bytransmitting a binary signal based on their identification value• A node drops out the competition if it detects a dominantstate while transmitting a passive state• Thus, the node with the LOWEST(change picture)identification value winsExamples: CAN, SAE J1850
Advantages:• High throughput under light loads• Local and global prioritizationpossible• Arbitration is part of the message - low overhead
Disadvantages:• Propagation delay limits bus length (2 tpd bit length)
• Unfair access - node with a high priority can "hog" thenetwork• Poor latency for low priority nodes
oComparison of MAC protocol tradeoffs:
Collision-based(Random access)protocols:• An unbounded number of collisions results in unboundedworst-case latency– Idea: use collision to signal start of a reservation CSMAprotocol –> works well• In general not constrained by bit speed /network length ratio (but IS constrained by message speed/network length ratio)
Token-based(Handshake) protocols:• Consumes bandwidth for explicit token passing– Master/Slave polling the worst – individual polling message – Token bus OK under heavy load, if token pass combined with transmission – Token ring is better, but requires special topology • Does not require precise oscillators, especially if used with self-clocking bits • Not specifically constrained by bit speed/network length ratio – But bus topologies are inefficient if network is longer
than a whole message time
Time-based(TDMA and Implicit token) protocols: • Longer timed intervals potentially waste bandwidth – Unused reserved time slices • Any timed interval requires an accurate oscillator at each node – Worst for TDMA – Relevant to Implicit token protocol as well • Constrained by bit speed/network length ratio
Bit dominance(binary countdown) protocols: • Excellent efficiency
– But must have compatible network medium • Constrained by network bit speed/network length ratio Local priority: • Flexible, straightforward to implement
Global priority: requires consensus of nodes to determine winner • Bit dominance does this “for free” • Implicit(time-based) tokens approximate this by very fast
(implicit) token pass to all nodes • Token ring approximates this by very fast (explicit) token pass to all nodes • Explicit token/handshake protocols in general have a difficult time doing this
Global fairness: requires ability to send non-prioritized messages
• Bit dominance and Random access must use emulation of another protocol to do this (e.g., polling) • Implicit token protocols do this by using rotating slots • Explicit token protocols do this as part of token passing –> no additional charge Protocols are optimized for different operating scenarios
Collision-based: • High number of possible transmitters • Low number of active transmitters • Arbitration overhead proportional to activity
Token-based, Time-multiplexed & Polled: • Moderate number of total transmitters • Handles worst case activity without a problem • Arbitration overhead relatively constant
Binary countdown: • Moderately large number of message types • Arbitration overhe ad constant• Global prioritization (but no mechanism for fairness)
•LAN principles
IP (or WAN) address (32-bit IPv4, 128-bit IPv6):
• network-layer address
• used to get datagram to destination network
MAC (or LAN or physical) address:• used to get datagram from one interface to another
physically-connected interface (same network)
• 48-bit MAC address (for most LANs) burned in the adapter ROM Analogy:– MAC address: like Social Security Number – IP address: like postal addressMAC flat address => portability– can move LAN card from one LAN to anotherIP hierarchical address => NOT portable – depends on network to which one attachesStarting at A, given IP datagramaddressed to B:– network layer: looks upnetw. address of B, finds B on same net. as A– link layer sends datagramto B inside link-layer frameEach IP node ( Host , Router ) onLAN has ARP module andARP Table:
IP/MAC address mappingsfor some LAN nodes:
TTL (Time To Live):time after which addressmapping will be forgotten (typically 20 min)
oLink layer vs network layer addressing:
> ARP: table, query, operation
A knows B's IP address, wants to learn physical address of B.• A broadcasts ARP query frame, containing B's IP address– all machines on LAN receive ARP query• B receives ARP packet, replies to A with its (B's) physicallayer address
• A caches (saves) IP-to-physical address pairs untilinformation becomes old (times out)– soft state: information that times out (goes away) unless
refreshedARP is “plug-and-play”: nodes create their ARP tables without
intervention from network administrator .
> Sending IP packet (encapsulated inside Ethernet frame) from one LAN to another
• A creates IP packet with source A, destination B• A uses ARP to get R’s physical layer address for111.111.111.110• A creates Ethernet frame with R's physical address as dest,Ethernet frame contains A-to-B IP datagram• A’s data link layer sends Ethernet frame• R’s data link layer receives Ethernet frame
• R removes IP datagram from Ethernet frame, sees it’sdestined to B• R uses ARP to get B’s physical layer address• R creates frame containing A-to-B IP datagram, sends to B
oCollision domain - the set of networked devices whose framescan collide with one another.Devices connected via coax, repeaters, shared hubs are in thesame collision domain (they can see each other’s frames -easy for sniffing devices to capture passwords/ sensitive data)All devices in a collision domain must share the bandwidth(more devices, less bandwidth)
oBridgesBridges and switches are store and forward class devicesThey read in a frame and then regenerate a complete new one, asif the box were two LAN boards back to backBridges/switches terminate a collision domain, enabling full
bandwidth for each port (each segment or collision domain).Store means delay, up to one whole packet time.Bridging has always been software-based and normally a bridgewould just have two ports used to connect the two LANsbeing bridged.
Frame switching is hardware-based and has many ports but allthe rules that apply to bridging also apply to switching andmore besides .oSwitchesLAN Frame switches can include FDDI, Token Ring or Ethernetswitches.Effectively, the switch provides single Collision Domains perswitch port and each port acts as a bridge port to the rest ofthe network.Forwarding tables are kept per port, different media, differentspeedsetc. can be configured on a port by port basis.The speed enhancement to the network is achieved through the'microsegmentation' of the large Collision Domain into manysmaller ones.Each port on an Ethernet switch is effectively a very fast bridgeport.
> Backpressure scheme: What for?, How?
Some switches allow you to implement a Backpressure schemewhereby, on a particular port, jamming frames can be sent toreduce traffic coming into the switch.
This stops one port hogging the backplane on a switch therebyeffecting other users.Obviously, you would not wish to implement this on a serverport, since thiswill affect many people and you would wishto keep as much of the switchprocessing capability for theattached servers.This is why so much play is made of the backplane capability ofa particular manufacturer's switch.
> Backplane capability vs aggregate forwarding rate
Switch Fabric is the capability of switch backplane to supportmaximum traffic load without blocking.It means the same matter with "Backplane", when you see the
different expression in Specification.Consider that a switch has 10 ports of 100 Mbps.To support full traffic load under full-duplex operation, the
switch backplane must provide at least 2 Gigabit of activity(10 ports x 100 Mbps x 2) to achieve.Then, say that Switch Fabric is 2 Gbps.A switch might have the capability to support full traffic load, ornot. Backplane Capability
– An internal backplane usually interconnects all switch
ports (has fixed capacity ).– If backplane capacity internal blocking may occur (undesirable)
Port mirroring
– Switch ports are in their own collision domain (can’t see
other ports’ unicast frames - good in general)
– Port mirroring (copying frames from one port to another)
needed to troubleshoot network via packet sniffers.
> Cut-through vs Store & Forward switching,
A Cut-through switch first reads the Destination address of a
frame and then sends the frame straight to the destination
before the rest of the frame has arrived at the switch.
The first 20 to 30 bytes of the frame need to be read to make sure
that the frame is not a collision fragment .
Until the destination address remains unknown, the switch
temporarily stores the frame.
Cut-through switching is fine for fixed speed networks such as all
10BaseT, and it is very fast.
However , if the switch has mixed speed ports such as 10/100
autosensing ports, then there is a bottle neck when packets are
moving across the switch fabric from a 100BaseT segment to
a 10BaseT segment.
Some switches, although they forward the frame as soon as they
read the destination address, they still read the frame up to the
CRC.
If there are a certain level of errors, they can be configured to
automatically change to a Store-and-Forward mode.
A Store-and-Forward switch, or 'buffered switch', stores each
complete frame in a buffer before forwarding it on to the
appropriate port.
This gets around the underflow or overflow situation that could
happen in a mixed speed environment.
It takes more time to examine the entire packet, but it allows the
switch to catch certain packet errors and collisions and keep
them from propagating bad packets through the network.
• Fragment-free( modified cut-through) switching
This is similar to Cut Through switching but here the frame is
checked a little further than the destination address to the
Length field in order to weed out collision fragments, before
it is forwarded.
Latency of a network increases as the network gets busier.
On a busy network, the backoffs (retransmits) that could occur
with Cut-through switches increase , thereby increasing
latency.
A Store-and-Forward switch on 10Mbps LAN delays a frame by
one frame time, obviously increasing latency, but there are no
backoffs.
• Ethernet
The original Ethernet:
• Idea by Robert Metcalfe, based on persistent CSMA/CD
• Standardised by Digital, Intel & Xerox(1978) –
Ethernet II or DIX, then by IEEE(1983) – 802.3
• Simpler, cheaper than token LANs and ATM
Persistent CSMA/CD operation
The 802.3 - 2002 standard is 1538 pages long! (mainly because of
various different Physical layers)
Current Ethernets (e.g. Gigabit Ethernet):
• Still offer persistent CSMA/CD
• But are usually used with full duplex point-to-point links,
where MAC is unnecessary!
Today , “Ethernet” is more of a brand than a MAC protocol.
A: sense channel(CSMA), if idle
then else o Framing:
> Preamble, Pad, CRC calculation, end of frame
Preamble: 7 Bytes of “10101010”, then 1B SOF = “10101011”.
• 1010… with Manchester enc. provides one transition per bit,
allowing the receiver to synchronise to the sender’s clock.
Data: ≤ 1500 B to limit RAM needed in NICs, back when RAM
was expensive . 0 B data field, to advertise source liveliness.
Pad: to ensure that transmission lasts long enough to hear whether
1st bit collided before transmitting last bit.
CRC calculation is run as each byte is received
– The calculation stops when the transmitted signal stops.
– CRC checking compares the calculated value with the
received bit pattern at the time the signal stops.
– A CRC error is sent to the controller chip and the
damaged frame is normally discarded.
• An Ethernet frame stops when the transmitted signal stops
– Length of a frame is implied by that end of signal
condition
> IEEE 802.3 vs Ethernet II(DIX): framing differences
Type / Length:
• DIX: Indicate higher layer protocol type (e.g. IP=2048).
• 802.3: Indicate length of Data+Pad+CRC. 802.3 can have
802.2 (LLC) as higher layer: indicates higher layer prot. type.
Data length ≤ 1500 -> helps to distinguish between DIX and 802.3
• DIX or 802.3 receiver can’t distinguish data from padding
-> higher layer protocol must do this (e.g. IP len field).
o Topologies and cabling
a. Coaxial cable.
– Maintenance is complicated by needing broad physical access.
– Data cable is distinct from twisted pair used for phones.
b. Computers connected using twisted pair to wiring closet.
– Hub in closet provides shared medium.
c. Switches become cheaper, and replace hubs in closets,
improving performance.
– Point-to-point links no longer need MAC.
– Later: Links become wireless and topology returns to original
distributed shared medium!
oSpeeds and framing: 10 Mbps / 100 Mbps / 1 Gbps
10Mbps & 100Mbps Ethernet frames look alike, except forclock rate:– Size: 64 B to 1518 B– transmit/receive on separate UTP pairs1Gbps Ethernet largely does too:– Size: 512 B (64 B+pad) to 1518 B– Simultaneous transmit/receive on same UTP pairs(full-duplex cables and switches): no collisions inbetween host and switch.> Twisted pair, Crossover cable
oError conditonsWhen working with Ethernet systems, it is useful to be familiarwith some common problems encountered in actual networks.The maximum transmission unit (MTU) is the maximum framesize allowed on the network.For the Ethernet, the MTU may range from 68 to 1500, withmany ISPs setting their lines to an MTU of 576.If the MTU is too small, large frames willbe fragmented into twoor more smaller frames, contributing to excess use of thebandwidth and increased collisions.
> Runt , Long, Giant, Dribble, Jabber – What causes these? How to distinguish these from normal frames?
Runt• A runt is any transmitted frame whose length is less than the
minimum frame size(64 B).
• Even a frame with a valid FCS is considered runt, if it’s FCS/CRC error, alignment error
Frame Check Sequence (FCS) Error, or CRC error
• This defines a frame which may or may not have the right
number of bits but they have been corrupted between the
sender and receiver, perhaps due to interference on the cable.
• The IEEE 802.3 says that there should be no more than 10-8
errors, i.e. 1 in 82 x 106.
Alignment Error
• Frames are made up of a whole number of octets.
• If a frame arrives with part of an octet missing, and it has a
FCS error, then it is deemed to be an Alignment Error.
• This points to a hardware problem, perhaps EMF on the cable
run between sender and receiver.
> Broadcast storm :
• An incorrect packet broadcast onto a network that causes
multiple stations to respond all at once, typically with equally
incorrect packets which causes the storm to grow
exponentially in severity.
• When this happens there are too many broadcast frames for
any data to be able to be processed.
• Broadcast frames have to be processed first by a NIC above
any other frames.
• The NIC filters out unicast packets not destined for the host
but multicasts and broadcasts are sent to the processor.
• If the broadcasts number 126 per second or above then this is
deemed to be a broadcast storm.
• An acceptable level of broadcasts is often deemed to be less
than 20% of received packets although many networks
survive well enough on higher levels than this.
• The performance lower-specified workstations may be
impacted by as little as 100 broadcasts/second.
• Some broadcast/multicast applications such as video
conferencing and stock market data feeds can issue more than
1000 broadcasts/sec.
> Collisions:
• Collisions are a normal occurrence on an Ethernet network.
• The more devices there are within a segment (Collision
Domain) the more collisions are likely.
• A badly cabled infrastructure can cause unnecessary
collisions due to a device being unable to sense a carrier and
transmitting anyway.
• If a collision rate is high then it may be worth while
considering segmenting the network by way of a bridge or
router.
• This reduces the chance of a collision occurring on each of
the segment thereby releasing more bandwidth for real
traffic. A good guide to use is collisions should not total more than
1% of the frames transmitted.
• Late Collision: Why is it worse than usual collision?
• A Late Collision occurs when two devices transmit at the
same time without detecting a collision.
• This could be because the cabling is badly installed (e.g. too
long) or there are too many repeaters.
• If the time to send the signal from one end of the network to
the other is longer than it takes to put the whole frame on to
the network then neither device will see that the other device
is transmitting until it is too late.
> Jam:
• On detection of a collision, the NIC sends out a Jam signal to
let the other stations know that a collision has occurred.
• A repeater, on seeing a collision on a particular port, will
send a jam on all other ports causing collisions and making
all the stations wait before transmitting.
• A station must see the jam signal before it finishes
transmitting the frame, otherwise it will assume that another
station is the cause of the collision.
• Jamming is a term used to describe the collisions
reinforcement signal output by the hub/repeater to all ports.
• The Jam signal consists of 96 bits of alternating 1s and 0s.
• The purpose is to extend a collision sufficently so that all
devices cease transmitting.
• Jamming is used when dealing with congestion.
• It is an to attempt to eliminate frame loss within the switch
by applying "back pressure" to those end stations or
segments that are consuming the switch buffer capacity.
• One way of accomplishing this is for the switch to issue an
Ethernet "jam" signal when buffers fill beyond a design
threshold level.
• Jam signals normally are the result of collision detection.
• When the sending end systems on the segment receive the
jamming signal, they will back off for a random time period
before attempting a retransmission.
• Each transmitting node monitors its own transmission, and if
it observes a collision (i.e. excess current above what it is
generating, i.e. > 24 mA) it stops transmission immediately
and instead transmits a 32-bit jam sequence.
• The purpose of this sequence is to ensure that any other node
which may currently be receiving this frame will receive the
jam signal in place of the correct 32-bit MAC CRC, this
causes the other receivers to discard the frame due to a CRC
error.
o Inter Frame Gap: why 96-bit length?
The inter-frame gap is a self-imposed quiet time appended to the
end of every frame.
This idle time gives the network media a chance to stabilize, and
other network components time to process the frame.
Following figure shows a sequence of frames separated by the
fixed-size inter-frame gap.
For 10-Mbps Ethernet, the inter-frame gap is 9.6 μs.
– This corresponds to 96 bit times (divide 9.6 μs by 100 ns
per bit).
– Thus, the 576 bits of a minimum-length Ethernet frame
are followed by 96 bit times of silence.
– Dividing 10 Mbps by 672 bits for each frame plus the
inter-frame gap gives a frame rate of 14,880 frames per
second (minimum sized frames).
– Because each frame is followed by 96 bits of silence,
there are a total of 14,880 × 96, or 1,428,480, bits of the
10-Mbps bandwidth (14.28%) lost as a result of the interframe
gap.
o Full duplex ethernet:
Recently a new mode of Ethernet operation called “Full-Duplex”
was defined in the IEEE 802.3x specification.
In this mode, all connections must provide one pair of wires for
transmit and one pair for receive.
This effectively doubles the throughput of the network.
However, it also limits connections to a point-to-point mode.
In point-to-point mode only two devices can be on the same
segment, thus:
– limiting a network to two nodes (using cross -cable), or
– incorporate the use of an Ethernet Switch.
Since the transmit lines of one station are tied to the receive lines
of the other station and visa versa, there is no longer a
possibility of collisions – CS & CD are not needed.
In full duplex mode, CSMA/CD protocol (contention stage) is
not needed, thus the main restriction for transmitting packets
is the “Interframe Delay” period.
Ethernet NICs have circuitry within them that allows full-duplex
operation and bypasses the normal CSMA/CD circuitry.
In addition, Congestion Control (operating similar to Xon/Xoff)
is turned on which 'jams' further data frames if the receive
buffer is filling up:
– A receiving station can send a packet to a sending station
to stop it sending data until a specified time interval has
passed.
• Industrial Ethernet
Emergence of Industrial Ethernet in the new millennium :
– Major automation OEM’s begin incorporating Ethernet
ports into their devices.
– All popular industrial “Fieldbus” protocols ported to
work over Ethernet.
– Major automation OEM’s begin to promote Ethernet for
process control and factory automation.
Ethernet designed for industry…
– Standard IEEE 802.3 in an industrialized design
– Higher temperature ranges
– Rugged and metal housing
– Fan-less products that withstand vibrations
– Industrial connectors and cables – CAT5E
– High speed redundancy
– EMC Immunity
Source: Bill King – Siemens Energy & Automation (ISA 2002 Conference Presentations)
• Good qualitative definition – but the devil is in the details …
– Higher temperature ranges – how much higher?
– EMC Immunity – which standards and what levels?
In addition to previous: higher reliability and availability!
What are industrial (factory floor ) users specifying?
The following excerpt was taken from an RFP specification
document from a major manufacturer who is intent to use
Ethernet on the factory floor:
– Required and Desired Features of Industrial Ethernet
Switches:
• The primary requirement for an industrial Ethernet
switch is that it be environmentally hardened to
operate under the same extremes in operating
conditions (temperature, vibration, humidity, etc.) as
an industrial PLC.
-> Industrial users are often taking the common sense approach
in specifying that the Ethernet networking equipment be us
robust as the IED’s connecting to it.
Enviroment:Office– Normal temperature range– Little dust , moisture and vibration– Hardly any mechanical loads or problems with chemicals
– Low EMC RequirementsIndustry
– Extended temperature range– Dust, moisture and vibration possible– Risk of mechanical damage or problems with chemicals– High EMC requirements
Installation:Office– Fixed basic installation in the building – Variable device connection at standard workstations– Cabling predominantly in star-topology
Industry– Plant dependent cabling and cable ducting– Field attachable connectors
up to IEC IP67 (waterproof)– Redundant cabling,often ring topologies
Data:Office– Large data packets– Medium network availability– Predominantly acyclic transmission– Real-time behavior not necesarryIndustry– Small data packets– Very high network availability– Predominantly cyclic transmission–Real-time behaviour necessaryQ. How do you make Ethernet “deterministic” or how do youprevent collisions from occurring on an Ethernet network?
Ans: You use an Ethernet switching hub (i.e. Switch) with
fullduplex ports – IEEE 802.3x!
oWhat changes would you need to make in your ethernet network, to implement/adopt prioritized messaging?
Standard Ethernet Frame:
Ethernet Frame with new VLAN :
IEEE 802.1Q & Priority IEEE 802.1p Tag