Please enable JavaScript.
Coggle requires JavaScript to display documents.
Module 9: QoS Concepts, image, image, image, image, image - Coggle Diagram
Module 9: QoS Concepts
-
-
9.3. Queuing Algorithms
- Introduces queuing algorithms as the mechanism used by routers and switches to decide the order in which packets are transmitted when a network interface is congested.
- These algorithms are fundamental to managing delay and jitter in a QoS environment.
- When an interface's output rate is saturated (congested), packets are temporarily held in a queue (buffer) before transmission.
- Queuing algorithms determine which packets drop (when the buffer overflows) and which packets transmit first (priority).
- 9.3.3 First In First Out (FIFO)
- FIFO is the default and simplest queuing mechanism; packets are sent in the exact order they arrive, without any regard for priority or traffic type.
- This method is only suitable for links with very low or no congestion.
- 9.3.4 Weighted Fair Queuing (WFQ)
- WFQ ensures that all traffic flows receive a fair share of the link's bandwidth, preventing any single conversation from monopolizing the queue.
- It assigns weights to prioritize traffic but ensures all low-volume flows (like interactive traffic) are transmitted quickly.
- 9.3.5 Class-Based Weighted Fair Queuing (CBWFQ)
- An enhanced WFQ that allows the creation of user-defined traffic classes based on criteria like protocol or port numbers.
- It guarantees a minimum bandwidth for each class, providing better control over resource allocation than standard WFQ.
- 9.3.6 Low Latency Queuing (LLQ)
- LLQ combines the guaranteed bandwidth of CBWFQ with a strict Priority Queue (PQ) for delay-sensitive traffic (like Voice).
- Traffic in the PQ is served before all other queues, ensuring minimal delay and jitter for real-time services.
9.4. QoS Models
- Introduces the three primary QoS models used to implement quality of service across a network: Best Effort, Integrated Services (IntServ), and Differentiated Services (DiffServ).
- Each model represents a different approach to resource reservation and traffic classification.
- 9.4.2 Selecting an Appropriate QoS Policy Model
- The choice of model depends on the size and complexity of the network and the specific QoS guarantees required.
- Simpler networks may use Best Effort, while complex enterprise or service provider networks require the scalability of DiffServ.
- Best Effort is the default model; all traffic is treated equally, without priority or resource reservation.
- It is the least complex model but offers no QoS guarantees and is prone to congestion and unpredictable delay.
- 9.4.4 Integrated Services (IntServ)
- IntServ uses the RSVP (Resource Reservation Protocol) to explicitly reserve bandwidth end-to-end for a specific application flow before data transmission begins.
- It offers strongest QoS guarantees but is not scalable due to the high overhead of maintaining state for every flow on every router.
- 9.4.5 Differentiated Services (DiffServ)
- DiffServ is the most common model; it classifies and marks traffic into different classes at the network edge.
- Routers within the network (the core) then apply specific per-hop behaviors (PHB) based on the marking, offering good scalability without per-flow state.
-
-
-
-
-
-
