How Important is QoS (Quality of Service) to decide on a solution of bandwidth?

When going through the decision making process for a solution of bandwidth .... for example, T1 or DS3 banda .... QoS is and should be an important consideration.

What (and should) QOS mean to you?

Quality of Service (QoS) Definition:

In computer / telecommunications networks, a traffic engineering term. Refers to mechanisms for resource reservation control, not the quality of the service itself. Possibility of priority to the applications, users, data flows, or to ensure a flow of performance data.

But you knew / knows.

As an alternative to QoS is of high quality communication of the "best effort" network, providing an excess of capacity, so that it is sufficient that provide for peak traffic loads. But high QoS is often confused with high performance, or quality of service received (for example, high bit rate, low latency, low bit error probability, and other arcane measures).

From a user perspective, however, QoS telephony and streaming video, you must try to measure the quality that the user actually experiences, called "Quality of Experience" (or QoE), based on the user's perception or their level of satisfaction. Thus QoS must keep track subscriber satisfaction, which of course is subjective and varies on a number of factors at the end user.

And when combined with the experience, you get a sort of hybrid measure that can be called "Service Quality of Experience" (or QoSE), which seeks to measure the user experience with respect to quality issues, regardless of bit rate or other arcane measures. None of these matter if the user is not satisfied.

This is what QoS means to me.

I want to measure the performance of QoS provided to the end user. If we refer when we speak of QoE QoS, we can now apply a model based on experience (QoE) to the traditional network measures of quality of delivery service (QoS). Adding to the traditional measures QoE QoS allows us to reiterate my first line "QoE is designed to measure application performance experienced by end users."

Since QoS measures the performance of the application, the measurement of QoS depending on application (Voice, VoIP, Video, Data, etc. ..). I think as QoS provides a good indication of the performance of delivery of the request, the measure must be used hand in hand with a certain type of model based on experience, when available, as QoE.

I believe that QoS in the eye of the beholder. Depending on the demands, the network, the level of traffic (not sure about this), the phase of the moon. Still another measure of QoS is Mean Opinion Score or MOS used in an attempt to quantify subjective voice application space. Then there is the ITU-T Emodel ... (As was the phone call for you?)

In the academic world thousands of research papers have been written and discussed. In the field we saw a large number of IEEE on QoS that all are based on a premise: a software layer / CPU centric fix-it on the best-effort IP mechanism comprising one or another scheme of packet prioritization and QoS must. The industry calls this soft QoS.

None of these systems are suitable for the traffic that passes through the domains of multiple operators. MPLS over IP is a core technology, what we need for end-to-end QoS is an end-to-end. Hard QoS.

So as you begin a project with a clean slate for global QoS. My starter for ten (for the sake of avoiding unnecessary complexity) would allow to insist on only two classes of QoS: best-effort and the best quality. Best-effort needs no introduction - this is what we are stuck with the Internet today - no guarantees, you get what you get, give or take. Best-quality should, by definition, IP networks allow for all latency critical applications such as voice, videoconferencing, database replication gaming ... on a global scale.

And this is where the next killer apps just might come from if we are able to deploy two classes of QoS in a scalable globally, so self-management .......