Concept Of Quality Of Service Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

QOS was basically introduced in telecommunication networks. With an increase in internet capabilities and introduction of real time streaming media and VOIP etc applications, QOS has become a major research area. The traditional best-effort service of the internet had limited ability and could not guarantee packet delivery or an error free transmission. QOS guarantees proper transmission of an error free, low jitter and bit rate in packets. It ensures proper and uninterrupted data flow which is most essential in streaming media and VOIP applications. QOS reserves resources for certain applications when there is increase in the flow of traffic in a network.

The main function of QOS is to prioritise application software and users. By this QOS ensures every user/application and their requirements are serviced without any delay and to their satisfaction. Quality of Service encompasses all the characteristics of a system or a connection in terms of audio quality, video quality, noise levels, the time a connection takes to respond, allocation of resources to the applications. It also requires applications to be co-operative in sharing resources. Apart from multimedia users, QOS has a distinct place in distributed systems and high security and reliability seeking defence purpose systems. QOS is a major requirement to satisfy all these necessities.

Quality of Service is often measured by the rate of user satisfaction or the Happiness quotient of the user. Quality of Service is therefore given high significance because it is in direct correlation with user happiness.

Concept of QOS in Systems:

QOS in system deals in resource sharing among applications and users within the system. Every user has unique requirements which vary over different applications. Sometimes a user may prioritise video stream over online updates, or may chose video chat over data downloads. QOS ensures these prioritising seamlessly for ease of the user. Even with regard to applications, in some applications higher bandwidth is just a option to perform better whereas in some applications it is a necessity. Therefore Quality of Service reserves some set of resources which are common and have same utility for diverse applications to expand their effectiveness.

Quality of Service can also be application specific for different systems. Systems can be generalised in terms of their requirements like timing, reliability, security and ease of use. For examples defence and banking systems require higher security and reliability whereas traffic monitoring systems have higher timing needs. The ability to prioritise sharing resources depending on specific requirements of different systems is the main functionality of Quality of Service.

QOS in Networks:

Network capacity and availability is put under tremendous test with the advent of streaming media, VOIP, IP telephony etc. Guaranteed service and error free transmission of data is a huge challenge in these conditions. These challenges are met by the introduction of QOS in networks. QOS in networks deals with allocation of resources in the network. End to end transfer of data requires high reliability on both source and destination. QOS manage resources throughout this transfer. It depends on support from all components involved in a network like routers, switches, and Interface cards to support it. QOS ensures these components provide co-operation during the transmission of data. Packet loss, packet delay, jitter, high bit rates and low latency are some major issues in a Network. Many Bandwidth management techniques are used in the network to improve the quality of service. Response time of the connection is increased with the application of these techniques. QOS in network is generally ensured by over providing of the resources before hand or by per hop behaviour depending on the type of requirement and the QOS model being used.

Problems in QOS:

Quality of service deals with many problems during its functioning. They can be categorised depending on the reason affecting them. They can be caused because of the lack of stability and availability of services. Delay and factors like dependability, scalability, usefulness, and maintenance are major factors which cause degrading of the quality of service. These problems are addressed by ensuring right services and QOS tools both on the network and system side. Quality of service is an important factor for high performance of the network and optimal use of the resources to user satisfaction.

4. QOS Architecture:

The architecture of QOS is represented in figure 1 below. QOS has 3 basic parts in its implementation in a network comprising of a client and server, along with many connected systems. QOS is applied in these parts in various techniques which are listed below.

Inside the network node QOS uses various QOS tools like scheduling, queuing, traffic shaping, congestion management etc.

Queuing- Is a tool used to decide the mechanism to be used when there is a queue overflow or underflow. It also ensures that packets marked with higher priority are given precedence ahead of others. While doing this it also has to make sure that packets with lower priority are discarded ahead of the ones with higher priority if there need be. Packets with high priority are moved ahead in the queue and the tool reserves some space for them when the queue is becoming full.

Shaping- It is a tool used to shape the flow of packets in the network. This tool is used to prevent overflow in heavy traffic. Examples of shaping technique are Token Bucket Algorithm and Leaky Bucket Algorithm. They buffer the traffic and keep them in a waiting state and release them when the traffic eases.

Congestion Management- It is a tool used in bursty traffic when amount of data packets surpass the capacity of the link. Congestion management tool include priority queuing, custom queuing, weighted fair queuing, and class based weighted fair queuing.

Figure 1: QOS Architecture [2]

QOS synchronises data transfer from source to destination using its Identification and Marking technique.

To prioritise the traffic, QOS has to first recognise its type, check if the packets in it are marked. This process is known as classification.

After identifying the type, QOS reserves the resources and transfers them to their destination.

Synchronisation is an important requirement in a network transfer to ensure packets are delivered without error or delays. It also ensures sequencing of the marked packets to avoid any out of order delivery of the packets.

Its policy management, accounting functions.

Policing technique is used when the traffic shaping technique cant prevent the overflow. Policing tool then restricts the traffic usually by discarding the packets.

Most of the network giants today have introduced their own quality of service policy managers which help in policing the traffic and also resource allocation.

5. QOS Broker:

QOS also adopts a QOS Broker, who brokers and negotiates for the system resources in the home system as well as in the worldwide network. QOS Broker in the local system maintains a record of all the requirements of the system to help allocate resources according to their needs. In the network it negotiates between the network and the operating system to ensure efficient network transfers. QOS Broker works in the same manner as a human broker. A human broker has prior knowledge of what the buyer wants and what a seller has to offer. He then coordinates between them and works out a scheme which is satisfactory to the both. Similarly a Network QOS Broker negotiates between a buyer (endpoint websites) who want to trade resources with a seller (isolated site). These resources have QOS parameters which specify the characteristics of applications and also their priorities. Parameters can be on both sides, operating side and network side. On the operating system side along with characteristics it also includes communication topology, processing time. On the network side parameters include specifications of bandwidth, error reports, jitter etc.

Figure 2: QOS Broker

6. Priority Levels in QOS:

Every application is assigned with a priority level depending upon user requirements. Best effort services are by default the lowest priority types. All system applications running in the background are given priority level 1. Streaming videos come with higher priority but they have lesser when compared with VOIP and video conferencing applications. These applications also require low latency and jitter to perform to their optimum. Network control applications in layer three are the next level of traffic. The traffic type with highest priority is Layer 2 network control reserved traffic. This type experiences the lowest latency and jitter and absolutely error free.

Priority Level

Traffic Type

0 (lowest)

Best Effort




Standard (Spare)


Excellent Load

(Business Critical)


Controlled Load

(Streaming Multimedia)


Voice and Video

(Interactive Media and Voice)

[Less than 100ms latency and jitter]


Layer 3 Network Control Reserved Traffic

[Less than 10ms latency and jitter]

7 (highest)

Layer 2 Network Control Reserved Traffic

[Lowest latency and jitter]

Priority level [1]

Quality Of Service Models:

The two models in Quality Of Service are:

Integrated Service (IntServ)

Differentiated Service (DiffServ)

These two models provide QOS an extranet. On any kind of network, Integrated Service is a general approach to QOS and Differentiated is an approach for large scale networks. The real push for these enhanced service architecture came in 1990's, after large scale video conferencing experiments over the internet. Real time resources such as video conferencing are very sensitive to the timeliness of data and these applications do not provide good result in the internet, where the latency cannot be predicted. The delay and jitter requirements of these applications require a new way of providing service that can provide some level of resource guarantee to the applications.

Hence to overcome these bandwidth problems, queuing delay and jitter problem, Integrated Service and Differentiated Service models were developed. These two architectures introduce a number of new concepts and ways which are important to QOS support in the world of internet.

Some of the concepts and primitives of QOS are as follows:

Frameworks for resource allocation which supports application guarantee and service differentiation.

In addition to the existing Best Effort Services, new service models for the internet were developed.

There was a need for a language that provides resource guarantee and its requirements.

Resource allocation enforcing mechanism.

These two services i.e Integrated Service (IntServ) and Differentiated Service (DiffServ), represents the two different solutions for the systems in a network. Resource assurance through resource reservation for individual application flow is provided by IntServ whereas DiffServ use a combination of edge policing, provisioning and traffic prioritization.

The IntServ architecture is based on per-flow resource reservation. To receive resource assurance, an application must make a reservation before it can transmit traffic onto the network. The DiffServ architecture is very simple, it offers different network service levels to a packet, and thereby it enables scalable service quality in the internet without the need for per-flow state. DiffServ routers apply pre-provisioned Per-Hop Behaviors (PHBs) to packets according to the encoded forwarding class. Let us discuss about these two Quality Of Service models in detailed in the coming section.

Integrated Service (IntServ):

IntServ model was the first step to enhance the internet with QOS functionalities. MBONE experiment conducted by researchers took them into an advance mode to think over the development of IntServ. The MBONE was a multicast network built on the internet for audio and video conferencing. From MBONE architecture, researchers concluded that the significant augmentations to the internet architecture were required to provide back bone support for the real time applications over the internet.

In 1990's Algorithm development and mechanism for supporting advanced resource allocation was made as true progress by the networking research group. Fair queuing algorithm was used to deal with the delay and the bandwidth guarantee for packet scheduling. Therefore by these ideas strict resource allocation and high resource utilization was out upon.

The development of Integrated Service model was a great technical achievement for the internet group. This new concept of IntServ shows considerable deviation from the datagram model on which the internet was built. Many new ideas were developed including new service model, RSVP protocol, routing specifications and scheduling algorithms.

The Internet Engineering Task Force (IETF) setup many working groups to evaluate by comparing with a standard service models and set of rules for the IntServ model.

The main functions of IntServ working community are as follows:

The IntServ model is responsible for stating the architecture, service models, flow specifications and other elements such as admission control, flow recognition and packet scheduling.

The IntServ over Specific Link Layers (ISLL) describes the methods required to apply IntServ capabilities within SLL technologies like Ethernet and ATM.

The RSVP working community establishes a standards for resource reservation setup applications set of rules that reserves a specific state in a network.

The Integrated Services model has been qualified by resource reservation. In the real-time applications, the applications must first decide the paths and reserve resources before data are transmitted. RSVP is a set of rules for signaling and deciding paths for reserving the resources.

Integrated Services (IntServ) Model Architecture:

The Integrated Services (IntServ) model was developed to provide individualized QOS guarantee to individual session. Here reservations are made upon per simplex flow. Applications request reservations for network resources which are granted or denied based on resource availability. Senders specify the resource requirements via a PATH message that is routed to the receiver. In turn the Receivers reserve the resources with a RESV message that follows the reverse path.

The Basic architecture of IntServ is shown below in the diagram and the second figure describes the RSVP signaling.

Fig: Architecture

In Fig 1 the sender sends a PATH Message to the receiver by setting the qualities of the traffic. Each arbitrate router along the path sends the PATH Message to the next hop mentioned by the routing protocol. After receiving a PATH Message, the receiver sends a RESV Message to request resources for the transmission. Every arbitrate router in the path may reject or accept the request of the RESV Message. If the request is declined, the router will respond by sending an error message to the receiver, and the signaling process will stop. If the request is granted, bandwidth link and spacing of the buffer are assigned for the flow of the data and the related flow state message will be setup in the router.

IntServ Components:

The Integrated Services Model can be divided into two parts - the Control and Data Planes

Control Plane:

Route Selection - Identifies the route to follow for the reservation

Reservation Setup - Installs the reservation state along the selected path

Admission Control - Ensures that resources are available before allowing a reservation

Data Plane:

Flow Identification - Identifies the packets that belong to a given reservation

Packet Scheduling - Enforces the reservations by queuing and scheduling packets for transmission

IntServ-Service Models:

Applications using IntServ can request two basic service-types:

Guaranteed Service:

Provides guaranteed bandwidth and queuing delays end-to-end, similar to a virtual-circuit. Applications can expect hard-bounded bandwidth and delay.

Controlled-Load Service:

Provides a Better-than-Best-Effort service, similar to a lightly-loaded network of the required bandwidth. Applications can expect little to zero packet loss, and little to zero queuing delay. These services are mapped into policies that are applied via CB-WFQ, LLQ, or MDRR.

Differentiated Service (DiffServ):

The Internet Engineering Task Force (IETF) developed DiffServ, the most illustrious result for providing QOS over internet network. This model is capable of giving different levels of services for the internet group. The traffic flows are thus differentiated and grouped into different forwarding classes. This type of working is achieved by Assured Forwarding Per-Hop Behaviour (AFPHB's) mechanism. In this working, data packets are observed and tagged according to a service margin. In the congestion time, data packets tagged for a rich quality of service and are given preference treatment using queue priority at the cost of other data packets which do not fit the rich QOS representation.

In DiffServ, there is a special code point known as DiffServ Code Point (DSCP) which is situated in the IP data packet header and this DSCP is used to find out the PHB. Each PHB has their own DSCP field associated with it, which is used to find out the behaviour of the DiffServ in which the data packets to be received.

The DiffServ Model specifies an approach that offers a service better than Best-Effort and more scalable than IntServ. Traffic is classified into one of five forwarding classes at the edge of a DiffServ network. Forwarding classes are encoded in the Differentiated Services Codepoint (DSCP) field of each packet's IP header. The DiffServ model describes the utilization of the DSCP and the PHBs. The PHBs gives the forwarding behavior of a DS-compliant node.

Differentiated Services (DiffServ) Model Architecture:

Fig: Basic DiffServ Model Architecture

The above figures show the basic DiffServ architecture. In the DiffServ model, Traffic is classified into one of five forwarding classes at the edge of a DiffServ network, forwarding classes are encoded in the Differentiated Services Codepoint (DSCP) field of each packet's IP header. The IP header of this DSCP uses 6 bits of the header i.e the DSCP field sets 64 different values for each header of the data packet field. Here the DiffServ router applies pre provisioned Per-Hop Behaviour (PHB's) to the data packets according to the encoded forwarding classes. Here a Per-Hop Behaviour is a way of forwarding a particular flow of traffic marked with a particular DSCP.

In the figure 2, the DS-domain is composed of DS ingress nodes, DS interior nodes and DS egress nodes. An ingress or egress node can be a DS boundary line node, which combines the two DS domains. Typically, all DS ingress and egress nodes can be distinguished as an edge nodes, as these nodes acts as an intermediate point between the DS-domain and the non-DS-aware network.

Basically, the DS edge node executes the traffic conditioning. A traffic conditioner basically divides the incoming data packets into pre-defined combinations, meters the data packets so as to identify and obliges to traffic parameters (and identifies whether the data packet is in the range, or out of range), it then marks them properly by writing or re-writing the DSCP field, and later it shapes (buffers to meet the considered flow rate) or drops the data packet in case there occurs congestion. Figure below illustrates the traffic conditioner at the boundary of a DS-domain. A DS Internal node forces the befitted PHB by engaging policy or shaping techniques, and sometimes re-marking out of range data packets, by looking on the policy or the SLA.

In the above DiffServ Traffic Conditioner Block (TCB),

Classifier: Chooses a data packet in a traffic group based on the limit of some part of the data packet header.

Meter: examines deference to traffic arguments (ie: token bucket) and passes the results to the marker and shaper/dropper to enable the action for in/out-of-range packets.

Marker: Writes or re-writes the DSCP field value

Shaper: Holds some data packets which are non-resistant with the field range.