NET4007: Exam Set #1

exam Set based off of review slides 1

What is the Nyquist Theorem?

It states that to correctly sample a signal, the sampling rate must be at least twice the maximum frequency content in the signal.

What is quantization in digital signal processing?

Quantization is the representation of signal amplitudes by discrete values, introducing rounding errors called quantization noise.

Define PCM (Pulse Code Modulation).

PCM converts audio signals into a digital format by sampling and quantizing. Variants include DPCM and ADPCM.

What is the main difference between lossless and lossy compression?

Lossless compression retains the exact original data, while lossy compression provides an approximation for higher compression ratios.

What is the LZW algorithm used for?

It is an adaptive, dictionary-based compression technique, used in formats like GIF, V.42 bis for modems, and UNIX compress.

How does the LZW algorithm work?

LZW builds a dictionary of variable-length strings and encodes these strings with fixed-length codes.

What happens when the LZW dictionary reaches its size limit?

The dictionary is either flushed or the least recently used (LRU) entries are removed to allow further encoding.

Why is LZW efficient for repetitive patterns?

It replaces repeated patterns with shorter codes by expanding the dictionary dynamically.

What is QoS in networking?

QoS (Quality of Service) manages bandwidth, delay, jitter, and packet loss to meet application performance requirements.

Name the three QoS models in Cisco IOS.

Best Effort, Integrated Services (IntServ), and Differentiated Services (DiffServ).

What are the main Cisco IOS QoS tools for traffic management?

Priority Queuing (PQ), Custom Queuing (CQ), Weighted Fair Queuing (WFQ), and Class-Based Weighted Fair Queuing (CBWFQ).

What is Low Latency Queuing (LLQ)?

LLQ adds a priority queue to CBWFQ to handle real-time traffic like voice and video with minimal delay.

How does Weighted Random Early Detection (WRED) work?

WRED stochastically discards packets to prevent congestion before queues fill completely based on IP precedence (8 profiles) and DSCP (64 profiles).

What is the role of Modular QoS CLI in Cisco IOS?

It simplifies QoS configuration by allowing policy definitions separate from interface-level commands.

What is MPEG, and why is it significant?

MPEG (Moving Picture Experts Group) develops compression standards for audio and video, used in multimedia streaming and storage.

Describe the main compression techniques used in MPEG.

MPEG employs techniques like motion estimation, transform coding (e.g., DCT), and entropy coding (e.g., Huffman Coding).

What is the role of I-frames, P-frames, and B-frames in MPEG compression?

I-frames are complete images.
P-frames predict changes from previous frames.
B-frames predict changes between surrounding frames.

How does MPEG achieve high compression ratios?

By reducing spatial redundancy within frames and temporal redundancy between frames.

What is Sound?

Sound is a wave phenomenon involving molecules of air being compressed and expanded, typically perceived as longitudinal pressure waves.

What does the Nyquist Theorem dictate about sampling rates?

To recover the original sound, the sampling rate must be at least twice the maximum frequency content of the signal.

What is quantization noise?

It is the rounding error introduced when amplitudes are approximated to the nearest quantization interval during digitization.

What does the Signal-to-Noise Ratio (SNR) measure?

The SNR measures the ratio of the power of the correct signal to the noise, often expressed in decibels (dB).

What is compression?

Compression reduces the number of bits needed to represent certain information. It can be lossless or lossy.

What is entropy in information theory?

Entropy measures the average amount of information contained per symbol in a source and provides a lower bound for average codeword length.

What are the steps of the Shannon-Fano Algorithm?

1. Sort symbols by frequency.
2. Divide symbols into parts with approximately equal frequencies.
3. Repeat until all parts contain one symbol.

What is the main principle behind Huffman Coding?

Huffman Coding creates minimum redundancy codes, assigning shorter codes to more frequent symbols and longer codes to less frequent symbols.

What is the purpose of Run-Length Coding (RLC)?

RLC compresses data by encoding sequences of repeated symbols as a single value and count, reducing data size.

What is lossy compression?

Lossy compression approximates the original data, yielding a much higher compression ratio at the cost of some fidelity.

What does the Discrete Cosine Transform (DCT) achieve in transform coding?

The DCT reduces redundancy by decorrelating data, allowing efficient coding of significant information in fewer components.

What are the benefits of VoIP networks?

VoIP networks use bandwidth efficiently, reduce costs, integrate telephony with business applications, and provide mobility with software-based phones.

What causes jitter in VoIP?

Jitter occurs when voice packets arrive at varying rates or out of order, disrupting timing and order dependencies.

What is Cisco's Low Latency Queuing (LLQ)?

LLQ prioritizes real-time traffic, ensuring minimal delay for voice and video applications.

How does the Cisco IOS Differentiated Services (DiffServ) model work?

DiffServ classifies flows into traffic aggregates and assigns per-hop behaviors to manage QoS efficiently at scale.

What are the drawbacks of the IntServ model?

IntServ requires continuous signaling and is not scalable to large implementations, such as the public internet.

How does RTP Header Compression reduce bandwidth usage?

It compresses the IP, UDP, and RTP headers from 40 bytes to as little as 2 or 4 bytes, conserving bandwidth on slow links.

What is the role of policy maps in QoS?

Policy maps define QoS policies for traffic classes identified using class maps, associating actions with these traffic classes.

How does Adaptive Multi-Rate (AMR) speech coding conserve bandwidth?

AMR conserves bandwidth by adjusting bit rates and useing discontinuous transmission (DTX), voice activity detection (VAD), and comfort noise generation during silence periods.

What is the purpose of Call Admission Control (CAC)?

CAC limits concurrent voice calls to prevent WAN oversubscription and poor quality due to excess voice traffic.

What is the difference between IntServ and DiffServ?

IntServ guarantees per-flow QoS with explicit resource reservations, while DiffServ aggregates flows and applies class-based QoS for scalability.

What does a Linear Predictive Coding (LPC) do?

LPC vocoders extract salient features of speech directly from the waveform, rather than transforming the signal to the frequency domain

Linear Predictive Coding (LPC) Features

• time-varying model of vocal-tract sound

• transmits only a set of parameters → small bit rate

• “linear” = the speech signal is calculated as a function of current speech plus second term in previous model

What is Comfort Noise?

Comfort noise (or comfort tone) is artificial background noise used in radio and wireless communications to fill the silence in a transmission resulting from voice activity detection or from the clarity of modern digital lines.

When to Use RTP Header Compression?

Only on slow links (less than 2 Mbps) If bandwidth needs to be conserved

Converged Network Quality Issues

1. Lack of bandwidth

2. End-to-end delay

3. Variation of delay

4. Packet Loss

Methods for Implementing Cisco QoS Policy

• Legacy CLI

• MQC

• Cisco AutoQoS

• Cisco SDM QoS Wizard

What is a Modular QoS CLI

A command syntax specific for configuring QoS policy. It reduces configuration steps and time and is a uniform CLI accross major Cisco IOS platforms

Identify three novel multimedia applications in wireless or wireline networks. Discuss why you think these multimedia applications are novel

a. Cloud Gaming; low latency networks

b. Telemedicine platforms; remote monitoring, secure networks

c. Smart Video Surveillance: multimedia-streaming for real-time application

Identify three problems with current wireless or wireline networks in supporting multimedia applications. List some possible solutions.

a. Latency Issues: Implement edge computing

b. Security Concerns: use encryption

c. Packet Loss: Implement protocols to mitigate loss

Your task is to design a system that transmits smell over the Internet. Suppose we have a smell sensor at one location and wish to transmit to Aroma Vector (say) to a receiver to reproduce the same sensation. List the major challenges in this system and possible solutions.

i. Standardization of Smell: create a new standard format

ii. Latency and Syncronization: Implement buffering mechanisms

III. Safety: FDA-Approved chemicals for smell reproduction

Assume that you want to increase the number of bits per sample in the quantization from 8 to 10. Discuss the difference between the new system and the original one in terms of required bandwidth and peak signal and peak quantization noise ratio (PSQNR).

a. Bandwidth: →increasing from 8 to 10 increases the data rate by 10/8 = 1.25 (meaning 25% more bandwidth needed)
b. PSQNR = 6.02 x N → 8bits = 48 → 10bits = 60 =→ Increased by 12 dB

Encode “OTTAWA” using Huffman coding algorithm.

a) Count the occurrences: O=1. T=2. A=2, W=1

b) Min-heap priority queue: O=1, W=1, T=2, A=2

c) Huffman Tree:

i) 2

/ \

O W

ii) 4

/ \

2 T

/ \

O W

iii) 6

/ \

A 4

/ \

2 T

/ \

O W

d) Assign binary codes:

A=0, T=11, O=100, W=101

e) Encode the word “OTTAWA”:

Using the binary codes from before = 100 11 11 0 101 0 = 1001111010100

What are the major differences between a vocoder and a general coder?

Vocoder is specialized for speech while a general coder is used for a diverse set of data, without specializing on a specific one

What is sound?

Sound is a macroscopic wave phenomenon involving air molecules being compressed and expanded, creating a longitudinal pressure wave perceived as sound.

What is digitization in the context of sound?

Digitization converts an analog signal to a stream of numbers, typically integers, for efficient digital representation.

What are the key steps in digitizing sound?

1. Sampling: Measure at evenly spaced intervals (determined by sampling rate).
2. Quantization: Round amplitude to the nearest discrete value.

What is differential coding?

Differential coding transmits the difference between current and previous samples, reducing data requirements.

What is ADPCM, and how is it adaptive?

Adaptive Differential PCM adjusts the quantizer and predictor dynamically to suit varying signal characteristics.

What is entropy in information theory?

Entropy measures the average information content per symbol, setting a theoretical limit for compression efficiency.

What is the Shannon-Fano Algorithm?

A top-down coding method dividing symbols into parts with approximately equal frequencies, assigning shorter codes to common symbols.

What is Huffman coding?

A bottom-up compression algorithm creating optimal variable-length codes for symbols based on frequency.

How does Run-Length Coding (RLC) work?

RLC encodes runs of repeated values as a single value and count, reducing redundancy effectively for repetitive data.

What is the Lempel-Ziv-Welch (LZW) algorithm?

LZW is an adaptive, dictionary-based compression technique that encodes repeated patterns with fixed-length codes.

What is the purpose of transform coding?

Transform coding reduces redundancy in image or audio data by decorrelating components for efficient compression.

What is the Discrete Cosine Transform (DCT)?

DCT converts spatial data into frequency components, separating significant information for compression.

What is Vector Quantization (VQ)?

VQ uses a codebook of vectorized samples to encode data efficiently by matching to the nearest code vector.

What are vocoders?

Vocoders compress speech by modeling it, either in time (e.g., LPC) or frequency (e.g., formant vocoders).

What is Linear Predictive Coding (LPC)?

LPC encodes speech by modeling vocal tract dynamics and transmitting coefficients representing its shape and excitation.

What is the difference between Code Excited Linear Prediction (CELP) and Algebraic CELP?

CELP matches excitation vectors to speech samples from a codebook, while ACELP uses algebraic structures for efficiency.

What is Adaptive Multi-Rate (AMR) coding?

AMR adjusts bit rates dynamically based on network conditions, optimizing bandwidth and quality.

What are the three major steps in QoS?

• Classification

• Policing

• Shaping

What is the difference between Policing and Shaping?

Policing; Limits bandwidth by discarding traffic, Should be used on higher-speed interfaces, Can be applied inbound or outbound

Shaping; Limits excess traffic by buffering, Recommended for slower-speed interfaces, Can only be applied in the outbound direction