Network Technologies¶

5-mark answer and 10-mark extension

For 5 marks - ISDN rates and operation¶

Integrated Services Digital Network (ISDN) provides end-to-end digital access for voice, circuit/packet data, text, and images through standardized user-network interfaces. A B channel carries a \(64\,\text{kbit/s}\) bearer; the D channel carries signaling and may carry low-rate packet data. Basic Rate Interface (BRI) is \(2B+D_{16}\), giving \(2(64)+16=144\,\text{kbit/s}\) of B/D payload for a home or small office. Primary Rate Interface (PRI) on E1 is \(30B+D_{64}\): \(30(64)+64=1.984\,\text{Mbit/s}\) of bearer/signaling within the \(2.048\,\text{Mbit/s}\) E1 line; TS0 supplies framing/overhead.

ISDN uses out-of-band D-channel signaling for call setup. The user requests a call on D, the network authenticates/analyzes the number, reserves a B channel, alerts the called party, connects the bearer on answer, supervises the call, and releases the B channel through D-channel messages.

Add for a 10-mark ISDN/BISDN extension¶

The reference configuration separates terminal and network responsibilities. TE1 is an ISDN terminal; a non-ISDN TE2 reaches ISDN through a terminal adapter (TA). The R point lies between TE2 and TA, S between TE1/TA and customer switching/concentration such as NT2, T between NT2 and the physical network termination NT1, and U is the two-wire access from NT1 toward the carrier line termination. Where NT2 is absent, S and T are commonly combined as S/T. Exact ownership and physical coding at U are national/implementation conventions.

The distinction between payload and gross line rate is essential; BRI physical rates and framing depend on the reference point/line code. H channels aggregate higher rates for wideband services in ISDN specifications.

Broadband ISDN (BISDN) was the broadband extension intended for integrated voice, video and high-speed data, historically selecting ATM cell switching as its transport basis. Narrowband ISDN centers on \(64\,\text{kbit/s}\) B channels and circuit-switched access; BISDN uses high-rate virtual paths/channels and QoS classes for multimedia. ISDN's strengths were digital quality, standard access, rapid signaling and simultaneous services, but limited BRI capacity and complex/tariff-dependent deployment led broadband IP access to supersede it. BISDN/ATM influenced QoS and virtual-circuit design but did not become today's universal residential access.

Practice target: 9 minutes for BRI/PRI or 18 minutes for the full reference diagram, call path, payload/line-rate table, BISDN comparison, and convention caveat.

5-mark answer and 10-mark extension

For 5 marks - ATM cell core¶

Asynchronous Transfer Mode (ATM) is connection-oriented statistical cell switching for integrated voice, video and data. Higher-layer information is adapted and segmented into fixed 53-byte cells, each containing a 5-byte header and 48-byte payload. Fixed length bounds serialization and simplifies high-speed hardware switching, though payload efficiency is only \(48/53\times100\approx90.57\%\) before adaptation overhead.

At an ATM UNI header, the fields are GFC 4 bits, VPI 8 bits, VCI 16 bits, PTI 3 bits, CLP 1 bit, and HEC 8 bits, totaling 40 header bits. VPI identifies a virtual-path bundle; VCI identifies one virtual channel within it. PTI distinguishes user/management and congestion-related payload types, CLP marks lower-priority cells for preferential discard, and HEC checks/corrects limited header errors and aids cell delineation.

Add for a 10-mark ATM switching extension¶

Before transfer, signaling or provisioning establishes a virtual-channel connection through ATM switches. At each input, the switch reads the incoming port plus VPI/VCI, consults a translation table, queues the cell for an output, replaces VPI/VCI with locally significant outgoing labels, recomputes HEC, and transmits it. Many VCs share a VP, allowing a provider to route/manage a bundle by VPI while endpoints distinguish services by VCI. Labels are link-local rather than globally fixed addresses.

The 5-byte network-node interface (NNI) header follows a different convention: it omits GFC and expands VPI from 8 to 12 bits; VCI 16, PTI 3, CLP 1 and HEC 8 remain. Thus field widths must be identified as UNI or NNI, not quoted as universal. ATM Adaptation Layers map variable application units to 48-byte payloads and reassemble them; loss of one cell can damage a larger adapted unit.

Traffic contracts describe peak/sustained rate and burst tolerance; policing, shaping, scheduling and admission control protect QoS. ATM's small cells reduce the time a voice cell waits behind a long data unit, and VP switching, CLP, OAM and hardware lookup support managed backbones. Against these benefits stand \(5/53=9.43\%\) header overhead, segmentation/reassembly complexity, cell-loss sensitivity and expensive control. IP/MPLS and carrier Ethernet largely replaced ATM, although ATM remains historically important and appeared in BISDN and access/backhaul systems.

Practice target: 9 minutes for definition and UNI fields or 18 minutes for both interfaces, translation flow, VP/VC hierarchy, service classes, efficiency, and caveats.

5-mark answer and 10-mark extension

For 5 marks - PDH and SDH core¶

Plesiochronous Digital Hierarchy (PDH) multiplexes tributaries whose clocks are nominally equal but not phase/frequency locked. Multiplexers insert justification/stuff bits to absorb tolerance. European rates are exactly E1 \(2.048\), E2 \(8.448\), E3 \(34.368\), and E4 \(139.264\,\text{Mbit/s}\). Because tributary locations are not directly visible at the highest level, extracting an E1 from E4 normally requires E4 -> E3 -> E2 -> E1 demultiplexing.

Synchronous Digital Hierarchy (SDH) derives network timing from synchronized references and maps tributaries into a structured synchronous frame with pointers and extensive overhead. The base STM-1 is \(155.52\,\text{Mbit/s}\); standard multiples include STM-4 \(622.08\,\text{Mbit/s}\), STM-16 about \(2.488\,\text{Gbit/s}\) and STM-64 about \(9.953\,\text{Gbit/s}\). An add-drop multiplexer can access tributaries without completely demultiplexing every intermediate level.

Add for a 10-mark transport comparison¶

PDH's independently clocked inputs are multiplexed in stages; the receiver removes framing and justification at each stage before recovering a low-rate tributary. Different European, North American and Japanese hierarchies, limited overhead, awkward cross-connection, and weak built-in performance management complicate multivendor networks. Approximate voice-channel counts such as E2 120 or E4 1920 are planning shorthand, not a direct multiplication that reproduces the line rates because framing and justification add overhead.

SDH maps PDH or synchronous tributaries into virtual containers, adds path/section overhead, and places them in a repeating STM frame. Pointers locate payload that may float slightly relative to the frame, avoiding wholesale bit-by-bit stuffing. A terminal multiplexer forms STM signals; an ADM inserts/extracts selected tributaries; a DXC switches paths among ports; regenerators restore long-haul signals. Working/protection rings can switch traffic around a fiber or node failure, while overhead supports alarms, orderwire, performance monitoring and remote management.

SDH therefore reduces equipment, cabling and restoration time and improves interoperability and management. It still needs careful synchronization: bad timing can cause pointer activity and slips, and protection reserves capacity. Stated STM decimal rates and rounded Gbit/s values should not be confused with binary data-rate units. Modern packet/OTN systems often carry Ethernet above or alongside SDH, but the hierarchy remains foundational in transport networks.

Practice target: 9 minutes for definitions/rates or 18 minutes for the add-drop path, elements, exact hierarchy, comparison table, advantages, and caveats.

Exam-ready answer

Digital Subscriber Line (DSL) sends broadband digital data over the existing balanced copper local loop by using frequencies above the POTS voice band. At the customer, a splitter/low-pass filter sends low frequencies to the telephone and high frequencies to the DSL modem/CPE. The same pair reaches an exchange-side splitter: voice goes to the PSTN switch, while data goes to a DSLAM, which terminates and aggregates many loops into the provider IP/backhaul network.

ADSL is asymmetric because residential demand is normally download-heavy, so it assigns more spectrum/subcarriers to downstream than upstream. In a common ADSL Annex A frequency-division plan, POTS occupies roughly \(0\)-\(4\,\text{kHz}\), a guard region separates it from data, upstream is approximately \(25.875\)-\(138\,\text{kHz}\), and downstream approximately \(138\,\text{kHz}\)-\(1.104\,\text{MHz}\). DMT divides spectrum into \(4.3125\,\text{kHz}\)-spaced subcarriers; training measures each tone, bit loading assigns more bits to high-SNR tones, and FEC/interleaving combats impulse noise. Exact usable tones, overlap, pilot, band edges and rates are annex/profile/vendor dependent, so these are the common Annex A values, not universal DSL constants.

HDSL and SDSL provide broadly symmetric upstream/downstream service, while VDSL/VDSL2 uses wider spectrum for much higher rates over shorter loops. ADSL operation is: synchronize and train modems, estimate channel/SNR, negotiate rate and coding, carry framed data while POTS remains available, then retrain or reduce bits if noise changes. Its advantage is reuse of installed copper and simultaneous phone/data; limitations are attenuation with loop length and gauge, crosstalk, bridge taps, poor joints, AM/RF interference, and distance from the DSLAM. Fiber access gives higher and more distance-stable capacity, but ADSL remains an economical legacy last-mile application.

Practice target: 8 minutes; draw modem-splitter-loop-splitter-DSLAM, label the common bands, explain DMT/asymmetry, and state that values depend on annex/profile.

Exam-ready answer

A numbering plan assigns unique, structured addresses to subscribers, networks, geographic areas and services. Under ITU-T E.164 an international number contains country code plus national significant number and has at most 15 digits; the displayed “+” represents the international access procedure and is not one of those digits. A good plan is unique, expandable, easy to analyze, supports emergency/service codes and portability, and supplies enough prefix information for routing and charging.

A routing plan converts the analyzed destination and service class into an outgoing route. A fixed plan always selects one path; alternate routing overflows to another trunk group; hierarchical routing passes local/tandem/transit levels; and dynamic routing reacts to load, failure, quality, cost, or policy. During a call, the originating exchange collects digits, normalizes them, queries portability/service data where necessary, selects an admissible route, attempts a free trunk, and retries an allowed alternate on congestion. The objectives are low blocking/setup delay, efficient trunk use, survivability, loop avoidance, and agreed inter-operator policy.

A charging plan defines when and how usage becomes a charge: flat fee, event, duration, volume, destination/distance, service/QoS, roaming, and peak/off-peak tariff are examples. On answer or another tariff-defined trigger, switches/gateways create CDRs containing calling/called identities, start/answer/end times, duration, route/trunk, service, cause and charge class. Mediation validates and correlates records; rating applies tariff and taxes; billing or prepaid control debits the customer.

These plans interact but are distinct: numbering identifies “whom,” routing chooses “how,” and charging determines “how much.” For example, a mobile number is analyzed, portability data may redirect it to the current network, an interconnect route carries it, and the CDR is rated by the subscriber's plan. Number portability means a prefix need not reveal current physical route, and emergency calls, billing triggers, rounding, tariffs and retained CDR fields are regulatory/operator conventions requiring security, privacy and fraud controls.

Practice target: 8 minutes; define all three plans, trace one number-analysis-to-CDR call, and include E.164 length, routing types, charging factors, and portability caveat.