Networking Concepts Quiz

Basic Voice Encoding

• Analog signals are sampled, quantized into binary, and compressed to reduce bandwidth.

Digital to Analog Conversion

• The steps for converting digital signals to analog signals are: decompress the samples, decode samples to voltage amplitudes to rebuild the PAM signal, and then reconstruct the analog signal from the PAM signals.

Common Voice Codec Characteristics

• ITU-T standards specify various codecs and their associated bit rates.
• Examples include G.711 (PCM with 64 kbps), G.726 (ADPCM with 16, 24, and 32 kbps), G.728 (LDCELP with 16 kbps), G.729 (CS-ACELP with 8 kbps), and G.729A (CS-ACELP with less computation and 8 kbps).

Mean Opinion Score

• The Mean Opinion Score (MOS) evaluates the quality of voice calls based on user ratings from 1 to 5 (lowest to highest).
• A score of 4.0 equals "Toll Quality."
• Different levels of distortion (imperceptible, just perceptible, and annoying) are associated with each speech quality rating.

DSP Used for Conferencing

• Digital Signal Processors (DSPs) can support single-mode or mixed-mode conferences.
• Mixed mode supports different codecs for flexibility.
• Single mode requires only one codec for all participants.
• Mixed mode conferences per DSP are fewer than single mode conferences.

Voice Transport in VoIP Networks

• Analog phones connect to analog voice gateways.
• Voice gateways convert between analog and digital signals.
• After call setup, the IP network provides packet-by-packet delivery, shared bandwidth, and variable delays.

Jitter

• Voice packets enter the network at a constant rate but can arrive at destinations at different rates or in the wrong order, which causes jitter.
• Receiving routers need to ensure steady delivery and maintain order of packets to avoid delays or queuing problems.

VoIP Protocol Issues

• IP doesn't guarantee reliability, flow control, or error correction.
• VoIP uses TCP or UDP transport layer protocols to manage reliability and sequencing (order of packets).
• TCP offers reliability but consumes extra bandwidth, and retransmission for lost packets isn't always needed for voice.
• UDP offers no reliability, but is efficient as voice data doesn't need to be reordered.
• RTP (Real-time Transport Protocol) is built on UDP and provides the required functionality for the characteristics of voice packets.

Protocols used for VoIP

• Voice does not need reliability, although needs reordering, and time-stamping
• UDP, with RTP, is used as an alternative to TCP, and saves network overhead

Voice Encapsulation

• Digitized voice is encapsulated into RTP, UDP, and IP packets.
• Typically, 20 ms of voice is packetized into a single IP packet.

Voice Encapsulation Overhead

• VoIP packets are small and sent at a high rate.
• IP, UDP, and RTP header overheads are enormous, especially compared to the small size of the payload.

RTP Header Compression

• Compresses IP, UDP, and RTP headers, reducing bandwidth use.
• Configured on a link-by-link basis.
• Reduces header size substantially.
• Saves considerable bandwidth.

When to Use RTP Header Compression

• Use on slow links (less than 2 Mbps) if bandwidth needs to be conserved.
• Consider the disadvantages of CRTP (Compound RTP), which include processing overhead and additional delays.
• Tune CRTP-set the number of sessions to be compressed (default is 16).

Packetization Period Impact

• Higher packetization periods result in larger IP packet sizes, which add to the payload, but lower packet rates, which reduce IP overhead.

Data-Link Overhead

• Ethernet, Frame Relay, and MLP protocols have differing data-link overheads (in bytes).

Security and Tunneling Overhead

• IP security (IPsec) and tunneling protocols can add overhead to VoIP packets.
• Encapsulation of original frames into another protocol enlarges packets and increases bandwidth requirements.
• Increased bandwidth requirements are important for voice packets because of their small packet size and high transmission rate.

Total Bandwidth Calculation Procedure

• Gathering packetization information (period or size) and codec bandwidth. Gathering link information (CRTP, data-link protocol, and IPsec protocols). Calculating packetization size or period. Summing packet size, all headers, and trailers. Calculating packet rate. Calculating total bandwidth.

Bandwidth Calculation Example

• Formula for calculating bandwidth in kbps, using total packet size, bytes per packet, IP overhead, packetization period, and packet rate.
• Example values are included.

Quick Bandwidth Calculation

• Formula for calculating total packet size (bits).
• Provides parameters and values for G.711 and G.729 codecs, and examples for Frame Relay.

Enterprise Voice Implementations

• Gateways, gatekeepers, Cisco Unified Call Manager, and IP phones are constituents of enterprise voice networks.

Deploying CAC

• Call Admission Control (CAC) helps limit concurrent voice calls within WAN resources.
• Needed due to QoS problems in setting up too many voice calls.

Cisco Unified CallManager Functions

• Headquarters manage call processing, dial plan administration, signaling, phone features, directories, and XML services; they also provide an interface to external applications.

Example: Signaling and Call Processing

• Cisco Unified CallManager cluster handles signaling for call setup and call processing.

Enterprise IP Telephony Deployment Models

• Provides different deployment models (single site, multisite, distributed, and clustering over WAN) with different characteristics for each model.

Single Site

• CallManager, applications, and DSP resources are located together.
• IP WAN is not used for internal voice traffic, and the PSTN is used for external calls.

Multisite with Centralized Call Processing

• CallManager servers and applications are centralized, but DSP resources are distributed among other sites.
• Voice signaling over IP WAN and data traffic over IP WAN.

Multisite with Distributed Call Processing

• All components (CallManager, applications, and DSPs) are located at each site.
• All inter-site calls use the WAN (signaling and media) and use PSTN when WAN is down.

Clustering over WAN

• All CallManager and DSP resources for a deployment are deployed over multiple sites.
• All calls use the WAN (signaling and media) with a 40 ms or less delay.

VoIP QoS

• Introduction of QoS, its requirements, and problems to solve.
• QoS is needed to help interactive voice, video conferencing, and other applications receive high quality of service.

Converged Network Realities

• Constant small-packet voice flow competes with bursty data flows.
• Critical traffic must have priority.
• Voice and video are time-sensitive.
• Network outages need to be minimized

Converged Network Quality Issues

• Lack of bandwidth, end-to-end delay (fixed and variable), and variation from delay cause some of the problems affecting quality issues in a VoIP network.

Increasing Available Bandwidth

• Upgrading the link (best solution but expensive).
• Improving QoS for critical traffic.
• Compressing Layer 2 frames and headers.

Using Available Bandwidth Efficiently

• Using advanced queuing mechanisms, and header compression improve bandwidth usage for voice and interactive traffic.

Types of Delay

• Processing, queuing, serialization, and propagation delays need to be considered.

Ways to Reduce Delay

• Upgrading the link, forwarding important packets first, reprioritization and frame/header compression.

Reducing Delay in a Network

• Customer and ISP (Internet Service Provider) routers perform TCP/RTP header compression, LLQ, and reprioritization (according to the QoS policy)

Impacts of Packet Loss

• Problems like voice breaking up, jerky video, corrupted files, and hold messages occur when packet loss is experienced in a VoIP network.

Types of Packet Drops

• Output queues filling up and congestion on a link are common reasons for packet loss or drops in a VoIP network.

Ways to Prevent Packet Loss

• Upgrading the link or increasing bandwidth usage to meet traffic demands.
• Preventing congestion by dropping less-important packets.

What is Quality of Service?

• Network managers need to control delay, jitter, and packet loss to manage bandwidth allocations and to provide desired network application performance.

Different Types of Traffic Have Different Needs

• Real-time applications (voice) are especially sensitive to QoS metrics such as delay, jitter, and packet loss.
• Management of bandwidth allocations deliver desired quality of experience in application performance.

Cisco IOS QoS Tools

• Queue management (PQ, CQ, WFQ, and CBWFQ)
• Congestion management (WRED)
• Link efficiency (fragmentation & interleave)
• Traffic shaping and policing (RTP & CRTP)

Priority Queuing

• PQ puts traffic into queues based on priority (high, medium, normal, and low).
• Classification determines queue length for priority.
• Output hardware interface hardware manages priority scheduling.

Custom Queuing

• Traffic is classified, interface buffer resources are managed, and link bandwidth is allocated in proportion to traffic demands.

Weighted Fair Queuing

• Uses weighted fair scheduling to share network resources among different flows based on weights set.
• Weights are determined by QoS (requested by the applications or flows) and the throughput of each flow.

Weighted Random Early Detection

• WRED stochastically discards packets at random when congestion begins to increase on a link.
• Provides a method to avoid packet loss if the queue occupancy increases.

Implementing QoS

• Steps include identifying traffic types, dividing traffic into classes, defining QoS policies for each class.

Step 1: Identify Types of Traffic and Their Requirements

• Network audit: Identify traffic types and network bandwidth requirements.
• Business audit: Determine the importance of each traffic type for business operations.
• Setting service levels: Determine the required response time for different traffic types.

Step 2: Define Traffic Classes

• Classifying traffic based on characteristics (voice, video, data) and providing predefined priorities, which are used in QoS policy and configurations.
• The important issues in determining important traffic factors are low latency, bandwidth requirements (guaranteed), guaranteed delivery, and/or no guaranteed delivery.

Step 3: Define QoS Policy

• Implementing policy for specific features, service levels, and prioritizing each class (voice, best-effort).
• Provides wide-ranging definition for specific levels of QoS for different classes of network traffic.

Quality of Service Operations

• Classify traffic, establish queues, and selectively drop or prioritize traffic by using QoS features and tools.

Three QoS Models

• Best-effort: No QoS is applied to packets.
• Integrated Services (IntServ): Applications signal QoS requirements.
• Differentiated Services (DiffServ): Network recognizes and classifies traffic classes that require QoS.

Best-Effort Model

• No QoS (Quality of Service).
• Simple, highly scalable.
• No service guarantees.

Integrated Services (IntServ) Model Operation

• Guaranteed and predictable network behavior for applications with multiple service levels.
• Signaling protocol (RSVP) to reserve resources based on specified QoS parameters by application.
• QoS parameters are linked to packet streams.

IntServ Functions

• Control plane: Routing selection, admission control, and reservation setup.
• Data plane: Flow and packet scheduling.

Benefits and Drawbacks of the IntServ Model

• Explicit admission control.
• Per-request policy admission control.
• Continuous signaling.
• Large, complex implementations aren't scalable.

The Differentiated Services Model

• Overcomes IntServ's limitations.
• Soft QoS (rather than hard QoS).
• Classifies flows into aggregates.
• Minimizes signaling and state maintenance requirements.

Methods for Implementing QoS Policy

• Legacy CLI (Time consuming, individual interface configuration)
• MQC (Uses configuration modules, good for QoS fine-tuning)
• Cisco AutoQoS (Applies a possible QoS configuration to interfaces)
• Cisco SDM QoS wizard (Good for simplified QoS configurations)

Modular QoS CLI

• Command syntax reduces configuration steps and time.
• Configures policy, not individual commands.
• Uniform CLI for different QoS platforms.
• Separates classification engine from policy.

Modular QoS CLI Components

• Define traffic classes by using a class map.
• Define QoS policies by using a policy map to define actions based on traffic classes.
• Apply QoS policy to network interfaces using a service policy.

Step 1: Creating Class Maps

• A class map defines traffic classes based on match conditions.

Configuring Class Maps

• Configure class map mode & matching strategy with commands.

Classifying Traffic with ACLs

• Standard ACLs and extended ACLs allow classifying traffic by the source, destination, port.

Step 2: Policy Maps

• Policy maps define QoS policies for specific classes of network traffic.
• Provide commands to define case-sensitive policy names.
• Defines QoS policies for each class.

Configuring Policy Maps

• Configure policy maps by entering the per-class policy configuration mode.

Step 3: Attaching Service Policies

• Attach QoS policy maps to network interfaces (input or output).

Modular QoS CLI Configuration Example

• Example shows how to configure traffic classes, QoS policies, and apply policies to interfaces. This example uses MQC Configuration model.

MQC Example

• Example shows a voice traffic needing priority versus interactive data.

Basic Verification Commands

• Commands such as show class-map, show policy-map, and show policy-map interface type number to display QoS configuration.

Implement the DiffServ QoS Model

• Introduces the DiffServ QoS model which overcomes many limitations of best-effort and IntServ Models.

Classification

• Classifies traffic based on incoming interface, IP precedence (DSCP), source/destination address, or application.

Marking

• Marking is the component used to differentiate traffic into classes by tagging with certain values.
• Protocols used for QoS marking are CoS (802.1p), MPLS EXP bits, Frame Relay, DSCP and IP precedence for the network layer.

DiffServ Model

• Describes services associated with traffic classes, unlike IntServ, which defines per-flow state.
• Complex traffic conditioning is performed at the edge.
• No protocol state is required inside the core network infrastructure.

Classification Tools

• IP Precedence and DiffServ Code Points (DSCP) are classification tools used to differentiate traffic.

IP TOS Byte and DS Field Inside the IP Header

• Shows the position of the Type of Service (ToS) fields in IPv4 packet header.

IP Precedence and DSCP compatibility

• Shows how IP precedence in version 4 IP packets is compatible with DiffServ code point (DSCP) used in version 4 packets.

Queuing

• Introduction to queue structures in QoS.

Test your knowledge on networking delays, packet loss, and traffic management techniques. This quiz covers essential concepts related to how data packets interact in a network. Gauge your understanding of various methods to optimize network performance and reduce latency.