Phone system: Why quality of service matters - and how to improve it

At a time when so many everyday business operations - and more often than not, simultaneous operations - rely on an internet connection to be performed, quality of service is a million-pound issue. Maybe that important call for a contract worth hundred thousands of pounds will be interrupted at the worst moment. Maybe the CEO at the end of the line will misunderstand what you tell him because your voice has been garbled. Quality of service, sometimes shortened as QoS, is an essential indicator in VoIP driven office environments.

How quality of service is measured

Quality of service can be accurately measured by applying a few policies. Furthermore, signs of a poor QoS are unfortunately easy to spot.

Basic quality of service policies

Every task simultaneously performed using the common bandwidth should have sufficient resources to be done without suffering from data loss. However, because data networks were not originally built to channel voice conversations, and because some tasks require more bandwidth than others, it is essential that a priority policy should be manually programmed so that bandwidth-intensive tasks are automatically prioritised. 

Signs it can be improved

If such policies are not set, signs of a bad QoS can quickly become annoying. Here’s a typical example of an unfortunate escalation, and what causes it:

- latency, which occurs when data packets receipt is delayed for a few milliseconds;

- further latency, when this delay undergoes variations;

- data loss or packet loss, when the line is so congested that the packet is ignored. Quality is clearly damaged, and something must be done to bring it back to acceptable standards.

How it can be improved

Quality of service can be improved through complex technical enhancements, as well as with daily procedures.

Technical enhancements

Technical enhancements which can be undertaken to increase quality of service include:

- traffic shaping, through packet prioritisation, queuing at congestion points, or application classification;

- packets going through a gateway host can be expedited depending on preset reservation criteria;

- dedicated-connection switching technologies such as asynchronous transfer mode.

- improve the “contention” ratio, the number of users per unit of data capacity. It should be as close to 1:1 as possible.

Daily “hygiene”

But there’s no need to be an IT specialist to perform daily tasks which will ensure the system will keep delivering a satisfying quality. Routine tasks can include:

- checking the router: routers must be QoS-enabled to allow quality of service settings and monitoring. But also, routers age, much faster than their users believe. Connectivity performance should be tested regularly.

- restrain high-bandwidth operations like streaming videos, video conferences or other calls during highly important calls.