Quality of Service, what it is and how it works

[Lenovo]

In the telephony field defines the quality level of the service provided by the manager; in the networking field, the priority with which applications can access network resources

When we talk about Quality of Service (sometimes abbreviated as QoS, "Quality of service" in Italian) we refer to a series of indicators that help to establish the quality and general performance of a telephone network or a computer network (as can be a LAN or Internet). In particular, this evaluation is done from the "user point of view".

For some types of connections, QoS is a determining factor. Some particularly expensive web applications in terms of bandwidth - VoIP, audio-video streaming, video calls, telepresence and more - require the Tlc operator to guarantee precise quality standards that guarantee its execution without interruptions or problems of any kind.

The definition
In the telephony sector (and more generally in communication based on circuit switching), the Quality of Service is defined by ITU in 1994. In this document, subsequently revised and refined, the quality of service is based on indicators such as availability the service (ie the presence or absence of a cabinet providing the fixed telephony service), the noise level on the circuit (ie the interference on the transmission signal causing disservices), the sound level and the probability of finding a free line for start the communication.

In the field of computer networks (and more generally in communication based on packet switching), the term refers to control systems for the intelligent management of bandwidth in a given network. Specifically, the QoS determines the different priority levels with which different applications, users or data streams will be able to access available network resources: for some programs or protocols, for example, certain bit rate levels may be guaranteed (transmission frequency bit), delay (delay in bit transmission), jitter and probability of errors in data transmission. The possibility of establishing different priorities is essential especially if the network resources are scarce or in any case insufficient to guarantee the transmission of all the data flows that need it; in cases like these, to meet the minimum QoS limits, "band hungry" applications such as audio-video streaming, IPTV and VoIP will necessarily have to have a preferential channel with respect to other applications. A network or a data transport protocol that supports the Quality of Service will be able to stipulate a specific traffic contract with software and applications requiring bit rate, delay and jitter guarantees in order to function correctly.

 

 

 

Traffic quality
Some elements more than others help to define the quality of service in the field of computer networks. In a packet network the data usually considered are:

Delivery out of order (out-of-order). In packet-switched networks, it can happen that packets do not arrive at their destination in the same order they started. This happens because the data can follow different paths and use different times to "complete the path". In order for the destination node to correctly interpret the received data, it is necessary that the protocol used in the transmission also deals with the "reconstruction" of the original message, causing a general delay in the communication between the two nodes. From a quantitative point of view this indicator consists in an assessment of the probability that a data package will come out of order.

Delay (delay). The transit of a packet within a network usually takes a few milliseconds (abbreviated as ms, equivalent to one thousandth of a second). A problem is considered to be the case in which the communication delay exceeds ten ms. For this purpose two indicators are generally considered: the "average delay", which indicates the average time for data packets to cross the network, and "percentile delay", which indicates the percentage of data packets reaching the destination within a certain time maximum (which, in general, is decided according to the type of application involved).

 

Packet loss. It is not always said that all packages find their way. In fact, for several drawbacks, one or more packets may be lost, causing a general delay in communication. In these cases, the communication protocol requests the resending of the packets not received to the sending node, so as to be able to complete the message and interpret it correctly. Also in this case the factor that defines the QoS indicates the maximum percentage of packets lost compared to the total number of packets sent.

Transmission error. For signal interference or problems in the communication medium, packets can arrive at their destination damaged. The recipient node, by means of the communicative protocol used, will have to ask that they be sent again or, as it happens more and more often, try to reconstruct the message itself based on the already received packets, interpolating the missing information. Again, the value that defines this factor represents the maximum percentage of damaged packets compared to the total number of sent

Throughput. The QoS depends on the bandwidth actually available to the user. The throughput is precisely this factor: the amount of bandwidth that can actually be used by the user for data transmission. The maximum available bandwidth is set upstream by the ISP (Internet service provider) but can change, moment by moment, depending on the network resources available at the general level (for example, it can drop a lot in case of high network congestion), consequently the throughput can completely saturate the transmission capacity available in that data moment. Therefore, the higher the throughput guaranteed by the ISP, the more the associated QoS will be high

Applications
Web applications are divided between elastic and inelastic. The former have no particular needs and can also work on low-performance networks. The latter, on the other hand, have stricter operating conditions and require guarantees on the available band, on the average (or percentile) delay and on the percentage of error admitted by the network. As a consequence, the latter require a QoS that guarantees its correct functioning.

Among these programs and protocols we find live streaming of multimedia content, VoIP telephony, a critical application for the safety of people (such as remote surgery), video calls, industrial control systems and online video games in real time (such as Enemy Territory ).