INTEGRATED SERVICES DIGITAL NETWORKS

I. BACKGROUND

Baseband vs Broadband LANs

 baseband LAN -transmit digital signals directly on cable
	- single data conversation on channel at any given time
	- typically cheaper to implement
	- limited in capacity & length

 broadband LAN - use carrier frequencies to modulate signal
	- thus many different channels ==> allows for multiple data conversations 
	simultaneously
	- more expensive to purchase & maintain but provides higher bandwidth 
	transmission & less restrictive length limitations

 - most LANs use twisted pair wire or coaxial cable
 - some use of fiber optic cable for high-speed applications

 eg. Ethernet - 10 Mbps;  Token ring - 16 Mbps -> shielded twisted pair,
 4 Mbps -> unshielded twisted pair;  FDDI - 100 Mbps.

FIG. 1   LANs

Digital Telecommunications

 PBX - Private Branch Exchange
- telecommunications switching system owned by customer
- most widespread "customer-premises" equipment in use today
- most major PBX vendors now offer digital service
- take digitized voice or data on lines
- typically digital phone (not DTMF - dual-tone multifrequency dialing 
which are analog devices) converts voice into digital form for 
transmission to PBX (PCM)
- also can send computer data through digital phone

- LANs & PBXs being connected
- can then connect to packet-switching or frame relay networks for
 long-distance communication


 Facsimile
- originally analog communication for transmitting image of printed page
- newer digital fax machines -> image to 1s & 0s ==> can transmit at 
high-speed over digital communication nw's including PBXs, LANs,
 pkt-switching nw's, frame relay nw's, etc.

video transmission - transmission of moving images
- originally analog (eg. conventional cable tv)
- digital video trans. now possible but need very high bandwidth

High-Bandwidth Digital Transmission

- mentioned briefly at beginning of course
- main example is T-1 carrier service
- developed at AT&T
- provides 1.544 Mbps digital path on all common carriers
- typically 1.544 Mbps service divided into 24 separate channels 
  using time division multiplexing (TDM)

  T-1 frame 
- each frame is 193 bits long - 24 8-bit channels + 1 framing bit
- 8000 frames transmitted per second  i.e. 193 bits x 8000 frames/sec
 = 1.544 Mbps
- each channel is 64 kbps (voice) --> 64 kbps x 24 + 8 kbps

FIG. 2    T-1 frame

- carries data or digital voice using TDM
- by combining T-1 carrier with PCM => 24 voice conversations can be
 simultaneously transmitted over single T-1 circuit requiring only two 
 twisted pairs
  
Note 1:  with analog voice => 24 pairs but do need repeaters to prevent 
fading & special digital transmission equipment at each end of T-1 circuit
- usually free of static unlike analog where noise is amplified & carried
 with signal

Note 2: occasionally need to transmit some signals - so "steal" bit for 
these -> robbed bit signaling  => this channel has only 7 bits, not 8
- this does not affect voice quality - comparable to speck of dust on 
one frame of movie each second but if sending data, this would cause
 errors - so - with data, only use 7 bits/channel for 56kbps per channel

- in contrast, with analog + modem, very expensive to get even 19.2 kbps
 reliably so 56 kbps is good

 T-1 Multiplexers & Submultiplexing

  - any channel can be submultiplexed  eg. take five 9600 bps terminals 
on one channel (which is 64 kbps) => entire T-1 could carry 120 
simultaneous 9600 kbps terminal-to-host computer  connections, all 
on 2 pairs of wire

 - other higher bw standards - less common

FIG 3   Digital Transmission Standards

II. Integrated Services Digital Network - ISDN

- evolving set of standards for Digital Network carrying both voice & 
  data communications
- most residential & small business communications on analog local loops
- only large businesses with high volumes of telecommunications traffic 
  use digital transmission - i.e. T-1

- ISDN makes digital trans. available to all users, large & small using
  many different sizes of digital trans. paths

ITU-T taking lead in setting up collection of standards for ISDN
- plan to have end-to-end digital connections between all devices in
 a world wide integrated digital nw where terminals, computers &
 telephones can all easily attach to the nw

ISDN Channels

- there are different sizes of trans. paths called channels which can be 
 combined in different ways to satisfy user requirements

B-channel - bearer channel - 64 kbps digital channel - can carry data or
 digitized voice

D-channel - digital channel to carry signaling info - 16 or 64 kbps 
   depending on  type of circuit
 - much of current signaling on D channel similar to that on T-1 carrier
 (eg. on-hook/off-hook status, dialing info, busy signal) but T-1 typically
 uses in-band signaling i.e. signaling bits interspersed in same channel as
 data bits vs. ISDN - "out-of-band" signaling or "common-channel" 
 signaling i.e. separate channel for signaling

- since most signaling at beginning & end of connection, D-channel could
 be idle most of time ==> some low-priority user data can be sent on
 D-channel when it is free eg. email in pkts.

- B & D-channels basic building blocks of ISDN but there are other 
  channels def'd:

A-channel - low speed (< 16 kbps) data channel for older, pre-ISDN data
 devices  eg. today's dumb terminal

FIG 4 -ISDN channel types

H-channel - higher bit rates eg. high-speed trunk

Basic-rate Interface - BRI

- for business & residential users
- 2 B & D i.e. 2B + D --> 1 B - digitized voice;  1 B - high-speed data, D 
- signaling & low speed data pkts

- D channel carries 16 kbps
- B channel carries 64 kbps per channel   ==>    64 + 64 + 16 = 144 kbps
   +  48 kbps overhead = 192 kbps capacity but 144 kbps effective bw

* - provides home with simultaneous voice & data communications
    capability and advanced signaling capability

 B- channel --> can have 3 kinds of connections:
   i) circuit switched - user places call, circuit-switched connection made
   with another nw user and call establishment made over D-channel

   ii) packet-switched - user connected to pkt-switching node, data 
   exchanged via X.25

   iii) semi-permanent - connection to another user set up by prior 
   arrangement 
- equivalent to leased line

Note: 64kbps set as standard when needed for digitized voice - now 
can do with < 32 kbps

Primary-rate Interface (PRI)

- provides higher bw service for business users
  e.g. connect PBXs to central offices, PBXs to PBXs, PBXs 
  to LANs, LANs to LANs

 - 23 or 30 B channels for user data + single 64 kbps D-channel for 
  signaling

 - in North America & Japan --> T-1 popular => use 23B + D
 - capacity of each of 23 B-channels -> 64 kbps so have 24 64 kbps
   channels so T-1 equipment can work with PRI

 - since signaling done on D-channel, no bits taken from B-channel for this
 - so, each channel has full 64 kbps capacity

 - framing arrangement similar to T-1 required -> 24 x 64 kbps  +  8 kbps
  =  1.544 Mbps

 in Europe - 3-B + D -> 2.048 Mbps
  - is there incompatibility with NA nw's?   No - since each country or
 continent can use own method of digital transmission as long as there are
 gateways to connect different nw's

E-channel - signaling for circuit-switched traffic used at user-nw 
interface for multiple-access configurations

H-channels
   H0 - 384 kbps;   H11 - 1.536 Mbps;  H12 - 1.92 Mbps where latter
 2 for countries with 2.048 Mbps (ie 30 B + D)
   - used for fast fax, video, high-speed data, high quality audio

ISDN Equipment Functions & Reference Points

- on customer's premises - ISDN Terminal Equipment (TE),  Network
 Termination Equipment (NT)

Note:  for ISDN, terminal equipment refers to any equipment attached
 to end of ISDN circuit - e.g. dumb terminal, personal computer, or any
 device attached to ISDN nw that transmits or receives voice, data, or
 other information

TE1 - terminal equipment compatible with ISDN nw - e.g. ISDN-
compatible digital phone, ISDN-compatible computer port, ISDN-
compatible workstation

TE2 - device not compatible with ISDN e.g. RS-232-C dumb terminal
 -> need Terminal Adapter (TA) to convert

- have reference points to help define user-nw access configuration
        R - rate, S- system,  T - terminal,  U - user

ISDN services

- offers several features to attract users from analog nw
e.g. much information can  be carried on D-channel -> can display both
number & name of caller, can give user choice of answering or ignoring
call, or could have central office block all calls from given caller, if
number being called busy => "camp-on" i.e. indicate connect when
no longer busy

Overview

1. ISDN Physical Layer

   - represented to user at ref-point S or T
   - interface different for BRI & PRI
   - both analog & digital data transmitted using digital signals

Table 1   Digital Signal Encoding
FIG. 5   Digital Signaling Formats

 framing & multiplexing - two 64 kbps B channels, one 16 kbps 
 D channel -> 144 kbps data multiplexed over 192 kbps 
 interface at S or T --> rest of capacity -> framing

 synchronous time-division multiplexed (TDM) scheme
    - repetitive, fixed-length frames
    - each frame 48 bits long --> one frame every 250 microsec. => 192 kbps
    - each frame has 16 bits from each of two B channels & 4 bits from
      D channel - other bits are overhead for framing (balancing)

 Primary Rate User - Network Interface
  - multiplexes multiple channels across single trans. medium
  - only point-to-point configuration allowed
  - typically interface at T ref. point with digital PBX or other concentration
     device controlling multiple TEs & providing synchronous TDM facility for
     access to ISDN
  - 2 data rates defined for PRI --> 1.544 Mbps & 2.048 Mbps

FIG. 6   Primary Access Frame

  - 1.544 Mbps rate based on North America DS-1 trans. structure
  - 193 bit frames
  - each frame -> 24 8-bit time slots + framing bit
  - at data rate of 1.544 Mbps -> repeat once every 125 microsec 
     i.e. 8000 frames/sec so each channel supports 64 kbps
  - typically 23 B channels + 1 D channel
   - other arrangements too --> 24 B channels & various combinations
     of H channels

2. ISDN Data Link Layer

  - LAPD is for communication between subscriber & nw
  - all D channel traffic uses LAPD protocol
  - for B channel traffic - if pkt-switched connection -> LAPB used to 
    connect subscriber to pkt-switched node
   - if circuit switched -> end-to-end circuit between 2 subscribers -
    can use any protocol at link level for end-to-end data-link control

- there are 2 ISDN-related data link control protocols
i) one similar to LAPD
ii) LAPF - supports frame-mode bearer service

LAPD protocol - basic characteristics
 - modeled after LAPB protocol in X-25 & HDLC
 - user info & protocol-control info & parameters transmitted in frames

FIG 7   LAPD formats
FIG. 8   Pkt-Switching vs Frame Relay

Frame Relay

- most important technical innovation to emerge from standardization 
    work on N-ISDN
- streamlined technique for pkt-switching that operates at data link layer
- has significantly less overhead than traditional pkt-switching with
  X.25 interface
- initially designed as service & switching mechanism for N-ISDN
 -> now considered means of service & switching outside ISDN

  * - LAPF is used on top of frame relay for end-to-end error &
   flow control

- congestion control one of most difficult technical issues associated 
   with frame relay
- no mechanism for flow control & error control between user & nw
- LAPF frame does not contain control field & thus no sequence numbers
- allows for efficient data transfer but also for possibility of congestion

Table 2   Frame Relay Congestion Control Technique

3. Network Layer

- main purpose of nw layer is to establish, maintain & clear nw 
connections  e.g. - circuit switched connections on B channel; 
pkt-switched connections using D or B channel; user-to-user 
signaling using D channel

Signaling System Number 7 (SS7)
- set of specifications & protocols for internal control & nw intelligence
 of digital nw
- based on use of common-channel signaling
- designed specifically for ISDN

FIG 9   SS7 Protocol Architecture

 - signaling data link level - corresponds to OSI layer 1
     - specifies full-duplex physical link for (dedicated) SS7 traffic
     - principal option is 64 kbps digital link

  - signaling link level - corresponds to OSI layer 2
     - uses same principles as LAPD & LAPB but formats & some of
       procedures differ

  - signaling nw level - some of functions of OSI layer 3
     - msg handling e.g. discrimination, routing, distribution
     - nw management e.g. traffic management, route management, & link management


  - signaling connection control part (SCCP)
    - completes set of functions associated with OSI layer 3
    - has enhanced addressing capability over signaling nw level
    - supports reliable connection-oriented data transfer

  - ISDN user part (ISUP)
    - defines functions, procedures, & interexchange signaling info flows
       required to provide user-channel services & associated user facilities 
       for voice  & non-voice calls over ISDN

- SS7 is the standard for carrying ISDN signals between switches 
   in the nw
- in widespread use in USA
- SS7 info includes call's origin & destination phone numbers,
  instructions for nw handling call
- signaling technology often used to assist in call routing   e.g. if chain 
  of stores has one toll-free phone number => can route potential customer
  to nearest store

Note: most users still have analog phones so can't take advantage of 
many of these features
  - some users with analog phones have special display for Caller ID but 
    not all other features
  - availability of these features -> debates about privacy rights -> 
    various solutions proposed
  - basically endless number of features for customers from ISDN ->
    new features can be added later by changing the software

National ISDN - Bellcore standards

- goal of insuring consistent deployment of ISDN in user's entire
  private nw
- represented as National ISDN-N for "N" the version
- ISDN-1 & ISDN-2 already deployed, & plans exist for National
 ISDN-3

Video & ISDN

  - various standards committees developing ideal methods for transmitting moving
    & still images
- one such committee on video or motion pictures is ISO's Motion Pictures Experts
   Group or MPEG -> provides variety of standards  with differing quality levels
    depending on bit rates
- uncompressed broadcast quality video -> 21 Mbps bw
- MPEG II --> compressed --> 7 Mbps bw (but neither BRI or PRI ISDN can provide
   this bw)
- but broadcast quality not needed for video conferencing  e.g. using BRI
 -> use both B channels for video & audio

- also have PRI video conferencing using 23B channels
- can accomplish video conferencing by only sending part that moves
   e.g. mouth, hands
- since many users don't require broadcast quality video have what is 
   called Nx64 or multirate ISDN

	i.e. use multiple B-channels for single connection
	   H0 	 6 B-channels  -->     384 kbps
	   H11   24 B-channels -->  1.536 Mbps
	   H12   30 B-channels -->  1.920 Mbps

Future of ISDN?

- based on proven technologies
	BRI - similar to PBX digital phones
	PRI - 23B + D similar to T-1 carriers

- to be fully implemented -> new customer-premises equipment needed
- who will pay?  - in some countries, gov't

- even though T-1 & ISDN similar (PRI), since giving up one of 24 data
 channels, get incredible signaling power   e.g. 2 PBXs at 2 locations
 -> tie together with ISDN PRI & signaling between them would make
  it look like single PBX to users

  - but only small fraction of users of telephones use custom calling  services
  - will they want ISDN services?  - advocates say can use data  channel to send
      messages from office re turning on AC, VCR etc. but - since < 1/2 bank customers
      use ATMs & even fewer can program VCR -> doubts

III. Broadband ISDN (BISDN)

  - BRI & PRI originally considered very high bw interfaces but with improved
     transmission technology & fiber optic cable, much faster transmission rates now 
     possible

 - high bw of B-ISDN would allow digitized video signals together with digitized voice
   & data

- integrate telephone & cable tv nw's into single nw for businesses & residences  e.g.:

Business -> inexpensive high quality video conferencing -> digital telephones

Residential -> rather than go rent movie -> transmit & automatically deduct cost from
    bank account
Work -> office workers could work at home, attend meetings via video conferencing, 
   transfer work to boss' computer

-  N-ISDN -> adds data channel & signaling, uses digital voice  transmission, + other
   services, but B-ISDN could revolutionize way we communicate

- thus most observers see businesses but not residential users, using N-ISDN but unable
   to resist B-ISDN

 - infrastructure necessary for success significant -> fiber optic cable  replacing copper in 
   long distance lines
- "fiber to curb" popular since copper can support high bw for short  distances

 Key developments in technology needed for B-ISDN
  - optical fiber transmission --> low-cost, high-data rate trans. channels  for nw trunks &
    subscriber lines

 - microelectronic circuits that offer high-speed, low-cost building blocks for switching, 
   transmission, & subscriber equipment

- high quality video monitors & cameras, that, with production quantities,  can be low cost

 - these technological developments will support universal communications with following 
   characteristics:

- worldwide exchange between any 2 subscribers in any medium or combination of media
- retrieval & sharing of massive amounts of info from multiple sources, in multiple media, 
   among people in shared electronic environment
- distribution, including switched distribution, of wide variety of cultural, entertainment, &
   educational materials to home or office, virtually on demand

Table 3   Requirements for B-ISDN
Table 4   Principles of B-ISDN

 Transmission Structure - 3 services

  i) full-duplex 155.52 Mbps
    - supports all N-ISDN services i.e. one or more BRI or PRI
    - can support most of B-ISDN services - i.e. can support one or several video channels 
      depending on resolution & coding technique

  ii) asymmetrical service
      - subscriber to nw -> 155.52 Mbps
      - nw to subscriber -> 622.08 Mbps

  iii) full-duplex 622.08 Mbps
    - handles multiple video distribution provider

- experts expect (i) to be most popular

B-ISDN Protocols

 *Note: for B-ISDN transfer of info across user-nw interface uses asynchronous transfer 
 method (ATM) & thus is packet-based nw at interface & almost certainly for internal 
 switching

- recommendations (ITU-T) state B-ISDN will support circuit-switched mode -> will be
   over pkt -based transport mechanism
 - more on ATM below

 Overview

- physical layer - 2 sublayers -> physical medium sublayer, transmission  convergence
   sublayer

- ATM layer - pkt transfer capabilities - similar to X.25 but ATM uses common channel 
  signaling

- AAL layer - maps higher layer info into ATM cells & collects  from ATM cells for 
   delivery to higher layers - e.g. PCM - string bits -> cells for transmission

- ref model has 3 separate planes

i) user plane -> provides for user info transfer & associated controls  - e.g.. flow control, 
  error control

ii) control plane - call control & connection control functions

iii) management plane - plane management i.e. management functions related to system as
 whole & coordination between all planes
- layer management - i.e. management functions relating to resources & parameters in 
  protocol entities

B-ISDN Physical Layer

	
Table 5   B-ISDN Physical Medium Characteristics

CMI - binary 0 - always positive transition at midpoint of binary unit time interval i.e.
 signal at lower level for first half & 2nd half at higher level - binary 1 - constant level for 
 duration of bit time; alternates between high & low for successive binary 1's

optical - binary 1 -> emission of light; 0 -> no emission of light

transmission structure:

  i) continuous stream of cells with no multiplex frame structure

 - synchronization on cell-by-cell basis i.e. receiver must properly mark  53-octet cell 
   boundaries
 - alignment done using header error-control (HEC) field
 - if HEC calculation indicates no errors => assume proper alignment
 - occasional error ok, but if stream of errors => indicates alignment off => recover 
    alignment

ii) put cells in synchronous time-division multiplex envelope

- bit stream has external frame based on Synchronous Digital Hierarchy
 (SDH) --> in America frame structure is SONET (synchronous  optical nw)

- SDH frame may be used exclusively for ATM cells, or, may also carry other bit streams 
  not yet defined in B-ISDN
- SDH standard defines hierarchy of data rates all of which are multiples of 51.84 Mbps &
    including 155.52 Mbps & 622.08 Mbps


  SONET - optical transmission interface originally proposed by  BellCore & standardized
   by ANSI

- compatible version by ITU-T is SDH
- establishes standard multiplexing format using any number of  51.84 Mbps signals
   as building blocks

- establishes optical signal standard for interconnecting equipment from different suppliers
- establishes extensive operations, administration & maintenance (OAM) capabilities as part 
   of standard
- defines synchronous multiplexing format for carrying lower-level digital signals (DS1, 
   DS2, ITU-T standards)
- establishes flexible architecture capable of accommodating future applications


- basic SONET building block is STS-1 frame - consists of 810 octets, transmitted every 
  125 microsec --> overall data rate of 51.84 Mbps

- can logically view as matrix of 9 rows of 90 octets each, with transmission being one row 
  at time, from left to right, & top to bottom

- first 3 columns (3 octets x 9 rows = 27 octets) of frame --> overhead; 
   9 octets -->section-related overhead;  18 octets --> line overhead

FIG 10   SONET STS-1 Overhead Octets
Table 6   STS-1 Overhead Bits

Pointer Adjustment 

 - in conventional circuit-switched nw's, most multiplexers & telephone company channel 
   banks require demultiplexing & remultiplexing of entire signal just to access piece of info
   addressed to a node  e.g. for T-1 MUX B that receives data on single T-1 circuit from T-
   1 MUX A => passes on to MUX C

- in signal received, single DSO channel (64 kbps) addressed to node B;
 rest will pass on to C & then on into rest of nw

- to remove single DS0 channel, B demultiplexes every bit of 1.544 Mbps
 signal, removes data, remultiplex every bit

- some T-1 MUXs allow for drop-and-insert capability i.e. only part of  signal has to be
 demultiplexed & remultiplexed but this equipment  will not communicate with other 
 vendor's

SONET - standard drop-and-insert capability --> applies 64 kbps & higher data rates
- uses set of pointers that locate channels within payload & entire payload within frame
- thus, info can be accessed, inserted, & removed by simple adjustment of pointers
- pointer info contained in path overhead that refers to multiplex structure of channels 
   contained within payload
- synchronous payload environment (SPE) of STS-1 frame can float with respect to frame
- actual payload (87 columns x 9 rows) can straddle 2 frames - H1, H2 octets in line 
   indicate start of payload

FIG 11   Representative location of SPE in STS-1 frame

IV. Asynchronous Transfer Mode (ATM)

 * - most important technical innovation from standardization work on B-ISDN
    - also known as cell relay
*  - transmission technique using fixed-size cells
*  - less overhead than frame relay - designed to operate at significantly 
      higher data rates than frame relay

- one of most difficult technical issues for ATM is congestion control
 - even more so than for frame relay

- first work on standards assumed synchronous TDM would be used (as for ISDN):

	(j x H4) + (k x H2) + (l x H1) + (m x H0) + (n x B) + D

for D, B, H0, & H1 (H11 or H12) are N-ISDN and H2 --> 30-45 Mbps &
   H4 --> 120 - 140 Mbps

- but determined this was not best model even though would be  natural extension
   --> two reasons why not

i) does not provide flexible interface for variety of needs

 - at high data rates of B-ISDN --> could be wide variety of applications &
  many different data rates that must be switched but one or two fixed-rate 
  channel types do not give structure to support this

- and, many data (vs voice or video) applications are bursty in nature which better handled
  by pkt-switched approach

- fixed-rate of synchronous approach not amenable to rate adaption (data stream < 64kbps,
  for e.g., mapped onto 64 kbps data stream)  -> complex for 64 kbps channel but much
 more complex & inefficient to extend for channels of 10's to 100's of megabits/sec

  ii) at high-speed transmission, synchronous approach complicates switching system i.e.
  would need switches that can handle data streams of multiple high data rates vs. N-ISDN
 --> just 64 kbps data stream to switch

 - thus synchronous TDM rejected & asynchronous transfer mode (ATM) proposed

Note: "asynchronous" in ATM does not refer to physical transmission of bits --> 
  transferred synchronously
   - "asynchronous" refers to fact that next cell does not arrive at predictable, fixed rate 
    similarly to how characters arrive in asynchronous data transmission

* - consider bit stream to be continuous, but data sporadic - 2 methods to do this:
   
if no user data to be transmitted => either special pattern bits representing
	i) idle line, or,  ii) idle cells  transmitted
  
 - thus receiver gets continuous stream of bits - some of cells with user  payload, others 
    representing idle time (idle bits or cells)

 - physical transmission medium can add overhead bits to separate timing from data & to 
   help position cells within continuous bit stream

 - ATM good when combining data from many sources
  e.g. user with 64 kbps voice connection, 10Mbps Ethernet data  stream, 7 Mbps 
   compressed video transmission -> could break into   ATM cells & transmit on same
   ATM 45 Mbps link

- such a large pipe through which users can transmit data at varying rates 
--> referred to as "bandwidth on demand"
 - however, there can be difficulties - cf highway works fine except during rush hour

 - unlike N-ISDN where fixed channels set up for duration of call,  ATM accepts cells on
   one side & delivers them on other side of nw 
- so have classical statistical multiplexing problem of how to allocate bw according to 
   demands & needs of sporadic use  i.e. nw for low  bandwidth voice, high-bandwidth
    video or both at same time?

Pricing Issues

  - traditional nw's price voice calls on distance, connection time & time of day
  - ATM must consider type of service --> video call more than voice call but price not 
     proportional to number of bits transmitted  e.g. 2-hour movie using MPEG II video at
    7 Mbps -> about 50.4 gigabits
- with voice could talk 4 hours, 20 minutes for 1 gigabit!
- if charge $3 to view movie => 6 cents/gigabit but unreasonable to charge 6 cents for the 
   phone call

  - carriers will need to balance between call set-up, QOS or bw, call  duration, time-of-day 
    usage charges
  - one suggestion -> bandwidth contract - user pays certain rate to guarantee availability of 
    given bw, but can burst to higher speeds if additional demands & nw has space => 
    premium rates for bursting above guaranteed rate at high traffic times

ATM Overview

  - similar in concept to frame relay but at higher data rate & more stream lined in its 
    functionality
* - pkt-oriented transfer mode 
- like frame relay & X.25, allows multiple logical connections to be  multiplexed over 
   single physical interface
- info flow on each logical connection organized into fixed-size pkts --> cells
- no link-by-link error control or flow control

FIG 12 - ATM Transport Hierarchy

- transmission path level  
- extends between nw elements that assemble & disassemble payload of transmission 
   system
- for end-to-end communication --> payload = end-user info
- for user-to-nw communication --> payload could be signaling information

- digital section

- extends between nw elements that assemble & disassemble continuous bit or byte stream
  i.e. exchanges or signal transfer points in nw involved in switching data streams

- regenerator section

- e.g. repeater used to simply regenerate digital signal along transmission
 path that is too long to be used without such regeneration
 - no switching

Virtual Channels & Virtual Paths

virtual channel (VC) - logical connection in ATM
- analogous to virtual circuit in X.25 or frame relay logical connection
* - basic unit of switching in B-ISDN
 - virtual channel set up between two end users through nw & variable-rate, full duplex 
    flow of fixed-size cells exchanged over connection (i.e. end-to-end)

- virtual channels also used for user-nw exchange (control signaling)  & nw-nw exchange
  (nw management & routing)

Virtual path (VP)
- second sublayer of processing
* - bundle of virtual channels that have same endpoints => all cells flowing over all  virtual 
   channels in single virtual path switched together
- response to increased cost of control in high-speed networks
- VP technique helps contain control cost by grouping connections sharing common paths 
  through nw in single unit

- nw management actions applied to small number of groups of connections instead of large
 number of individual connections

FIG 13  -  ATM Connection Relationships
FIG 14  -  Call Establishment Using Virtual Paths
Table 7 -   Virtual Path/Virtual Connection Terminology


Note:  for individual virtual channel setup, control involves checking that there is a virtual 
path connection to the required destination node with sufficient available capacity to support 
VC with appropriate QOS  =>  storing required state info (VC/VP mapping)

-within nw, may be number of points at which VC switched & at these points, VCI may be
   changed
- thus, VCC consists of concatenation of one or more VC links with VCI remaining 
constant for extent of VC link & changing at VC switch points

- between endpoint & VC switching point, or, between 2 VC switch points, VP connection
(VPC) provides route for all VC links that share the 2 VPC endpoints

- at this level, may be internal switching such that VPC passes through one or more VP 
   switch points with VPI changing at each such point

- thus, VPC consists of concatenation of one or more VP links

- VP switches terminate VP links
- VP switch translates incoming VPIs to outgoing VPIs according to destination of VPC
- VCI values remain unchanged
- VC switches terminate VC links & necessarily VP switches

FIG 15  -  Representation of VP & VC switching Hierarchy

- end points of VCC may be end users, nw entities, or end user & nw entity
 - in all cases, cell sequence integrity preserved within VCC i.e. cells delivered in same 
   order in which they are sent

 - thus, VC switch must switch both VPs & VCs & so both VCI & VCI translation 
    performed


Control Signaling

 - in N-ISDN, D channel for control signaling of calls on B & H channels

 - in B-ISDN with ATM interface, no simple fixed-rate structure of H, B & D channels, so 
   more flexible arrangement for control signaling needed, and, need to allow for 
   establishment & release of 2 entities:  VCs & VPs

For VCs  - one or a combination of following:

i) semipermanent VC for control signaling

ii) if no pre-established call-control signaling channel  =>  one must be set up, so control-
   signaling exchange needed between user & nw on some channel => need permanent low 
   data-rate channel for this -->  meta-signaling channel

iii) meta-signaling channel used to set up VC between user & nw for  call-control signaling
   =>  user-to-nw signaling VC used to set up VCs to carry user data



   iv) meta-signaling channel also used to set up user-to-user signaling VC  -->  must be set 
    up on pre-established VP  =>  can be used to  allow 2 end users, without nW 
    intervention, to establish & release user-to-user VCs to carry user data

 For VPs  -->  3 methods

  i)  VP on semipermanent bases by prior agreement - so no control  signaling required
  ii) VP establishment/release may be customer-controlled  so customer uses signaling VC 
       to request VP from nw
  iii) VP establishment/release may be nw controlled  so nw establishes VP for its own 
       convenience - path can be nw-to-nw, user-to-nw, user-to-user

ATM Adaption Layer

 - ATM creates need for adaption layer to support info transfer protocols not based on ATM
 - e.g. PCM voice, LAPD

PCM - stream of bits  -->  assemble into cells for transmission & to read them out on 
 reception such that produce smooth, constant flow of bits to receiver

LAPD  -  standard data link control protocol for ISDN & B-ISDN  -->  map LAPD frames 
  into ATM cells, so usually segment one LAPD frame into number of cells on transmission
  & reassemble frame from cells on reception

 - by allowing use of LAPD over ATM, all of existing ISDN application & control 
    signaling protocols can be used on B-ISDN

AAL Services

 - general examples:
- handling of transmission errors
- segmentation & reassembly, to enable larger blocks of data to be carried in info field of
   ATM cells

- handling of lost & misinserted cell conditions
- flow control & timing control
 - to handle these introduced 4 classes of service based on whether timing relationship must 
   be maintained between source & destination, whether  application requires constant bit 
   rate, whether transfer connection-oriented or connectionless

 	e.g. class A --> circuit emulation
	       class B --> variable bit-rate video
	       class C & D  -->  data transfer

FIG 16  -  Service Classification for AAL

Bell South - Integrated Services Digital Network, Information Guide, 1995

- 128 kbps (uncompressed) ==> 2:1 compression - 256 kbps; moving to 4:1 & 8:1

- greater accuracy - basically error free so better fax & clearer phone - "quiet"

- can supply ISDN on same copper wire in use today

- advantage of D-channel (out-of-band signaling)

- speed of calls - 1-3 seconds after last digit dialed --> ring vs 10-30 seconds now

- "bond" together B channels to give rate needed

- with ISDN - expensive digital lines no longer only way to get digital speeds

   - modems to convert digital pulses of computer to analog pulses for analog nw, no longer 
     needed

   - digital signals can flow from one digital device to another, through totally digital, yet 
      inexpensive dialed connection

   - most often all available through existing phone copper wires

- LAN connectivity - limits of LANs (geographically & number of users) removed - e.g.
   large company - each floor, one LAN
 - ISDN link together as if one

- makes telecommuting more efficient - estimated 51 million work some at home

- individual connections during day ==> large file transfer at night over bonded links

- can get "overflow" lines for fraction of cost of leased lines when needed e.g. Avis uses 
   for backup

- can use for distributed security watch

- fax - 10 to 20 times faster than analog line

- whiteboards, video depositions, remote consultations for medicine (especially remote 
   areas), distance learning (interactive with classroom)

- Frank Sinatra recent album "Duets" - 6 of 13 artists appearing "phoned in" contributions 
   via ISDN their parts e.g. Aretha Franklin
 - Detroit; Tony Bennett - NY; Liza Minelli - Brazil

- ISDN capabilities at all competitive sites at Olympics

- ISDN in advertising - send proofs all over country/world --> e.g. company can finish in 4 
  hours what ordinarily took 7 days

- BP OIL --> ISDN BRI lines at  >500 stations - mostly for credit cards
 - takes 8 seconds vs 25-30 seconds for analog

- BellSouth's ISDN PRI Service provided through standard T-1 dedicated lines (point-to-
  point but have minute-by-minute control of each 64 kbps channel 

Asymmetric Digital Subscriber Line (ADSL)

- can provide high-speed Internet connections, telephone, video-on-demand, networking, 
   & fax on twisted-pair copper wiring (i.e. phone lines)

- theoretically deliver these simultaneously

- signal processing with digital filters make up for BW shortcomings of thin copper phone 
   wire

- developed by Bellcore Labs (NJ) -> also developed ISDN

- ADSL originally developed for interactive TV market in 1987

- now test adapters-> 6.14 Mbps downloading, 640 kbps  uploading

i.e. download speed > 200 x 28.8 kbps modem

- does not work well with symmetric applications -> e.g.videoconferencing which requires
   large data streams to & from sites

- predictions that sales will go from $50M this year to $2.5 Billion by 2000
-originally developed for residential real-time video -> but companies now want for remote 
  offices & telecommuting

ISDN vs ADSL

- took about 20 years to develop ISDN

- when phone companies upgraded to digital lines => could handle ISDN

- eventually realized solution for customers who wanted more BW

- say ISDN will meet the needs of 90% of customer base -> 6 Mbps speed of ADSL is 
  "like using a blow torch to heat a cup of coffee"

- 2 competing systems for ADSL (comparable to VHS vs Betamax battle):

1) CAP - Carrier Amplitude & Phase Modulation
 - works on 2 channels - if asymmetrical => larger one -> download, smaller one -> upload

- cheaper

2) DMT - Discrete Multitone Technology

- spectrum divided into 256  4-KHz carriers

- variable number of bits loaded onto each carrier independently

- many carriers, so flexible

- ANSI adopted ADSL standard based on DMT

- DMT appears to be faster than CAP

- DMT prototype transceivers have high transmission speeds, simplicity, reliability, good 
  line noise characteristics & seems to be more robust than CAP

- CAP supporters are requesting ANSI include CAP standards with ADSL

- ADSL will help telephony providers compete in Internet services market against cable tv 
   companies

B-ISDN vs IP

- Internet responds to "killer apps:"
  
previously: email, file transfer, remote terminal access

currently:  WWW

future? - real-time voice & video

- ATM, fiber to curb, SONET developed to support BISDN

- ATM has become core NW to interconnect routers e.g. Internet MCI in Atlanta, biggest 
  provided of backbone Internet services, installed  nationwide core NW to interconnect its 
  routers

- now most long-haul U.S. Internet traffic runs over ATM

- ISDN digitizes the loop but that meant nationwide ISDN needed to be developed at great 
  expense since, 64 kbps channel is end-to-end

- now BISDN just gearing up to run on ATM & SONET to integrate any service onto 
  single NW

but original version of BISDN depends on single world-wide NW

- on the other hand, Internet is network-of- networks architecture which is impossible to 
  integrate -> but IP works over it

- ATM & SONET now becoming preferred switching & transmission technologies for 
  wide area broadband parts of Internet

i.e. "BISDN is happening - except it's spelled IP"