Data Communication

ISDN (Integrated Services Digital Network) is a set of ITU-T standards (I-series) for a fully digital network that carries voice, data, fax and video over the same digital subscriber line, using standard interfaces and common-channel signalling. It gives end-to-end digital connectivity.

Features

• Digital transmission from the user to the exchange.
• Several services on one line, with several simultaneous calls.
• Out-of-band signalling on a separate D channel (DSS1 to the user, SS7 inside the network).
• Standard user–network interfaces and fast call setup.

Channels

Access types

• Basic Rate Interface (BRI): 2B + D = 2×64 + 16 = 144 kbps (192 kbps with framing at the S/T point); for homes and small offices.
• Primary Rate Interface (PRI): 30B + D = 2.048 Mbps (Europe/Nepal, E1) or 23B + D = 1.544 Mbps (North America, T1); D = 64 kbps; for PBXs and large users.

Architecture (reference model)

        R        S        T        U
 [TE2]---[TA]--+
               |--[NT2]---[NT1]=======[LT/ET]
 [TE1]---------+  (PBX)   (NTU)  2-wire  exchange
               S bus    4-wire   loop

• TE1: ISDN terminal; TE2: non-ISDN terminal with TA (terminal adapter).
• NT1: physical/line termination at the customer site; NT2: intelligent device such as a PBX.
• R, S, T, U: reference points between these units; U is the two-wire loop to the exchange.

Services

• Bearer services: circuit-mode and packet-mode transport.
• Teleservices: telephony, fax, videotex, teletex.
• Supplementary services: caller ID, call waiting, call forwarding, conference.

Advantages and limitations

• Faster and better-quality connections than analog modems; several services on one pair; quick call setup.
• Data rates are low by modern standards; costly equipment; largely replaced by DSL and fibre. Broadband ISDN (B-ISDN), based on ATM, was the high-speed extension.

Soft switching

Soft switching is a switching method in which call control is done by software on a general-purpose server (the softswitch or call agent), separate from the hardware that carries the media. The softswitch controls media gateways over an IP network, so one software platform can handle PSTN, VoIP and mobile calls. It is the core of next-generation networks (NGN).

Architecture

+---------------------------------------------+
| Application layer: app servers, IN, billing |
+----------------------+----------------------+
                       | SIP / APIs
+----------------------+----------------------+
| Control layer: SOFTSWITCH (call agent/MGC)  |
+-----+----------------+---------------+------+
      | H.248/MGCP     | SIGTRAN       | SIP
+-----+-----+   +------+-----+   +-----+-----+
|  Media    |   | Signalling |   | IP phones |
|  gateway  |   |  gateway   |   | / SIP UA  |
+-----+-----+   +------+-----+   +-----------+
      |  transport layer: IP/MPLS core   |
   PSTN trunks       SS7 network

1. Softswitch / Media Gateway Controller (MGC): call control, routing, signalling interworking, number analysis.
2. Media Gateway (MG): converts TDM voice (PSTN) to RTP/IP packets and back; done under control of the MGC via H.248/Megaco or MGCP.
3. Signalling Gateway (SG): converts SS7 signalling to IP (SIGTRAN, M3UA/SCTP) for the softswitch.
4. Application server: value-added services (voicemail, prepaid, conference, IN services).
5. Media server: tones, announcements, IVR, conference mixing.
6. Transport (IP/MPLS) network that carries both media and signalling.

Functions

• Call setup, supervision and release (call control).
• Signalling interworking: SS7, ISUP, SIP, H.323, H.248.
• Routing and number translation.
• Control of media gateways and allocation of resources.
• Service invocation through application servers.
• Generation of CDRs for billing; authentication and authorisation.

Management

The softswitch network is managed by an OSS/NMS using FCAPS:

Management uses SNMP and web/CLI interfaces, often remotely.

Advantages: lower cost, open interfaces, easy addition of new services, scalability, and one platform for voice and data.

IP telephony (Voice over IP, VoIP) is the transmission of voice calls as data packets over an IP network (internet or private IP network) instead of over dedicated circuits of the PSTN. Voice is digitised, compressed, put into IP packets and sent; at the far end the packets are reassembled and played out.

Basic operation

 Speech -> A/D + codec -> packetise (RTP/UDP/IP)
                              |
                         IP network
                              |
 Speech <- D/A + decoder <- jitter buffer <- depacketise

1. Digitisation: the analog voice is sampled at 8 kHz and coded.
2. Compression (codec): G.711 (64 kbps), G.729 (8 kbps), G.723.1 (5.3/6.3 kbps), Opus.
3. Packetisation: e.g. 20 ms of speech per packet, with RTP, UDP and IP headers.
4. Transmission over the IP network; packets may take different paths.
5. Jitter buffer at the receiver smooths variable delay.
6. Decoding and playout; lost packets are concealed.

Signalling sets up and ends calls: SIP (IETF) or H.323 (ITU-T), plus MGCP/H.248 for gateways. RTP/RTCP carry voice and quality reports.

Components

• IP phones and softphones (terminals).
• Gateway: converts between PSTN (TDM) and IP for calls to normal phones.
• Gatekeeper (H.323) / SIP proxy and registrar: address translation, admission control, routing.
• IP-PBX / softswitch: call control for an enterprise or carrier.
• MCU: multipoint conferencing.

Call configurations

Quality of service issues

• Delay: one-way delay should be below about 150 ms (ITU-T G.114) for good conversation.
• Jitter: variation in packet delay; handled by jitter buffers.
• Packet loss: causes gaps; handled by loss concealment, FEC.
• Echo: needs echo cancellers.
• QoS techniques: DiffServ priority marking, MPLS, bandwidth reservation.

Advantages

• Much lower cost, especially for long distance and international calls.
• One network for voice, video and data (convergence).
• Better bandwidth use through compression and silence suppression.
• Rich features: video, presence, conferencing, mobility of number.

Disadvantages

• Quality depends on network; delay and jitter can be noticeable.
• Needs power and internet; emergency calls and location are harder.
• Security threats: eavesdropping, spam over IP telephony, denial of service.

VoIP is now the basis of enterprise telephony, operator NGN/IMS cores and VoLTE in 4G mobile networks.

Soft switching is a switching technique in which call control is performed by software on a server called a softswitch, separated from the hardware that carries the voice. It is the core of next-generation (IP-based) networks.

     [Softswitch / Call agent]
      |H.248/MGCP     |SIGTRAN
 [Media gateway] [Signalling gateway]
      |                 |
 PSTN trunks        SS7 network
      \_____ IP / MPLS core ____/

Main elements

• Softswitch (Media Gateway Controller): call setup, routing, signalling and billing records.
• Media gateway: converts TDM voice to IP packets (RTP).
• Signalling gateway: carries SS7 signalling over IP.
• Application and media servers: services, announcements, conferencing.

Protocols: SIP, H.323, H.248/Megaco, MGCP, SIGTRAN.

Advantages

• Lower cost and space than circuit switches.
• Open, standard interfaces; vendor independence.
• Fast introduction of new services.
• Same network carries voice, video and data.
• Easy to scale by adding servers and gateways.

A datagram network is a packet-switched network in which each packet (datagram) is treated independently; there is no connection setup. Every packet carries the full source and destination address, and each router decides the route for it separately using its routing table. The internet (IP) is a datagram network.

       +--R2--+
 A ----R1     R4---- B
       +--R3--+
 packets 1,3 via R2; packet 2 via R3
 B may receive 1,3,2 (reordering)

Features

• Connectionless: no setup or teardown phase.
• Each packet may follow a different path.
• Packets may arrive out of order, be duplicated or lost; higher layers (e.g. TCP) fix this.
• Routers keep no per-connection state; routing tables depend only on the destination address.
• Resources are not reserved, so delay varies.

Advantages

• No setup delay; good for short, bursty traffic.
• Robust: if a router or link fails, packets are routed around it.
• Efficient use of links.

Disadvantages

• Larger header (full address) in every packet.
• Variable delay, reordering, no guaranteed QoS.

ADSL (Asymmetric Digital Subscriber Line) is a DSL technology that gives high-speed internet over the existing copper telephone pair while the same pair still carries ordinary telephone (POTS) calls. It is asymmetric: downstream (to the user) is much faster than upstream, which suits web browsing and video. Standard: ITU-T G.992.1 (up to about 8 Mbps down, 1 Mbps up); ADSL2+ (G.992.5) gives up to 24 Mbps down.

Working principle: frequency division of the copper line

The copper pair can carry frequencies up to about 1.1 MHz over short loops. ADSL divides this band:

 POTS  guard   upstream        downstream
 |---|------|----------|------------------------|
 0   4k    25k       138k                    1104k Hz

Modulation: DMT (Discrete Multitone). The band is divided into 256 sub-channels of 4.3125 kHz each. Each sub-channel uses QAM, and the modem measures the signal-to-noise ratio of each sub-channel and loads more bits on good ones and fewer (or none) on noisy ones. This adapts to line length and noise. (Some early systems used CAP modulation.)

Modem connection and topology

 Customer premises            Telephone exchange
 [PC]--Eth--[ATU-R modem]       +----------+
              |                 | Splitter |--> PSTN
 [Phone]--[Splitter]==copper====|  (POTS)  |    switch
                     pair       |   DSLAM  |--> ISP /
                   (< ~5 km)    | (ATU-C)  |   internet
                                +----------+

• ATU-R (ADSL Transceiver Unit, Remote): the ADSL modem at the user's home; connects to PC or router by Ethernet/Wi-Fi/USB.
• Splitter / microfilter: a low-pass/high-pass filter that separates voice (below 4 kHz) from data, so telephone and internet work at the same time.
• Local loop: the existing twisted pair; speed falls with distance (useful up to about 5 km).
• DSLAM (DSL Access Multiplexer): at the exchange; contains many ATU-C modems, one per line, and aggregates their traffic onto a high-speed link (ATM or Ethernet) to the ISP network.
• POTS splitter at the exchange sends voice to the telephone switch.

The topology is a star: each subscriber has a dedicated line to the DSLAM, so bandwidth on the loop is not shared, unlike cable modem networks.

Operation

1. On power-up the ATU-R and ATU-C perform training: they test each DMT sub-channel and agree the bit loading.
2. User data from the PC is framed, error-coded (Reed–Solomon, interleaving) and modulated onto the upstream sub-channels.
3. The DSLAM demodulates and forwards the data to the ISP; return traffic is sent on the downstream sub-channels.
4. Voice continues in the 0–4 kHz band through the splitters, unaffected.

Advantages: uses existing copper, always-on, simultaneous voice and data, dedicated loop. Limitations: speed depends on distance and line quality; upstream is limited.

Routing is the process by which a node (router or packet switch) decides the outgoing link for a packet so it reaches its destination. Routing algorithms are of two kinds: non-adaptive (static) and adaptive (dynamic).

Non-adaptive (static) routing

Routes are computed in advance (off-line) and loaded into the routers; they do not change with current traffic or topology, unless an administrator changes them.

Examples

1. Shortest path routing: the network is a graph with link costs (distance, delay); Dijkstra's algorithm finds the least-cost path from each source, and the tables are fixed.
2. Flooding: every incoming packet is sent on every outgoing line except the one it came on; hop counters or sequence numbers stop endless copies. Very robust, used in military networks and for broadcasting link-state updates.
3. Flow-based routing: uses known average traffic between node pairs and link capacities to choose routes that minimise mean delay.
4. Fixed/static routes configured by hand (e.g. default route).

Features: simple, low processing and no routing traffic, predictable; but cannot react to failures or congestion.

Adaptive (dynamic) routing

Routes are changed automatically in response to changes in topology and traffic load. Routers exchange information and recompute paths periodically or when a change occurs.

Types by where decisions are made

• Centralised: a routing control centre collects information and sends tables to all nodes.
• Isolated: each node decides from local information only, e.g. hot-potato routing (send on the shortest queue) or backward learning.
• Distributed: nodes exchange information with neighbours.

Main distributed algorithms

1. Distance vector routing (Bellman–Ford): each router keeps a table of the best known distance to every destination and the next hop, and periodically sends it to neighbours. Used in RIP. Simple but suffers from the count-to-infinity problem and slow convergence.
2. Link state routing: each router learns its neighbours and link costs, floods link state packets to all routers, builds the full topology and runs Dijkstra. Used in OSPF and IS-IS. Converges fast, larger memory and CPU.

   A --2-- B --1-- D
   |       |       |
   5       2       1
   |       |       |
   C --3-- E ------+
 If link B-D fails, adaptive routing
 finds A-B-E-D; static routing drops.

Comparison

Switching is the method of connecting a sender to a receiver through intermediate nodes of a network, so that every pair of users does not need a dedicated link. Data communication uses three main switching techniques: circuit switching, message switching and packet switching (datagram and virtual circuit).

               Switching
     +-------------+--------------+
  Circuit       Message         Packet
                           +------+------+
                       Datagram    Virtual circuit

1. Circuit switching

A dedicated physical path is set up between the two users before data transfer and kept for the whole call. The telephone network (PSTN) uses it.

Three phases

1. Circuit establishment: signalling sets up a path through the switches.
2. Data transfer: data flows continuously at a fixed rate over the reserved path.
3. Circuit disconnect: the path is released.

 A --[S1]======[S2]======[S3]-- B
   setup    dedicated path   release

• Space-division or time-division (TDM) switches are used.
• Advantages: fixed, small delay; no congestion once connected; data arrive in order; good for real-time voice.
• Disadvantages: setup delay; channel capacity is wasted during silence; both ends must use the same rate; a busy network blocks new calls.

2. Message switching

There is no dedicated path. The whole message is sent with a destination address to the next node, which stores it, checks it and forwards it when the next link is free (store-and-forward). Used in early telegraph and e-mail relay systems.

 A --msg--> [N1 store] --msg--> [N2 store] --msg--> B

• Advantages: good line efficiency (links shared); no blocking (messages queued); priorities and code/speed conversion possible; broadcast possible.
• Disadvantages: long and variable delay; nodes need large storage (disk); not suitable for interactive or real-time traffic.

3. Packet switching

The message is divided into small packets (e.g. up to 1500 bytes) each with a header. Packets are stored and forwarded through the network; links are shared statistically. Packets are small, so they can be held in memory and pipelined over successive links, giving much less delay than message switching.

(a) Datagram approach

• Each packet is routed independently; no setup.
• Packets may take different routes, arrive out of order or be lost; the destination reorders them.
• Example: IP / the internet.
• Robust to node failure, no setup delay; but variable delay and no guaranteed order.

(b) Virtual circuit approach

• A logical route (virtual circuit) is set up first with a call-request/call-accept exchange; each packet carries a short VC identifier.
• All packets follow the same path in order; the circuit is cleared at the end.
• Examples: X.25, Frame Relay, ATM, MPLS (label switched paths).
• Allows QoS and in-order delivery; but setup delay and failure of a node breaks the VC.

 Datagram:      1 -> R1-R2-R4    2 -> R1-R3-R4
 Virtual circ.: 1,2,3 -> R1-R2-R4 (all same path)

Delay comparison (sending a long message over 3 hops)

 Circuit:  |setup|=========data========|
 Message:  |==msg==|==msg==|==msg==|
 Packet:   |p1|p2|p3|
              |p1|p2|p3|
                 |p1|p2|p3|   (pipelined)

Packet switching delivers sooner than message switching because packets overlap on successive links.

Comparison of techniques

Choice of technique

• Real-time, constant-rate traffic (voice) suits circuit switching.
• Bursty computer data suits packet switching, the basis of today's networks.
• Modern networks combine both ideas: e.g. VoIP carries voice in packets but MPLS virtual circuits give it a guaranteed path.

A communication protocol is a set of agreed rules and conventions that govern how data are formatted, transmitted, received and interpreted by two or more communicating entities. Without a common protocol, devices can send bits to each other but cannot understand them, just as two people need a common language to talk.

Key elements of a protocol

1. Syntax: the format and structure of data, e.g. first 8 bits are the address, next 8 bits the control field, then data.
2. Semantics: the meaning of each field and the action to take, e.g. what a "NAK" means.
3. Timing: when data should be sent and how fast; sequencing and speed matching.

Example: HDLC frame

| Flag | Address | Control | Data | FCS | Flag |
  8b      8b        8b      var   16b    8b

Why a protocol is required (functions)

1. Encapsulation: adds headers (address, control, error-check bits) to user data.
2. Segmentation and reassembly: breaks long messages into blocks of suitable size and rebuilds them at the receiver.
3. Connection control: sets up, maintains and releases logical connections (e.g. TCP three-way handshake).
4. Ordered delivery: sequence numbers allow the receiver to put data back in the correct order.
5. Flow control: prevents a fast sender from overrunning a slow receiver (stop-and-wait, sliding window).
6. Error control: detects errors (CRC, checksum) and recovers by retransmission (ARQ) or correction.
7. Addressing: identifies source, destination, ports and connections.
8. Multiplexing: many connections share one link or one lower-layer connection.
9. Synchronisation: both ends agree on timing, state and frame boundaries.
10. Transmission services: priority, QoS, security (encryption, authentication).
11. Interoperability: equipment from different vendors and countries can work together.

Layered protocol architecture

Because communication is complex, protocols are arranged in layers; each layer uses the services of the layer below and serves the layer above. Peer layers talk using their protocol.

  Host A                          Host B
 Application  <-- HTTP -->    Application
 Transport    <-- TCP  -->    Transport
 Network      <-- IP   -->    Network
 Data link    <-- Eth  -->    Data link
 Physical     ============    Physical

Types of protocols

• Direct or indirect: data passed directly between two stations, or through switches/routers.
• Monolithic or structured (layered).
• Symmetric (peer-to-peer) or asymmetric (client/server, master/slave).
• Standard (ISO, ITU-T, IETF) or non-standard (vendor specific).

Example

When a browser loads a web page, HTTP defines the request and reply format; TCP divides data into segments, numbers them, and retransmits lost ones; IP routes each packet; Ethernet frames carry packets on the LAN with a CRC. If any of these rules were missing, data would be lost, mixed up or not understood.

ISDN

ISDN (Integrated Services Digital Network) is an ITU-T standard digital network that provides end-to-end digital connections for voice, data, fax and video over the same subscriber line, using standard user–network interfaces and out-of-band (D-channel) signalling.

Architecture

        R        S        T        U
 [TE2]---[TA]--+
               |--[NT2]---[NT1]=======[LT/ET]
 [TE1]---------+  (PBX)   (NTU)  2-wire  exchange
               S bus    4-wire   loop

Functional groups

• TE1 (Terminal Equipment 1): ISDN-compatible terminal (digital phone, ISDN PC card).
• TE2: non-ISDN terminal (analog phone, RS-232 device).
• TA (Terminal Adapter): converts TE2 signals to ISDN format.
• NT1 (Network Termination 1): layer-1 device at the customer premises; terminates the two-wire loop, line coding, power, timing.
• NT2: intelligent layer 2–3 device such as a digital PBX or LAN switch that concentrates terminals.
• LT/ET: line termination and exchange termination at the central office.

Reference points

• R: between TE2 and TA (non-ISDN interface).
• S: between TE1/TA and NT2 (4-wire, up to 8 terminals on a bus).
• T: between NT2 and NT1.
• U: between NT1 and the exchange (2-wire loop, 160 kbps for BRI using 2B1Q coding).

Services offered

1. Bearer services: transport of information without change: circuit-mode (64 kbps speech, unrestricted digital) and packet-mode (X.25 on B or D channel).
2. Teleservices: complete services including terminal functions: telephony, telefax (Group 4), teletex, videotex, videotelephony.
3. Supplementary services: added to basic services: calling line identification, call waiting, call forwarding, call hold, conference call, closed user group, advice of charge.

Channels

Interfaces

• BRI: 2B + D = 144 kbps user rate (192 kbps at S/T with overhead).
• PRI: 30B + D = 2.048 Mbps (Europe, Nepal) or 23B + D = 1.544 Mbps (North America).

ISDN (Integrated Services Digital Network) is a fully digital network defined by ITU-T (I-series Recommendations) in which voice, data, fax and video share the same digital subscriber line. Signalling is carried separately on a D channel, giving fast call setup and many supplementary services.

Main features

• End-to-end digital connection over the existing copper loop.
• Standard user–network interfaces (BRI, PRI) and reference points (R, S, T, U).
• Several devices (up to 8) on one line and simultaneous calls.
• Common-channel signalling: DSS1 (Q.931) on the D channel to the user, SS7 in the network.

Types of ISDN channels

1. B channel (bearer channel), 64 kbps

• Carries user information: PCM voice, circuit-switched data, packet data (X.25), fax.
• Four kinds of connection on B: circuit-switched, packet-switched, semi-permanent (leased), and multiplexed lower-rate data.
• Two B channels can be combined (bonding) to give 128 kbps.

2. D channel (data/signalling channel), 16 or 64 kbps

• 16 kbps in BRI, 64 kbps in PRI.
• Carries common-channel signalling for the B channels using LAPD (layer 2) and Q.931 (layer 3).
• When idle, it can carry low-speed packet data and telemetry.

3. H channels (high-speed)

Used for video-conferencing, high-speed data, LAN interconnection and fast fax.

Channel structures (access interfaces)

 BRI:  | B1 64k | B2 64k | D 16k |  = 144 kbps
 PRI:  | 30 x B 64k | D 64k | sync | = 2.048 Mbps

BRI suits homes and small offices; PRI connects PBXs and large users. B-ISDN (ATM based, 155/622 Mbps) was the broadband extension.

ISDN (Integrated Services Digital Network) is a digital network standardised by ITU-T that provides end-to-end digital connections for voice and non-voice services over the subscriber line, through a small set of standard user–network interfaces.

ISDN reference configuration and interfaces

        R        S        T        U
 [TE2]---[TA]--+
               |--[NT2]---[NT1]=======[LT/ET]
 [TE1]---------+  (PBX)   (NTU)  2-wire  exchange
               S bus    4-wire   loop

Functional groups: TE1 (ISDN terminal), TE2 (non-ISDN terminal), TA (terminal adapter), NT1 (line termination at customer site), NT2 (PBX or controller), LT/ET (exchange side).

Interfaces (reference points)

When there is no NT2, S and T merge into the S/T interface.

User–network access interfaces

1. Basic Rate Interface (BRI): 2B + D(16 kbps) = 144 kbps; with framing, 192 kbps at S/T. For homes and small offices.
2. Primary Rate Interface (PRI): 30B + D(64 kbps) = 2.048 Mbps (E1) in Europe/Nepal; 23B + D = 1.544 Mbps (T1) in North America. For PBXs and large customers.

ISDN channels

A virtual-circuit (VC) network is a packet-switched network in which a logical connection is set up between source and destination before data transfer. All packets of the session follow the same route and arrive in order, although link capacity is still shared with other users (it is "virtual", not a dedicated circuit). Examples: X.25, Frame Relay, ATM, MPLS.

Three phases

1. Setup: a call-request packet travels through the switches; each switch makes an entry in its table and assigns a virtual circuit identifier (VCI) for the incoming and outgoing port.
2. Data transfer: each packet carries only the short VCI; switches look it up and forward (label swapping).
3. Teardown: a clear packet removes the table entries.

 Switch table:
 In port | In VCI | Out port | Out VCI
    1    |   14   |    3     |   22
 A --[VCI 14]--> S1 --[VCI 22]--> S2 --> B

Features

• Short header (VCI instead of full address).
• In-order delivery; resources and QoS can be reserved at setup.
• Setup delay; if a switch fails, the VC is lost and must be re-established.

Types: switched VC (SVC, set up per call) and permanent VC (PVC, configured by the operator).

IP switching is a technique that combines the speed of Layer-2 (ATM/label) switching with the flexibility of Layer-3 IP routing. Normal routers examine the full IP header of every packet and look up a large routing table, which is slow. An IP switch routes only the first packets of a flow; then it sets up a fast hardware path for the rest.

 +-------------------------------+
 | IP switch controller          |
 | (routing, flow classifier)    |
 +---------------+---------------+
                 | GSMP
 +---------------+---------------+
 | ATM / label switching fabric  |
 +-------------------------------+

Operation

1. Packets first go on a default path to the controller, which routes them normally.
2. The controller identifies long-lived flows (same source, destination, ports).
3. Using IFMP (Ipsilon Flow Management Protocol) it asks the upstream node to send that flow on a new label/VC.
4. Through GSMP it sets up a connection in the switching fabric.
5. Later packets of the flow are switched in hardware without IP lookup; short flows stay routed.

Advantages

• High throughput and low delay.
• QoS for selected flows.
• Uses low-cost switching hardware.

This idea led to tag switching and MPLS, used in today's ISP backbones.

DSL

DSL (Digital Subscriber Line) is a family of technologies that give high-speed digital data over the ordinary copper telephone pair (local loop). A voice call uses only 0–4 kHz, but the copper pair can carry frequencies up to a few MHz over short distances; DSL uses this extra bandwidth with advanced modulation. The family is often written xDSL.

Features: always-on connection, uses existing copper, dedicated line to the exchange (not shared), speed falls with loop length and noise.

ADSL

ADSL (Asymmetric DSL), ITU-T G.992.1, gives a higher downstream rate than upstream, matching internet use, and keeps the telephone working on the same pair.

Frequency plan

 voice   upstream      downstream
 |---|--|---------|----------------------|
 0  4k 25k      138k                  1104k Hz

Modulation: DMT (Discrete Multitone): the band is split into 256 sub-carriers of 4.3125 kHz; each uses QAM with bits loaded according to its SNR.

Connection

 [PC]-[ADSL modem]-+                 +-> PSTN switch
                 [Splitter]==copper==[Splitter]
 [Phone]-----------+    loop        +-> DSLAM -> ISP

• ADSL modem (ATU-R) at the customer.
• Splitter/microfilter separates voice and data.
• DSLAM at the exchange holds the ATU-C modems and connects to the ISP.

Advantages: cheap broadband using existing lines, simultaneous voice and internet. Limitations: distance limit (about 5 km), low upstream rate, line-quality dependent.

ADSL (Asymmetric Digital Subscriber Line) is a DSL technology that sends high-speed data over the existing copper telephone pair. It is called asymmetric because the downstream rate (exchange to user) is much higher than the upstream rate (user to exchange). This suits home users, who download far more than they upload.

Key features

• Uses the same twisted pair as the telephone; voice and data work at the same time.
• Typical rates (ITU-T G.992.1): up to about 8 Mbps downstream and 640 kbps–1 Mbps upstream. ADSL2+ (G.992.5) reaches up to 24 Mbps downstream.
• Works up to about 5.5 km from the exchange; the rate falls as the loop gets longer.
• Always-on connection; no dial-up needed.

Frequency division of the line

 0-4 kHz     25-138 kHz        138-1104 kHz
+-------+   +----------+   +------------------+
| Voice |   | Upstream |   |    Downstream    |
+-------+   +----------+   +------------------+

The line uses DMT (Discrete Multitone) modulation: the band up to 1.104 MHz is split into 256 sub-channels of 4.3125 kHz each. Each sub-channel carries QAM, and more bits are loaded on clean sub-channels and fewer on noisy ones.

ADSL system

Phone --+                            +--> PSTN switch
        |                            |
     [Splitter]==copper pair==[Splitter]
        |                            |
PC--[ATU-R modem]           [DSLAM (ATU-C)]--> Internet

• Splitter / microfilter: separates the low-frequency voice band from the high-frequency data band.
• ATU-R: ADSL modem at the customer side.
• DSLAM: at the exchange, holds many ATU-C modems and joins the data of many users onto the ISP backbone.

Advantages: reuses existing copper, no new cabling; dedicated line (not shared like cable); voice and Internet together. Limitations: speed depends on distance and line quality; low upstream rate; crosstalk in bundled cables.

Example: Nepal Telecom's ADSL broadband service gave home users Internet over their existing landline pair.

DSL (Digital Subscriber Line) is a family of technologies that send high-speed digital data over the ordinary copper twisted pair of the telephone local loop. A voice channel uses only 0–4 kHz, but the copper pair can carry signals up to about 1 MHz or more over short distances. DSL uses this unused bandwidth with advanced modulation, so the same line gives telephone and broadband Internet together.

Basic arrangement

Home                          Exchange
Phone--+                     +--> Voice switch
       [Splitter]--loop--[Splitter]
PC--[DSL modem]              +--> DSLAM --> ISP

• DSL modem (ATU-R) at the user's end.
• Splitter/filter separates voice and data bands.
• DSLAM (DSL Access Multiplexer) at the exchange collects data from many lines and connects to the Internet.

Types of DSL (xDSL)

Modulation: DMT (Discrete Multitone) is the standard for ADSL/VDSL; CAP (Carrierless Amplitude Phase) and 2B1Q were used in early ADSL and HDSL/ISDN.

Advantages

• Uses the existing telephone copper, so deployment is cheap and fast.
• Always-on; voice and data at the same time.
• Each user has a dedicated loop to the exchange, so bandwidth is not shared on the access link.

Limitations

• Rate falls sharply with loop length and line quality.
• Crosstalk between pairs in the same cable; loading coils and bridge taps must be removed.
• Lower capacity than fiber (FTTH), which is now replacing DSL.

Routing is the process of choosing a path for packets from the source node to the destination node through a network of routers (packet switches). The routing algorithm is the part of the network layer software that decides on which output line an incoming packet should be sent. It builds and updates the routing table of each node.

Desirable properties of a routing algorithm

• Correctness and simplicity.
• Robustness: keeps working when links or nodes fail or traffic changes.
• Stability: routes should converge to a fixed set and not oscillate.
• Fairness and optimality: gives fair service to all and minimises a cost such as delay or hop count.
• Low overhead: little control traffic and processing.

Optimality principle and sink tree

If router J is on the optimal path from I to K, then the optimal path from J to K also lies on the same route. Hence the optimal routes from all sources to one destination form a tree rooted at the destination, called the sink tree. Routing algorithms try to find and use sink trees.

Classification

           Routing algorithms
          /                  \
 Non-adaptive (static)    Adaptive (dynamic)
  - Shortest path          - Centralized
  - Flooding               - Isolated (hot potato)
  - Flow-based             - Distributed
                              . Distance vector
                              . Link state

1. Non-adaptive (static) routing: routes are computed in advance (offline) and loaded into routers; they do not change with traffic or topology.

• Shortest path routing (Dijkstra): the network is a graph; each link has a weight (distance, delay, cost); the path of least total weight is chosen.
• Flooding: every incoming packet is sent on every outgoing line except the one it came from. A hop counter or sequence numbers stop endless copies. Very robust (used in military networks and to spread link-state packets) but wasteful.
• Flow-based routing: uses both topology and the known average traffic between node pairs to choose routes that minimise average delay.

2. Adaptive (dynamic) routing: routes change automatically as topology and load change.

• Centralized: a Routing Control Centre collects information from all nodes, computes routes and sends tables to routers. Optimal but a single point of failure.
• Isolated: each node decides using only local information, e.g. hot potato routing sends the packet on the shortest output queue.
• Distributed: each node exchanges information with others.
  • Distance vector (Bellman-Ford): each router keeps a table of (destination, distance, next hop) and periodically sends it to its neighbours. A router updates its entry as $D_x(y) = \min_v \{c(x,v) + D_v(y)\}$. Simple, but slow convergence and the count-to-infinity problem. Used in RIP.
  • Link state: each router discovers its neighbours and link costs, floods a link state packet to all routers, so every router knows the whole topology and runs Dijkstra's algorithm to find shortest paths. Fast convergence, no count-to-infinity; used in OSPF and IS-IS.

3. Hierarchical routing: in large networks routers are grouped into regions (areas, domains). A router knows the full detail of its own region and only one route to each other region, which keeps tables small.

4. Broadcast and multicast routing: send one packet to all or to a group of hosts, using spanning trees or reverse path forwarding.

Comparison of the two main dynamic methods

Metrics used

Hop count, delay, bandwidth, load, reliability and cost. In telephone-type data networks the route is fixed at call set-up (virtual circuit); in datagram networks each packet is routed separately, so routing must be fast and adaptive.

Good routing gives low delay, high throughput, balanced load and survives failures, which is why it is the central function of the network layer.

Both are dynamic (adaptive) routing methods. In link-state routing every router learns the whole map and runs Dijkstra's algorithm; in distance-vector routing every router only exchanges distance tables with its neighbours and uses the Bellman-Ford equation.

Example network used below (link costs shown):

        A
     2 / \ 6
      /   \
     B --3-- C
     |     / |
    7|   2/  |4
     |   /   |
     D --1-- E

Link-state routing (Dijkstra)

1. Each router finds its neighbours (HELLO packets) and measures the cost of each link.
2. It builds a link state packet (LSP): its ID, sequence number, age and the list of neighbours with costs.
3. LSPs are flooded to all routers; each router stores them in a link-state database, so all routers have the same topology.
4. Each router runs Dijkstra with itself as root:
   • Put the source in the permanent set with $D=0$; set other $D$ = link cost or $\infty$.
   • Repeatedly make permanent the tentative node with the least $D$, then for each neighbour $v$ update $D(v) = \min(D(v),\ D(u) + c(u,v))$.

Dijkstra from A:

Routing table of A: to B via B (2), C via B (5), D via B (7), E via B (8). Used in OSPF and IS-IS.

Distance-vector routing (Bellman-Ford)

1. Each router keeps a vector of its best known distance and next hop to every destination; at start it knows only its direct links.
2. Periodically (e.g. every 30 s in RIP) it sends its vector to its neighbours only.
3. On receiving vectors, router $x$ updates each destination $y$ using

$$D_x(y) = \min_{v \in \text{neighbours}} \{\, c(x,v) + D_v(y) \,\}$$

4. Repeating this, the tables converge to the shortest paths.

For A, neighbours are B (cost 2) and C (cost 6), with converged vectors $D_B = (A2, C3, D5, E6)$ and $D_C = (A5, B3, D2, E3)$:

$$\begin{aligned} D_A(C) &= \min(2+3,\ 6+0) = 5 \ \text{(via B)} \\ D_A(D) &= \min(2+5,\ 6+2) = 7 \ \text{(via B)} \\ D_A(E) &= \min(2+6,\ 6+3) = 8 \ \text{(via B)} \end{aligned}$$

This is the same result as Dijkstra, reached in 3 exchange rounds (A's vector after rounds: E = 10, 9, 8).

Comparison

Here "asynchronous / synchronous DSL" is read as the usual asymmetric and symmetric DSL.

Asymmetric vs symmetric DSL

Modems used in ADSL

An ADSL link has an ATU-R (ADSL Transceiver Unit – Remote) modem at the customer and an ATU-C (Central) modem inside the DSLAM at the exchange. A splitter at each end separates voice (0–4 kHz) from data.

PC--[ATU-R]--+--copper loop--+--[ATU-C in DSLAM]--ISP
Phone--------+   (splitter)  +--------------PSTN

Two modulation methods have been used:

1. CAP (Carrierless Amplitude/Phase)

• A form of QAM in which the carrier is not sent; the signal is built digitally by two orthogonal filters.
• Bands: voice 0–4 kHz, upstream about 25–200 kHz, downstream about 250 kHz–1.1 MHz.
• Simple and cheap, but one wide band is badly hit by noise at any frequency. Used in early ADSL.

2. DMT (Discrete Multitone) – the ITU-T G.992.1 standard

• The 0–1.104 MHz band is divided into 256 sub-channels of 4.3125 kHz.
• Channel 0 is voice; channels 1–5 are guard band; about 25 channels (≈26–138 kHz) carry upstream and the rest (≈138 kHz–1.1 MHz) carry downstream, some kept for control.
• Each sub-channel uses QAM. During training the modem measures SNR on each sub-channel and loads more bits on good ones and fewer or none on noisy ones (bit loading).
• Implemented with FFT/IFFT (like OFDM).

amplitude
 |V|    upstream   |      downstream
 |o|  |||||||||||  ||||||||||||||||||||||
 +-+--+---------+--+--------------------+--> f
 0 4k 26k     138k                   1104k

DMT adapts to each line, so it gives higher and more reliable rates than CAP and is now the standard ADSL modem technique.

A routing algorithm decides the output line on which a router forwards each packet so that it reaches its destination. Routing algorithms are of two main kinds: static (non-adaptive) and dynamic (adaptive).

Static (non-adaptive) routing

Routes are computed in advance and loaded into the routers (by the administrator or an offline program). They do not change with traffic or topology until changed manually.

• Shortest path routing: the network is drawn as a graph with link weights (distance, delay, cost); Dijkstra's algorithm finds the least-cost path, which is stored in the table.
• Flooding: each packet is sent on all output lines except the input one; a hop counter limits copies. Very robust, used where reliability is vital or for broadcasting.
• Flow-based routing: uses known average traffic between nodes and link capacities to pick routes with least average delay.

Advantages: simple, no routing overhead, predictable, secure. Disadvantages: cannot react to failures or congestion; unsuitable for large changing networks.

Dynamic (adaptive) routing

Routers exchange information and change routes automatically when topology or load changes.

• Centralized: one Routing Control Centre computes all tables.
• Isolated: each node uses only local data (e.g. hot potato: shortest queue).
• Distributed:
  • Distance vector (Bellman-Ford): each router sends its distance table to its neighbours periodically and updates using $D_x(y)=\min_v\{c(x,v)+D_v(y)\}$. Simple but slow convergence and count-to-infinity. Example: RIP.
  • Link state (Dijkstra): each router floods its link costs to all routers, builds the full map and computes shortest paths. Fast convergence. Example: OSPF.

Example: if link B–C fails, a static router keeps sending packets to B–C and they are lost; a dynamic router learns of the failure and shifts traffic to another path.

Comparison

Both are packet-switching methods. In datagram switching each packet is routed independently; in virtual circuit switching a logical path is set up first and all packets follow it.

A call from a PC running Skype to an ordinary telephone crosses two different networks: the packet-switched Internet (IP switching) and the circuit-switched PSTN. A softswitch and gateways join them.

PC (Skype client)
   | RTP/UDP/IP voice + SIP signalling
   v
Internet routers (IP switching)
   |
   v
+---------------+  H.248  +------------------+
| Media Gateway |<------->|    Softswitch    |
| pkt <-> TDM   |         | call control,    |
+---------------+         | routing, billing |
   | E1 / PCM             +------------------+
   v                               | SIP
PSTN exchange <--SS7 ISUP-- Signalling Gateway
   |
   v
Telephone

Role of IP switching

• Voice is digitised and compressed by the Skype client (e.g. SILK/G.729), put in RTP/UDP/IP packets (about 20 ms of speech each).
• Internet routers switch these packets hop by hop using the destination IP address, with no fixed circuit. Packets reach the operator's gateway (Skype's "SkypeOut" partner or the telecom operator).

Role of the softswitch

A softswitch is call-control software running on servers that does the job of a classic exchange's control part, separate from the media path.

• Receives the call request (dialled number) from the Skype server/SIP.
• Authenticates the user and checks credit (billing).
• Chooses the route and the gateway nearest the called number.
• Controls the media gateway using MGCP/H.248 (Megaco).
• Through the signalling gateway, converts SIP/IP signalling into SS7 ISUP messages (IAM, ACM, ANM, REL) for the PSTN.

Call steps

1. User dials +977-1-xxxxxxx in Skype; the request goes over IP to the operator softswitch.
2. Softswitch verifies the account, selects a gateway and trunk.
3. Signalling gateway sends ISUP IAM to the PSTN exchange; the exchange rings the phone and returns ACM.
4. When the phone is answered, ANM comes back; the softswitch tells the media gateway to connect the RTP stream to a TDM time slot.
5. Media gateway converts packets to 64 kbps PCM (G.711) and back, adding jitter buffering and echo cancellation.
6. At hang-up, BYE/REL messages release the slot; the softswitch records the call for billing.

So IP switching carries the voice packets across the Internet, while the softswitch provides the intelligence (signalling, routing, billing) and controls the gateways that join the Internet to the PSTN.

An adaptive routing algorithm changes its routes automatically when the topology or traffic changes. Link-state routing is adaptive because every router keeps measuring its link costs, floods a new link state packet (LSP) whenever a cost changes or a link fails, and every router then re-runs Dijkstra's algorithm on the updated map.

Steps of link-state routing

1. Discover neighbours (HELLO packets).
2. Measure the cost (delay/load) of each link, e.g. by ECHO packets.
3. Build an LSP (router ID, sequence number, age, neighbours and costs).
4. Flood the LSP to all routers.
5. Each router computes shortest paths with Dijkstra.

Steps 2–5 repeat periodically and on any change, which is what makes it adaptive.

Example network (normal state)

        A
     2 / \ 6
      /   \
     B --3-- C
     |     / |
    7|   2/  |4
     |   /   |
     D --1-- E

Dijkstra at A (cost, previous node):

Routes from A: C via A–B–C (5), D via A–B–C–D (7), E via A–B–C–D–E (8). All traffic leaves A on link A–B.

Change 1: link B–C becomes congested (cost 3 → 10)

B and C measure the higher delay and flood new LSPs. A re-runs Dijkstra:

New routes: C via A–C (6), D via A–C–D (8), E via A–C–D–E (9). Traffic to C, D and E now leaves on link A–C and avoids the congested link.

Change 2: link C–D fails

(B–C back to cost 3.) C and D stop getting HELLO replies and flood LSPs without that link. A re-runs Dijkstra and gets: C via A–B–C (5), D via A–B–D (2 + 7 = 9) and E via A–B–C–E (5 + 4 = 9).

The routing tables follow the state of the network without any manual change, so link-state routing (e.g. OSPF) is an adaptive routing algorithm.

ISDN (Integrated Services Digital Network) is an ITU-T standard digital network that gives end-to-end digital connections for voice, data, fax and video over the telephone subscriber line, through a small set of standard user–network interfaces.

Structure (reference model)

TE1 ------------ S --[NT2]-- T --[NT1]-- U --[LT/ET]
TE2 -R-[TA]----- S

• TE1: ISDN terminal (digital phone); TE2: non-ISDN terminal; TA: terminal adapter.
• NT2: customer switching (PBX, LAN); NT1: line termination at customer premises.
• LT/ET: line and exchange termination at the exchange.

Reference points: R (TE2–TA, non-ISDN e.g. RS-232), S (TE–NT2), T (NT2–NT1), U (NT1–exchange, 2-wire loop).

Channels

• BRI: 2B + D(16) = 144 kbps (192 kbps with framing).
• PRI: 30B + D(64) = 2.048 Mbps (Europe/Nepal) or 23B + D = 1.544 Mbps (USA).

Services

• Bearer services: pure transport – circuit-mode 64 kbps speech/data, packet-mode, frame-mode.
• Teleservices: complete services incl. terminals – telephony, Group 4 fax, teletex, videotex.
• Supplementary services: add-ons – calling line ID, call waiting, call forwarding, conference, closed user group.

In a circuit switch a dedicated end-to-end path is set up for the whole call, as in the PSTN. In a packet switch data is broken into packets that share links and are stored and forwarded at each node, as in the Internet.

Circuit:  A ==[S1]==[S2]==[S3]== B   (one fixed path)
Packet:   A --[S1]--[S2]-------- B
               \     /
                [S4]                 (packets may differ)

Flow control is the set of procedures that stop a fast sender from overrunning a slow receiver (or an intermediate node) with more data than it can buffer and process. The receiver tells the sender how much it may send, or when to stop.

1. Stop-and-wait flow control

• The sender sends one frame and waits for an acknowledgement (ACK) before sending the next.
• The receiver controls flow simply by delaying the ACK.
• Simple, but poor efficiency on long or fast links. Link utilisation:

$$U = \frac{1}{1 + 2a}, \qquad a = \frac{t_{prop}}{t_{frame}}$$

2. Sliding window flow control

• The sender may send up to W frames before receiving an ACK. Frames carry sequence numbers (modulo $2^k$, window ≤ $2^k - 1$).
• Each ACK slides the window forward; the receiver can also send RNR (Receive Not Ready) to close the window.

 sent+ACKed | sent, not ACKed | can send | cannot
  0 1 2     |  3 4 5          |  6 7     | 8 9 ...
            |<------- window W = 5 --->|

• Utilisation is 1 if $W \ge 2a + 1$, else $U = \dfrac{W}{1+2a}$. Used in HDLC, X.25 and TCP.
• Combined with error control: Go-back-N ARQ and Selective-repeat ARQ.

3. Credit-based (window) end-to-end control

The receiver advertises a credit or window size (e.g. TCP receive window). The sender may send only that many bytes; a window of 0 stops it.

4. Rate-based control

The source is limited to an agreed rate. Leaky bucket gives a constant output rate; token bucket allows limited bursts. Used in ATM and QoS networks.

5. Start–stop (on/off) control

• XON/XOFF characters in software (DC1/DC3) on serial links.
• RTS/CTS or DTR/DSR hardware lines in RS-232 modems.
• Choke packets / back-pressure: a congested node tells the source or previous node to slow down.

Levels of flow control

Good flow control avoids buffer overflow, packet loss and retransmissions, and keeps throughput high.

ISDN (Integrated Services Digital Network) is an all-digital network defined by ITU-T that carries voice, data, fax and video over one subscriber line using standard interfaces. Its architecture has the user–network access, the ISDN exchange, and separate networks for circuit, packet and signalling traffic.

 Customer premises            |    Network
TE1 -S-+                      |
       [NT2]-T-[NT1]-U--------+--[LT/ET  ISDN exchange]
TE2-R-[TA]-S-+                |     |   |    |
                              |  circuit packet SS7
                              |  network network (CCS)

Functional groups

• TE1: ISDN-ready terminal (digital phone, ISDN PC card).
• TE2: non-ISDN terminal (analogue phone, RS-232 PC); needs a TA (terminal adapter).
• NT2: intelligent customer equipment – PBX, LAN controller; does switching and concentration.
• NT1: physical termination of the 2-wire loop; converts 2-wire U to 4-wire S/T, line coding, timing, power.
• LT/ET: line and exchange termination at the ISDN exchange.

Reference points: R (TE2–TA), S (TE–NT2), T (NT2–NT1), U (NT1–LT).

Access interfaces

• BRI: 2B + D = 2×64 + 16 = 144 kbps; for homes and small offices.
• PRI: 30B + D = 2.048 Mbps (E1) or 23B + D = 1.544 Mbps (T1); for PBXs and businesses.

Network side: the exchange routes B channels into the circuit-switched network or a packet-switched network, and carries D-channel signalling (Q.931 / LAPD) into common channel signalling (SS7) between exchanges. Signalling is thus out-of-band, giving fast call set-up and supplementary services.

A PC-to-phone call is a VoIP (Voice over IP) call in which one end is a packet network (the Internet) and the other end is the circuit-switched PSTN. The telecom operator joins the two with a VoIP gateway controlled by a gatekeeper/softswitch.

System arrangement

PC + softphone (mic, speaker)
   | voice in RTP/UDP/IP; SIP or H.323 signalling
   v
Internet (ISP routers)
   v
+------------------- Operator (e.g. NT) -----------+
| Gatekeeper/Softswitch: login, number translation,|
|   routing, billing (prepaid account / PIN)       |
|              | H.248 / MGCP control              |
| VoIP Gateway: packets <-> 64 kbps PCM (E1 TDM)   |
+--------------+-----------------------------------+
               | E1 trunk + SS7 ISUP signalling
               v
         PSTN local/tandem exchange --> called phone

Functions of each part

• PC softphone: samples the voice, encodes it with a codec (G.711 64 kbps, G.729 8 kbps, G.723.1 6.3 kbps), places 20–30 ms of speech in each RTP packet, adds UDP and IP headers. It also does silence suppression and has a jitter buffer for received packets.
• Internet: routes the packets like any other IP traffic (best-effort, no reserved circuit), so delay, jitter and loss vary.
• Gatekeeper / softswitch: registers and authenticates the user (account and password or PIN), checks balance, translates the dialled E.164 telephone number into the correct gateway address, controls the gateway and records call details for billing.
• VoIP media gateway: converts packets to PCM samples and places them in an E1 time slot towards the PSTN, and the reverse for the other direction. It does codec transcoding (e.g. G.729 to G.711), echo cancellation, jitter buffering and packet-loss concealment.
• Signalling gateway: translates SIP/H.323 call signalling to SS7 ISUP (IAM, ACM, ANM, REL) used by the PSTN.

Call procedure

1. User logs in to the operator's VoIP service from the PC application.
2. User dials the phone number; a SIP INVITE / H.225 Setup goes to the softswitch.
3. Softswitch checks the account, chooses a gateway and seizes a free E1 channel.
4. Gateway sends ISUP IAM to the exchange; the exchange rings the phone and returns ACM; ringback tone is played to the PC.
5. When the called party answers (ANM), the speech path opens: PC ↔ Internet ↔ gateway ↔ PSTN ↔ phone.
6. On hang-up from either side, BYE/REL messages free the channel; the softswitch charges the account by duration.

Quality issues and remedies

In this way the call is carried as packets over the Internet up to the operator's gateway, and as a normal 64 kbps circuit over the PSTN from the gateway to the telephone.

ISDN offers standard user–network interfaces, defined by the channel structure (BRI and PRI) and by the reference points between functional groups at the customer premises.

1. Basic Rate Interface (BRI) – 2B + D

• Two B channels of 64 kbps for voice/data and one D channel of 16 kbps for signalling and low-speed packet data.
• User rate $2 \times 64 + 16 = 144$ kbps; with framing and synchronisation the S/T line rate is 192 kbps. On the U interface 160 kbps with 2B1Q coding.
• Works over the existing 2-wire telephone loop. Used by homes and small offices.

2. Primary Rate Interface (PRI)

• 30B + D (Europe, Nepal): 30 × 64 + 64 (D) + 64 (sync) = 2.048 Mbps on an E1.
• 23B + D (USA, Japan): 23 × 64 + 64 + 8 framing = 1.544 Mbps on a T1.
• D channel here is 64 kbps. Used to connect PBXs and LANs. B channels may be grouped into H channels: H0 = 384 kbps, H11 = 1536 kbps, H12 = 1920 kbps.

3. Reference points (physical interfaces)

 TE1 ------------- S --+
                       [NT2]-- T --[NT1]-- U --[LT/ET]
 TE2 -- R --[TA]-- S --+   (PBX)    (NTE)    exchange

The NT1 converts the 2-wire U interface to the 4-wire S/T bus. The D channel at these interfaces uses LAPD (Q.921) and Q.931 call control messages.

Significance of routing

Routing decides the path that packets take from source to destination through the network of switching nodes.

• Delivers packets to the correct destination across many possible paths.
• Chooses the best path (least delay, cost or hops), so it improves throughput and reduces delay.
• Balances load over links and avoids congested parts.
• Finds alternative paths when links or nodes fail, so the network is reliable.
• Uses network resources efficiently, reducing cost.

Significance of flow control

Flow control regulates how fast a sender transmits so that the receiver or an intermediate node is not overrun.

• Prevents buffer overflow and the resulting packet loss and retransmissions.
• Matches devices of different speeds (e.g. fast server, slow terminal).
• Helps avoid congestion and deadlock in the network.
• Keeps throughput high and delay stable; ensures fair sharing.
• Methods: stop-and-wait, sliding window (HDLC, TCP), XON/XOFF, RTS/CTS, credit and rate (leaky bucket) control.

Dynamic routing

Dynamic (adaptive) routing changes routes automatically to follow the current topology and traffic. Routers exchange routing information and recompute their tables periodically or when a change is detected.

Types

• Centralized: a Routing Control Centre gathers data from all nodes, computes routes and sends tables.
• Isolated: each node uses only local information, e.g. hot potato routing (send on the shortest queue).
• Distributed:
  • Distance vector: each router sends its distance table to neighbours; updates by Bellman-Ford, $D_x(y) = \min_v \{c(x,v) + D_v(y)\}$. Simple, but slow convergence (count-to-infinity). Example: RIP.
  • Link state: each router floods its link costs to all; every router builds the whole map and runs Dijkstra. Fast convergence. Example: OSPF.

 A ---2--- B ---1--- D
  \               /
   5             3
    \--- C -----/
 Normal: A->D via B (cost 3)
 B-D fails: A->D via C (cost 8), learnt automatically

Advantages: adapts to failures and congestion, no manual work, good for large networks. Disadvantages: control traffic overhead, more processing, possible routing loops or oscillation during convergence.

ISDN (Integrated Services Digital Network) gives digital end-to-end connections over the subscriber line for many services (voice, data, fax, video) through one standard interface. The connections and services ISDN offers can be described by (1) the types of connection it can set up, (2) the access channels and interfaces used, and (3) the service classes (bearer, tele- and supplementary services) defined in ITU-T I.200 series.

Architecture behind the connections

 Customer premises                   Network
TE1 --S--+
         [NT2]--T--[NT1]--U--+--[ISDN exchange]
TE2-R-[TA]-S--+              |   |     |     |
                             | circuit packet CCS/SS7
                             | switched switched signalling

• TE1 ISDN terminal, TE2 non-ISDN terminal with TA, NT2 PBX/LAN, NT1 line termination, LT/ET at the exchange.
• Reference points R, S, T, U separate these groups.
• The exchange sends user B-channel traffic to the circuit-switched or packet-switched network and the D-channel signalling to the common channel signalling (SS7) network.

1. Types of ISDN connection

1. Circuit-switched connection: a B channel (64 kbps) is set up on demand using D-channel signalling (Q.931 SETUP, ALERTING, CONNECT, RELEASE). Used for telephony, fax, video calls, file transfer.
2. Packet-switched connection: user data is carried as X.25 packets either on a B channel to a packet handler or directly on the D channel (low-speed data, e.g. POS terminals, alarm systems).
3. Semi-permanent (leased) connection: a B channel is set up by the operator for a long time without call signalling; acts like a 64 kbps leased line.
4. Frame-mode connection: frame relay service with LAPF framing over B or H channels.

2. Channels and access interfaces

• BRI = 2B + D = 144 kbps user rate (192 kbps line rate at S/T): homes, small offices.
• PRI = 30B + D = 2.048 Mbps (E1) or 23B + D = 1.544 Mbps (T1): PBXs, businesses.

3. ISDN services

(a) Bearer services – only transport the information across the network between interfaces (OSI layers 1–3); the user chooses the terminal.

• Circuit-mode: 64 kbps unrestricted digital, 64 kbps speech, 3.1 kHz audio (for modems and fax), 2×64 kbps, 384/1536/1920 kbps.
• Packet-mode: virtual call and permanent virtual circuit (X.25), connectionless.
• Frame-mode: frame relay and frame switching.

(b) Teleservices – complete services including terminal functions (layers 1–7):

• Telephony (3.1 kHz and 7 kHz wideband speech)
• Group 4 facsimile
• Teletex and telex
• Videotex, videotelephony, video conferencing

(c) Supplementary services – added to a bearer or teleservice, not sold alone:

• Calling line identification presentation/restriction (CLIP/CLIR)
• Call waiting, call hold, call forwarding (busy / no reply / unconditional)
• Three-party and conference calling
• Closed user group, direct dialling-in, advice of charge, user-to-user signalling

Example

A small office has a BRI line. One B channel carries a phone call (circuit-mode telephony teleservice) while the other carries a 64 kbps Internet/file-transfer connection; the D channel handles signalling for both and also shows the caller's number (CLIP supplementary service). A PC with only an RS-232 port connects through a TA at the R point.

Advantages of ISDN connections

• Fully digital, so clear speech and error-free data.
• Fast call set-up (1–2 s) because of out-of-band D-channel signalling.
• Several services and calls on one line at the same time.
• Standard interfaces let terminals from different makers work together.

Flow control in a data communication network is the technique that controls the rate at which a sender sends data so that the receiver (or an intermediate node) is not flooded with more data than its buffers and processor can handle. Without it, frames are lost, must be retransmitted and throughput falls.

Why needed

• Sender and receiver differ in speed and buffer size.
• Receiver must also process data (check errors, pass to higher layers).
• Network nodes have limited buffers; overflow causes congestion.

Main methods

1. Stop-and-wait: send one frame, wait for ACK, then send the next. Simple but inefficient when propagation delay is large ($U = 1/(1+2a)$, $a = t_{prop}/t_{frame}$).
2. Sliding window: sender may have up to W unacknowledged frames; ACKs slide the window. Receiver can send RNR to stop. Efficient; used in HDLC, X.25, TCP.

  [0 1 2][3 4 5 6 7][8 9 ...]
  ACKed   window=5   not yet allowed

3. Start–stop: XON/XOFF characters or RTS/CTS hardware lines.
4. Credit / rate based: receiver grants credits (TCP window); or source is limited to a rate (leaky/token bucket).

Levels: hop-by-hop (between adjacent nodes, data link layer), network access (host to network) and end-to-end (between hosts, transport layer).

ISDN (Integrated Services Digital Network) is an ITU-T standard all-digital network that provides voice and non-voice services over the subscriber line through a small set of standard user–network interfaces.

Basic services in ISDN

1. Bearer services – carry information between user–network interfaces without changing it (OSI layers 1–3).

• Circuit-mode: 64 kbps unrestricted data, 64 kbps speech, 3.1 kHz audio, multiples of 64 kbps.
• Packet-mode: X.25 virtual calls and permanent virtual circuits on B or D channel.
• Frame-mode: frame relay.

2. Teleservices – complete services including the terminal (layers 1–7).

• Telephony, Group 4 fax, teletex, videotex, videotelephony.

3. Supplementary services – extra features added to the above.

• Caller ID (CLIP), call waiting, call forwarding, call hold, conference calls, closed user group, advice of charge.

Channel services (access)

B = 64 kbps user channel; D = signalling and low-speed packets; H channels (384, 1536, 1920 kbps) for high-speed data.

ISDN architecture

 Customer premises                   Network
TE1 --S--+
         [NT2]--T--[NT1]--U--+--[ISDN exchange]
TE2-R-[TA]-S--+              |   |     |     |
                             | circuit packet CCS/SS7
                             | switched switched signalling

Functional groups

• TE1: ISDN terminal (digital phone, ISDN fax).
• TE2: non-ISDN terminal (analogue phone, PC with RS-232).
• TA: terminal adapter; converts TE2 signals to ISDN.
• NT2: intelligent premises equipment (PBX, LAN controller) – switching, concentration, layer 2–3 functions.
• NT1: physical termination of the line – 2-wire to 4-wire conversion, line coding, timing, power feeding.
• LT/ET: line and exchange termination at the central office.

Reference points

• R: TE2–TA (non-ISDN).
• S: TE1/TA–NT2, 4-wire 192 kbps bus (up to 8 terminals).
• T: NT2–NT1.
• U: NT1–exchange over the 2-wire loop (2B1Q, 160 kbps).

Network side: the ISDN exchange connects to a circuit-switched network, a packet-switched network and a common-channel signalling (SS7) network. User signalling on the D channel (Q.931) is converted to SS7 ISUP between exchanges, giving fast call set-up and the supplementary services.

Ethernet

Ethernet is the most widely used LAN technology, standardised as IEEE 802.3. It was developed at Xerox PARC (Metcalfe, 1973) and later by DEC–Intel–Xerox. All stations share one channel (a bus, or a hub/switch in star wiring) and access it with CSMA/CD (Carrier Sense Multiple Access with Collision Detection). Classic Ethernet runs at 10 Mbps with Manchester coding; Fast Ethernet (100 Mbps), Gigabit and 10 Gigabit Ethernet followed.

Ethernet (802.3) frame

+--------+---+----+----+-----+---------+-----+
|Preamble|SFD| DA | SA | Len | Data+Pad| FCS |
|   7    | 1 | 6  | 6  |  2  | 46-1500 |  4  |
+--------+---+----+----+-----+---------+-----+
            (bytes)

• Preamble 10101010… for clock sync; SFD 10101011 marks frame start.
• 48-bit destination and source MAC addresses.
• Data 46–1500 bytes (padded), so frame size is 64–1518 bytes.
• FCS: 32-bit CRC for error detection.

CSMA (Carrier Sense Multiple Access)

In CSMA a station listens before talking: it senses the channel and transmits only if the channel is idle. This reduces collisions compared with ALOHA, but collisions still happen because a signal takes time (propagation delay) to reach other stations; two stations may both sense idle and start together.

Persistence methods when the channel is busy:

• 1-persistent: keep sensing; send as soon as it becomes idle (probability 1). Used by Ethernet. Many collisions if several are waiting.
• Non-persistent: if busy, wait a random time and sense again. Fewer collisions, more delay.
• p-persistent: (slotted channels) when idle, send with probability p, otherwise defer to the next slot.

CSMA/CD – working principle of Ethernet

CSMA/CD adds collision detection: a station keeps listening while transmitting, and if it detects a collision it stops at once instead of sending the whole damaged frame.

1. Station with a frame senses the channel (carrier sense).
2. If busy, wait until idle (1-persistent), then wait the inter-frame gap (9.6 µs).
3. Transmit the frame while monitoring the line.
4. If no collision during the whole frame: success.
5. If a collision is detected (signal on the line differs from sent signal / voltage too high):
   • stop sending the data and send a 32-bit jam signal so all stations notice the collision;
   • increment attempt counter $n$; if $n > 16$, give up and report an error;
   • otherwise wait a random binary exponential backoff time, then go to step 1.

Binary exponential backoff: after the $n$-th collision choose a random integer

$$K \in \{0, 1, \dots, 2^{m} - 1\}, \quad m = \min(n, 10)$$

and wait $K \times$ slot time. Slot time = 512 bit times = 51.2 µs at 10 Mbps. The waiting range doubles after each collision, so the load adapts.

     Frame ready
         |
   Sense channel <------------------+
         | idle                     |
   Transmit + listen                |
     |           |                  |
 no collision  collision            |
     |           |                  |
  Success    send jam, n = n+1      |
             n > 16 ? --yes--> abort|
                 | no               |
         wait K x 51.2 us ----------+

Minimum frame size: a sender must still be transmitting when news of a collision returns from the far end, so frame time ≥ round-trip time ($2\tau$). For 10 Mbps and 2500 m (with 4 repeaters) the slot is 512 bits, hence the 64-byte minimum frame.

Example: stations A and C both find the bus idle and start at almost the same time. Their signals collide; both detect it, send jam and stop. After the first collision each picks K from {0, 1}; if A picks 0 and C picks 1, A sends at once and C, on sensing a busy channel, waits until A finishes.

Features: simple, cheap, no central control, works well at light and moderate load; at heavy load collisions reduce efficiency. Modern switched full-duplex Ethernet has no shared medium, so CSMA/CD is not needed there.

An internet (internetwork) is a set of separate networks (LANs, MANs, WANs) joined together by interconnecting devices so that any host can communicate with any other host. The global Internet is the largest example: home and office LANs connect to ISPs, ISPs connect to regional and national backbones, and these connect to international backbones.

Structure of an internet

     LAN-1 segment A        LAN-1 segment B
  H--H--H--[Repeater]--H--H--H
                |
            [Bridge]------ LAN-2 (H--H--H)
                |
            [Router R1]
                |  WAN link (leased line)
        +-------+--------+
        |   ISP / WAN    |
        |  [R2]----[R3]  |  backbone routers
        +---+--------+---+
            |        |
        [Router]  [Gateway]
            |        |
          LAN-3   Different network
        (H--H--H) (e.g. mainframe SNA,
                   X.25, email system)

Hosts (H) on LANs reach other LANs through bridges; LANs reach distant networks through routers and WAN links; networks with different protocol suites are joined by gateways.

Devices and the OSI layers they work at

 OSI layer      Device
 7 Application  +
 6 Presentation | Gateway (all layers)
 5 Session      |
 4 Transport    +
 3 Network      Router
 2 Data link    Bridge (and switch)
 1 Physical     Repeater (and hub)

1. Repeater

• Works at the physical layer.
• A signal weakens (attenuation) and is distorted as it travels. A repeater receives the weak signal, regenerates the original bit pattern and sends it on, so the LAN can be made longer.
• It connects two segments of the same LAN; it does not read addresses or filter traffic, so all frames and collisions pass through. Both segments are one collision domain.
• Example: Ethernet 10Base5 segments of 500 m are joined by up to 4 repeaters for 2500 m total. A hub is a multiport repeater.
• Note: a repeater regenerates, it does not merely amplify (an amplifier also boosts noise).

2. Bridge

• Works at the physical and data link layers.
• Connects two LANs or LAN segments (same or different MAC types, e.g. Ethernet to Token Ring with translation).
• Reads the MAC address in each frame and uses a forwarding table:
  • destination on the same side as source → filter (discard);
  • destination on the other side → forward;
  • unknown destination → flood.
• Transparent (learning) bridges learn addresses from the source field of frames; the spanning tree algorithm removes loops when several bridges exist.
• Divides the network into separate collision domains, so it reduces traffic and improves security. A switch is a multiport bridge.

3. Router

• Works at the network layer (layers 1–3).
• Connects different networks (LAN to LAN, LAN to WAN) that may use different technologies, and forwards packets using logical (IP) addresses.
• Keeps a routing table built by static entries or routing protocols (RIP – distance vector; OSPF – link state; BGP between ISPs) and chooses the best path by metrics such as hops, delay or cost.
• Does not forward broadcasts, so it separates broadcast domains; it can also fragment packets, do filtering (firewall), NAT and QoS.
• Example: the router at a campus connects the campus LAN to the ISP over fibre.

4. Gateway

• Works at all seven layers (mainly transport to application).
• Connects networks that use completely different protocol architectures, and performs protocol conversion: changing message format, addressing, codes and procedures.
• Examples: email gateway between SMTP and X.400; gateway between TCP/IP and IBM SNA; VoIP gateway between IP telephony and the PSTN; in common usage the "default gateway" is the router leading out of a LAN.
• Usually software on a dedicated computer; slower and more complex than routers.

Comparison

Together these devices let small LANs grow into campus networks and connect, through routers and gateways, to the global Internet.

Communication protocol

A communication protocol is a set of rules and conventions that govern the exchange of data between two communicating entities. Both ends must follow the same protocol to understand each other. A protocol defines three key elements:

• Syntax: format of data and control information (field order and sizes, e.g. a header of 20 bytes).
• Semantics: meaning of each field and the action to take (e.g. what an ACK or error code means).
• Timing: when and how fast data is sent, sequencing and speed matching.

Examples: HDLC (data link), IP (network), TCP (transport), HTTP and SMTP (application), Q.931 in ISDN.

OSI reference model

The Open Systems Interconnection (OSI) model was developed by ISO (ISO 7498, 1984) so that systems from different vendors can communicate. It divides the communication task into seven layers. Each layer performs a defined set of functions, uses the services of the layer below and gives services to the layer above. Layers on two machines talk to each other by peer protocols, while actual data moves down one stack, across the medium and up the other.

Layering principles: a layer for each different level of abstraction, well-defined functions, small interface flow across boundaries, and enough layers to separate functions but not too many.

 Host A                                  Host B
+----------------+                  +----------------+
| 7 Application  |<---------------->| 7 Application  |
| 6 Presentation |<---------------->| 6 Presentation |
| 5 Session      |<---------------->| 5 Session      |
| 4 Transport    |<---------------->| 4 Transport    |
| 3 Network      |<->[3]<---->[3]<->| 3 Network      |
| 2 Data link    |<->[2]<---->[2]<->| 2 Data link    |
| 1 Physical     |<->[1]<---->[1]<->| 1 Physical     |
+-------+--------+  router  router  +-------+--------+
        +========= physical medium =========+

Layers 1–3 are network-dependent (also run in intermediate nodes); layers 4–7 are end-to-end (only in hosts).

1. Physical layer

• Sends the raw bit stream over the medium.
• Defines mechanical (connectors, pins), electrical (voltage levels, bit duration), functional (meaning of pins) and procedural specifications.
• Line coding, bit synchronisation, data rate, simplex/half/full duplex, topology.
• Examples: RS-232/V.24, X.21, 10Base-T, ISDN I.430. Data unit: bit.

2. Data link layer

• Turns the raw line into a reliable link between adjacent nodes.
• Framing (start/end of frame, bit stuffing), physical (MAC) addressing.
• Error control with CRC and ARQ (retransmission), flow control (sliding window).
• In LANs it is split into LLC and MAC (medium access, e.g. CSMA/CD) sublayers.
• Examples: HDLC, LAPB, LAPD, PPP, IEEE 802.3. Data unit: frame.

3. Network layer

• Delivers packets from source host to destination host across many networks.
• Logical addressing (IP address), routing (choosing paths), switching, congestion control, fragmentation, internetworking.
• Connection-oriented (virtual circuit) or connectionless (datagram) service; accounting.
• Examples: IP, X.25 packet layer, ICMP. Data unit: packet.

4. Transport layer

• End-to-end (process-to-process) reliable delivery of the complete message.
• Segmentation and reassembly, port addressing, end-to-end error and flow control, connection management, multiplexing several sessions on one network connection.
• Hides network details from upper layers; gives the quality of service required.
• Examples: TCP, UDP, ISO TP0–TP4. Data unit: segment.

5. Session layer

• Establishes, manages and terminates sessions (dialogues) between applications.
• Dialogue control: who talks when (half/full duplex, token management).
• Synchronisation: inserts checkpoints so that after a failure transfer restarts from the last checkpoint, not from the start.
• Examples: RPC, NetBIOS, ISO 8327.

6. Presentation layer

• Concerned with syntax and semantics of the information exchanged.
• Translation between different codes (ASCII, EBCDIC) and data formats (ASN.1).
• Encryption/decryption for security and compression to reduce bits.
• Examples: JPEG, MPEG, SSL/TLS encryption, ASN.1.

7. Application layer

• Gives the user and application programs access to network services.
• Services: file transfer and access (FTAM, FTP), e-mail (X.400, SMTP), directory (X.500), virtual terminal, web (HTTP), network management.

Data flow and encapsulation

Each layer adds its own header (and the data link layer also a trailer) to the data from the layer above; the receiver removes them in reverse order.

          Data                       Layer 7
        AH|Data                      7  (headers
      PH|AH|Data                     6   added at
    SH|PH|AH|Data                    5   each layer)
  TH|SH|PH|AH|Data       = segment   4
NH|TH|SH|PH|AH|Data      = packet    3
DH|NH|....|Data|DT       = frame     2
0110100101110...         = bits      1

Summary table

Advantages of the layered OSI model

• Divides a complex task into small, manageable parts.
• Changes in one layer do not affect others (modularity).
• Equipment from different vendors can work together (open standards).
• Clear framework for teaching, design and troubleshooting.

The OSI model is mainly a reference model; practical networks mostly use the TCP/IP suite, whose layers map onto OSI (application ≈ 5–7, transport = 4, internet = 3, network access = 1–2).

IBM Token Ring is a LAN developed by IBM in the 1980s and standardised as IEEE 802.5. Stations form a logical ring and get the right to transmit by capturing a circulating token, so there are no collisions. It runs at 4 Mbps or 16 Mbps with differential Manchester coding.

Physical structure – star-wired ring

          +---------------+
   A -----|               |----- B
          |   MAU (hub)   |
   D -----|  ring inside  |----- C
          +---------------+
 Logical ring: A -> B -> C -> D -> A
 Each lobe = 2 pairs (send + receive)

• Stations connect to a MAU (Multistation Access Unit), a wiring concentrator, by shielded twisted pair (STP, IBM Type 1) or UTP.
• Inside the MAU the lobes are joined into a ring. A relay in the MAU bypasses a station that is switched off or faulty, so the ring is not broken. MAUs are joined by Ring-In / Ring-Out ports to make larger rings.
• Up to 260 stations on STP (72 on UTP).

Frame and token format

Token (3 bytes):
+----+----+----+
| SD | AC | ED |
+----+----+----+
Data frame:
+--+--+--+--+--+------+---+--+--+
|SD|AC|FC|DA|SA| Data |FCS|ED|FS|
|1 |1 |1 |6 |6 |  0-n | 4 |1 |1 |
+--+--+--+--+--+------+---+--+--+
AC byte = P P P T M R R R

• SD / ED: start and end delimiters (use code violations J, K of differential Manchester).
• AC (access control): PPP = priority, T = token bit (0 token, 1 frame), M = monitor bit, RRR = reservation.
• FC: frame control (data or MAC control frame).
• DA, SA: 6-byte addresses; FCS: 32-bit CRC.
• FS (frame status): A (address recognised) and C (frame copied) bits set by the receiver.

Working (token passing)

1. A free token circulates round the ring; each station repeats it bit by bit (1-bit delay).
2. A station with data waits for the token, changes the T bit to 1 (token becomes a start-of-frame), and sends its frame. It may hold the token for at most the token holding time (10 ms).
3. Each station repeats the frame. The destination copies it and sets the A and C bits in FS.
4. The frame returns to the sender, which removes it from the ring (source stripping), checks A and C for delivery, and releases a new free token.
5. At 16 Mbps, early token release lets the sender release the token right after its frame, so two frames may be on the ring at once.

Priority: a station with a high-priority frame writes its priority in the RRR bits of a passing frame; the next token is issued at that priority, so only stations with equal or higher priority may use it.

Ring maintenance – active monitor

One station is elected active monitor; others are standby monitors.

• Detects a lost token (no token within a timeout) and issues a new one.
• Removes orphan/circulating frames: it sets the M bit on each frame; a frame seen again with M = 1 is purged.
• Supplies the master clock and a latency buffer so the ring is long enough to hold a 24-bit token.
• Sends Active Monitor Present frames; beacon frames locate cable breaks.

Features

Advantages: deterministic access (bounded waiting time), good performance under heavy load, priority support, fault bypass in MAU. Disadvantages: more complex and costly than Ethernet, monitor and token management needed, now replaced by switched Ethernet.

The OSI (Open Systems Interconnection) reference model, published by ISO (ISO 7498, 1984), is a seven-layer framework that describes how data is exchanged between open systems from different makers. Each layer does a defined job, uses the services of the layer below and serves the layer above. Peer layers on two machines communicate using protocols; actual data passes down the sender's stack, over the medium and up the receiver's stack.

 Host A                       Host B
+----------------+    +----------------+
| 7 Application  |<-->| 7 Application  |
| 6 Presentation |<-->| 6 Presentation |
| 5 Session      |<-->| 5 Session      |
| 4 Transport    |<-->| 4 Transport    |
| 3 Network      |<-->| 3 Network      |  (also in
| 2 Data link    |<-->| 2 Data link    |  routers)
| 1 Physical     |<-->| 1 Physical     |
+-------+--------+    +-------+--------+
        +====== medium =======+

Functions of the layers

1. Physical: transmits raw bits; defines connectors, voltages, bit timing, line coding, data rate (RS-232, V.24, X.21). Unit: bit.
2. Data link: framing, MAC addressing, error detection (CRC) and correction by retransmission, flow control between adjacent nodes (HDLC, LAPB, PPP, IEEE 802; LLC and MAC sublayers). Unit: frame.
3. Network: routes packets from source to destination host across networks; logical addressing, routing, congestion control, internetworking (IP, X.25). Unit: packet.
4. Transport: reliable end-to-end delivery of whole messages; segmentation/reassembly, port addressing, end-to-end error and flow control, multiplexing (TCP, UDP, TP0–TP4). Unit: segment.
5. Session: sets up, manages and ends dialogues; dialogue control (who talks when) and synchronisation checkpoints for recovery.
6. Presentation: data representation – code conversion (ASCII/EBCDIC), encryption, compression (ASN.1, JPEG, TLS).
7. Application: network services for users – file transfer (FTAM, FTP), e-mail (X.400, SMTP), directory (X.500), virtual terminal, web.

Layers 1–3 are network-oriented (also run in routers/switching nodes); layers 4–7 are end-to-end, running only in hosts.

Encapsulation

Each layer adds a header (data link also adds a trailer) to the data it receives, and the receiver strips them in reverse order:

L7        AH|Data
L4     TH|...|Data          segment
L3  NH|TH|...|Data          packet
L2 DH|NH|TH|...|Data|DT     frame
L1 0101101001...            bits

Merits

• Breaks a complex job into small, independent layers; change in one layer does not affect others.
• Open, vendor-independent standards allow interoperability.
• Good tool for teaching, design and troubleshooting.

The practical Internet uses TCP/IP, whose four layers map onto OSI (application ≈ 5–7, transport = 4, internet = 3, network access = 1–2).

The Cambridge Ring is a slotted ring LAN developed at the University of Cambridge (UK) in the late 1970s. A fixed number of small, fixed-size mini-packets (slots) circulate continuously round the ring; a station sends data by filling an empty slot. The raw ring speed is 10 Mbps.

Structure

        [Monitor]
       /         \
   [R]-S1       S4-[R]
    |             |
   [R]-S2 ---- S3-[R]
 R = repeater, S = station unit + access box

• Repeaters (about 100 m apart) regenerate the signal; they are powered from the ring, so a station can be switched off without breaking the ring. Two twisted pairs are used.
• Station unit / access box: connects a host to its repeater and handles slot filling and response checking.
• Monitor station: sets up the slot structure at start-up, marks and clears lost or damaged slots, and checks parity; there is also an error-logging station.

Mini-packet (slot) format – 38 bits

+-+-+-+-------+-------+----------+--+-+
|S|F|M| Dest  |Source |   Data   |R |P|
|1|1|1|   8   |   8   |    16    |2 |1|
+-+-+-+-------+-------+----------+--+-+

• S: start bit (always 1); F: full/empty bit.
• M: monitor bit, used to remove orphan slots.
• 8-bit destination and source addresses (up to 255 stations).
• Data: 2 bytes per slot.
• R: 2 response bits; P: parity bit.

Working

1. A station that wants to send waits for a slot with F = empty.
2. It sets F = full, writes the addresses and 2 bytes of data.
3. The destination copies the data and sets the response bits to show: accepted, busy, rejected, or ignored (station absent).
4. The slot goes all round the ring back to the source, which reads the response and marks the slot empty.
5. The source may not reuse the same slot at once; it passes it on, so every station gets a fair share.
6. The monitor sets M as a full slot passes it; if a full slot returns with M already set (source failed to empty it), the monitor empties it.

Efficiency

Only 16 of 38 bits are user data:

$$\frac{16}{38} = 0.421, \qquad 0.421 \times 10\ \text{Mbps} = 4.21\ \text{Mbps}$$

so the useful capacity is at most about 4.2 Mbps; inter-slot gaps and the slot-return rule lower one station's rate further (around 1 Mbps).

Advantages and disadvantages

• Simple, fixed-size slots; fair and bounded access; no collisions.
• Immediate acknowledgement through response bits; powered repeaters improve reliability.
• Low data efficiency (large overhead per 2 bytes), small slots unsuitable for big transfers, single ring break stops the network.

The Cambridge Ring was widely used in UK universities and research labs and influenced later slotted-ring and ATM-style cell designs.

Network topologies

The topology of a network is the geometric arrangement of the nodes and the links that join them. It decides cost, reliability and the access method.

 Bus:   H---H---H---H        Star:     H   H
        ===============               \ /
                                  H--[Hub]--H
 Ring:  H---H                         / \
        |   |                        H   H
        H---H
 Mesh:  every node linked to every other
        (n(n-1)/2 links)

IBM Ring LAN

IBM Token Ring (IEEE 802.5) is a LAN in which stations form a logical ring and gain the right to send by capturing a circulating token; there are no collisions. Rates are 4 or 16 Mbps, coding is differential Manchester. Physically it is a star-wired ring – a hybrid topology:

          +---------------+
   A -----|               |----- B
          |   MAU (hub)   |
   D -----|  ring inside  |----- C
          +---------------+
 Logical ring: A -> B -> C -> D -> A
 Each lobe = 2 pairs (send + receive)

• Stations connect to an MAU by STP/UTP lobes; relays in the MAU bypass a dead station.
• Up to 260 stations per ring (STP).

Frame and token format

Token (3 bytes):
+----+----+----+
| SD | AC | ED |
+----+----+----+
Data frame:
+--+--+--+--+--+------+---+--+--+
|SD|AC|FC|DA|SA| Data |FCS|ED|FS|
|1 |1 |1 |6 |6 |  0-n | 4 |1 |1 |
+--+--+--+--+--+------+---+--+--+
AC byte = P P P T M R R R

• SD / ED: start and end delimiters (use code violations J, K of differential Manchester).
• AC (access control): PPP = priority, T = token bit (0 token, 1 frame), M = monitor bit, RRR = reservation.
• FC: frame control (data or MAC control frame).
• DA, SA: 6-byte addresses; FCS: 32-bit CRC.
• FS (frame status): A (address recognised) and C (frame copied) bits set by the receiver.

Working (token passing)

1. A free token circulates round the ring; each station repeats it bit by bit (1-bit delay).
2. A station with data waits for the token, changes the T bit to 1 (token becomes a start-of-frame), and sends its frame. It may hold the token for at most the token holding time (10 ms).
3. Each station repeats the frame. The destination copies it and sets the A and C bits in FS.
4. The frame returns to the sender, which removes it from the ring (source stripping), checks A and C for delivery, and releases a new free token.
5. At 16 Mbps, early token release lets the sender release the token right after its frame, so two frames may be on the ring at once.

Priority: a station with a high-priority frame writes its priority in the RRR bits of a passing frame; the next token is issued at that priority, so only stations with equal or higher priority may use it.

Ring maintenance – active monitor

One station is elected active monitor; others are standby monitors.

• Detects a lost token (no token within a timeout) and issues a new one.
• Removes orphan/circulating frames: it sets the M bit on each frame; a frame seen again with M = 1 is purged.
• Supplies the master clock and a latency buffer so the ring is long enough to hold a 24-bit token.
• Sends Active Monitor Present frames; beacon frames locate cable breaks.

Advantages: deterministic, fair access, good at high load, priorities. Disadvantages: costly and complex compared with Ethernet.

Ethernet is the most widely used LAN technology, standardised as IEEE 802.3. It was developed at Xerox PARC (Robert Metcalfe, 1973) and later specified by DEC, Intel and Xerox (DIX, 10 Mbps). All stations share a common channel and access it using CSMA/CD (Carrier Sense Multiple Access with Collision Detection). It uses Manchester coding at 10 Mbps.

Physical types

In 10BaseX, "10" = 10 Mbps, "Base" = baseband, the last part = cable type or length (×100 m).

Bus (10Base5/10Base2):
 T|==+=====+=====+=====+==|T   T = terminator
     H     H     H     H
Star (10Base-T):  H--[Hub/Switch]--H
                        |
                        H

Frame format

+--------+---+----+----+-----+---------+-----+
|Preamble|SFD| DA | SA | Len | Data+Pad| FCS |
|   7    | 1 | 6  | 6  |  2  | 46-1500 |  4  |
+--------+---+----+----+-----+---------+-----+

• Preamble (10101010…) for clock sync; SFD (10101011) marks start.
• 48-bit MAC destination and source addresses.
• Length/type field; data 46–1500 bytes (padded to 46).
• FCS: 32-bit CRC. Frame size 64–1518 bytes.

CSMA/CD operation

1. A station senses the channel; if busy it waits until idle (1-persistent).
2. It transmits and listens at the same time.
3. If a collision is detected, it stops, sends a 32-bit jam signal, and waits a random binary exponential backoff time: after the $n$-th collision it picks $K$ from $0$ to $2^{\min(n,10)} - 1$ and waits $K \times 51.2\ \mu\text{s}$ (slot time = 512 bit times at 10 Mbps).
4. After 16 failed attempts the frame is dropped.

The 64-byte minimum frame makes sure a sender is still sending when a collision from the far end (up to 2500 m with 4 repeaters) reaches it, i.e. frame time ≥ round-trip time.

Advantages and limitations

• Simple, cheap, easy to install and expand; no central control.
• Good performance at light to moderate load.
• At heavy load collisions waste capacity; access delay is not deterministic (unlike token ring); no priority.

Modern Ethernet uses switches and full-duplex links, so each link is collision-free and CSMA/CD is no longer needed; speeds have grown from 10 Mbps to 100 Gbps and beyond.