Skip to main content

Quality of Service (QoS) Models

 

Quality of Service (QoS) Models

QoS models define how network traffic is managed to meet performance requirements. The three primary models are:




1. Best-Effort Service

  • Description: No QoS guarantees; all traffic is treated equally ("first come, first served").

  • Characteristics:

    • No prioritization or resource reservation.

    • Suitable for non-critical applications (e.g., web browsing, email).

  • Pros: Simple, no overhead.

  • Cons: Poor performance under congestion (high latency, packet loss).

2. Integrated Services (IntServ)

  • Description: Provides per-flow QoS guarantees by reserving resources before transmission (like a virtual leased line).

  • Key Protocol: RSVP (Resource Reservation Protocol) signals bandwidth requirements.

  • Characteristics:

    • Guarantees strict latency, bandwidth, and jitter for critical apps (e.g., VoIP, video conferencing).

    • Uses admission control to reject flows if resources are insufficient.

  • Pros: High reliability for real-time traffic.

  • Cons: Scalability issues (each flow requires state maintenance).

3. Differentiated Services (DiffServ)

  • Description: Classifies traffic into priority groups (classes) rather than per-flow reservations.

  • Key Mechanism: Packets are marked with DSCP (Differentiated Services Code Point) in the IP header.

  • Characteristics:

    • Traffic is grouped (e.g., EF for Expedited Forwarding, AF for Assured Forwarding).

    • Routers apply Per-Hop Behaviors (PHB) based on class (e.g., priority queuing).

  • Pros: Scalable (no per-flow state), widely used in enterprise/ISP networks.

  • Cons: No absolute guarantees (relative prioritization only).


Comparison Table

ModelBest-EffortIntServ (RSVP)DiffServ (DSCP)
GranularityNonePer-flowPer-class
ScalabilityHighLowHigh
OverheadNoneHighModerate
Use CaseWeb, emailVoIP, telemedicineEnterprise/ISP traffic

Which Model to Use?

  • Best-Effort: Default for most internet traffic.

  • IntServ: Critical real-time apps (e.g., military comms, medical systems).

  • DiffServ: Balanced approach (e.g., prioritizing VoIP over HTTP in a corporate network).



Quality of service models:

Best-effort service

Integrated services

Differentiated services



Best-effort service model – The Internet was initially based on a best-effort packet delivery service. Best-effort is the default mode for all traffic. There is no differentiation among types of traffic.

The best-effort model is similar to standard mail service – it will get there when it arrives.





Advantages of the best-effort service model:

Highly scalable

No special mechanisms required



Disadvantages of the best-effort service model:

No service guarantees

No service differentiation

 

Integrated services (intserv model) are a QoS model that is older than QOS / Differentiated services model. IntServ model defines a signaling process by which an individual flow can request that the network reserve the bandwidth and delay needed for the flow.

 

The integrated services model uses RSVP for reserving bandwidth. RSVP stands for a resource reservation protocol. RSVP  is used for reserving bandwidth/QOS tools for a particular flow.

Here is a limitation of intserv model if reserved no other traffic can use it and it is not scalable each flow required separate admin configuration on the router.




 

 

The differentiated services model is designed to overcome the limitation of best-effort and intserv models. In Diffserv model network traffic is identified by classes and network QOS policy enforces differentiated treatment of traffic classes. You choose the level of service for each class.




Advantages of the Differentiated Service Model:

Highly scalable

Many levels of quality are possible

 

Disadvantages of the Differentiated Service Model:

No absolute service guarantee

Complex mechanisms

 

 

 

 

Comments

Popular posts from this blog

What is Classification and marking? What is IP Precedence and DSCP Values? What is AF and EF?

  Classification and Marking Classification and marking are the processes of identifying the priority of each packet. This is the first step of QOS control and should be done near the source hosts. Classification is the process of identifying and categorising traffic into classes, typically based upon incoming interface IP precedence, DSCP Source or Destination address. Application classification is the most fundamental QOS building block. Without classification, all packets are treated the same.   The table lists the criteria of classification                                  Marking  Marking is the QOS feature component that colours a packet so it can be identified and distinguished from other packets in QOS treatment. Commonly used markers: link-layer: COS (ISL, 802.1q), MPLS EXP bits,...

OSPF Tunnels Vs Virtual-links

OSPF VIRTUAL LINKS  An Open Shortest Path First (OSPF) autonomous system has all areas physically connected to the backbone area (Area 0). In some cases, if this is not possible, we can use a virtual link to connect to the backbone (Area 0) through a non-backbone area. We can also use the virtual links to connect two parts of a partitioned backbone (Area 0) through a non-backbone area. The area through which we configure the virtual link, known as a transit area, must have full routing information. Remember, the transit area cannot be a stub area. Virtual link must be configuring both sides with the same area ID and the corresponding virtual link OSPF neighbor router ID. Show IP OSPF neighbors command give you the router ID information. OSPF Transit Area from Tunnel When configuring OSPF with tunnel interfaces as transit areas, there are several important considerations to ensure proper routing functionality. Here's what you need to know: Key Concepts Transit Area : In OS...

What is of NBAR (network- based application recognition),classification QOS How to configure NBAR?

 When we implement Quality of Service (QoS ) the first step is classification, by default our router does not care about what kind of IP packets it is. Our router just looks at the routing table to find the correct destination IP address and forwards. When we configure QoS technologies like queuing, policing, or shaping before doing first we have to identify what kind of traffic is running on our routers so we can figure out what kind of application it belongs to. This is classification. What is Quality of Service (QOS)? What is Modular Quality of service CLI (MQC)? What is Classification and marking? What are IP Precedence and DSCP Values? What is AF and EF?     Once we identify our traffic (classify). Now we can mark and apply QoS policy to it. There are a few methods on IOS routers for classification: Header inspection It is a simple classification method and it works really good but has some disadvantages. Let’s say we configure our router all th...