The survey also revealed a possible cause for these worries - over a third (39%) of companies do not profile the performance of applications such as voice, prior to implementation, so cannot anticipate the effect that it will have on the network, and thus be reassured about quality and reliability.
Michael Allen, global director of performance solutions, Compuware said, "People might be surprised that quality and reliability is still such a big concern for so many, when VoIP is high on the agenda for most companies. However, when you take a closer look, this is not entirely unexpected, as many companies are not taking the appropriate steps prior to VoIP implementation. It's like driving a new car before the manufacturer has checked that all the components are fitted together properly and the safety checks are completed. Businesses must profile performance prior to deployment otherwise they have no way of knowing whether there will be negative side effects due to factors such as network design or application conflicts. With voice being such a high profile application organisations can't afford to take this risk."
Furthermore, companies are also failing to correctly implement strategies for monitoring the quality of VoIP calls once such a system is in place. The survey revealed that only 8% of organisations manage and monitor, or are prepared to manage and monitor calls at an individual level. It is only by monitoring at the individual level that organisations can gain the necessary visibility into the network and the VoIP calls running on it to be assured of the required level of quality and reliability. 37% of IT managers rely on retrospective user feedback and consequently can only resolve problems after they have occurred. 29% of companies use occasional synthetic testing, which simultaneously adds to the load on the network and also means that problems that occur at unusual times of day or night, when IT technicians are not running such tests, cannot be tracked.
The survey also revealed that 72% of IT departments are taking a "sledgehammer" approach to network monitoring: only looking at overall network utilisation rather than examining the individual behaviour and usage of each application. This means that the quality of service of VoIP may well suffer, even if the organisation is using Class of Service on a MPLS network, because IT departments will not have the necessary insight into application performance. For example, if there is a large amount of lag on a VoIP call, which may be due to a problem at one of the network nodes, this may have gone unnoticed because VoIP does not necessarily generate a large volume of traffic, even though call quality is suffering.
This approach is also reflected in IT managers' reactions to problems - 46% of respondents admitted to simply throwing more bandwidth at network utilisation problems rather than probing into any further detail to get to the heart of the problem.
"With real time applications such as VoIP coming onto the network, businesses need to change the way they are monitoring and managing network and application performance. They can't afford to be reactive any more and simply monitor utilisation. The problem is that many organisations are still using these simplistic techniques and hence they are quite right to be worried about quality and reliability," continued Allen. "If they used more appropriate and granular monitoring techniques they would no longer have to worry about facing the wrath of business users that are experiencing problems when making telephone calls."
The survey was of 300 IT directors from large enterprises across Europe, and was carried out by Omniboss, the enterprise division of market research company Vanson Bourne