Please enable JavaScript.
Coggle requires JavaScript to display documents.
Real-time Communication in Packet-Switched Networks, Aluno: Alexandre…
Real-time Communication in Packet-Switched Networks
Introduction
Unsatisfactory Approaches to Real-Time Communication
Circuit-switching
Buffering
Bandwidth
Goals for Real-Time Communication Techniques
Low jitter
Low latency
Ability to easily integrate non-real-time and real-time services
Adaptable to dynamically changing network and traffic conditions
Good performance for large networks and large numbers of connections
Modest buffer requirements within the network
High effective bandwidth utilization
Low overhead in header bits per packet or cell
Low processing overhead per packet within the network and the end system
Characteristics of Real-Time Traffic
Examples
Multimedia conference
Shared workspaces
Remote medical diagnosis
Telephony
Command and control systems
Distributed interactive simulation
Audio and video broadcasts
Games
Small packets
minimize packetization delays
limit effect of packet loss
Sources
constant bit rate (CBR)
variable bit rate (VBR)
on/off source
periodic with variable packet sizes
Hard-Real-Time Communication
General Remarks
Rate-based methods
Scheduler-based methods
Intolerant to packet losses
Real-Time Scheduling Theory
Analogous to OS scheduler
Earliest Due Date (EDD) = EDF
Traffic Characterization
QOS
delay
jitter
loss bounds
Peak-rate model
minimum inter-arrival time T_i
maximum packet length 𝜏_i
end-to-end deadline D_i
bandwidth p_i = 𝜏_i/T_i
exact for CBR, overstates VBR
Linear Bounded Arrival Process (LBAP)
maximum burst size σ_i
leaky bucket
Connection-Level Processinga
admission control
determines if resources are available (forward) and allocate it (backward)
Scheduler-Based Methods
EDD-D: Earliest due-date for delay
EDD-J: Earliest due-date for jitter
SRT: smallest response time
PCT: preemptive cut through
Rate-Based Methods
HRR: hierarchical round robin
S&G: stop-and-go
WFQ: weighted fair queueing
PGPS: packet generalized processor sharing
RCSP: rate-controlled static priority
Per-Packet Processing
Processing steps
Packet demultiplexing
Queue insertion
Queue multiplexing
Input regulation
Policies
work-conserving: output link will never be idle as long as there are packets waiting for link
non-work-conserving: reduce jitter
Scheduler-Based Methods
packet holding to decrease jitter
SRT and PCT: preemption
Rate-Based Methods
Stop-and-Go
HRR
RCSP
Implementation Requirements
buffer space
Summary
EDD-J: minimizes buffer and hitter
PCT minimizes end-to-end latency
S&G: jitter, buffer, delay proportional to frame size
Soft-Real-Time Communication
Application-Level Characteristics
QOS
delay
jiiter
max packet loss
Connection-Level Issues
Multiple Traffic Classes and Grades of Service
bound delays
late loss
best effort
Providing Statistical Guarantees on Delay and Loss
Source-based approach
Bounding approach
Observation-based approach
Best effort delivery
Synchronization of Soft-Real-Time Traffic
Per-Packet Processing
Priority Policies
Laxity-Based Policies
Conclusion
Routing
QOS similar to telephone
Fault tolerance
Error Control
Synchronization
Multicasting
Aluno: Alexandre Nadolni Bonacim