In this document, I have explained how VoIP in virtual environments can be achieved and what factors play an important role. Additionally, I described my own work on this topic. Now, what are the conclusions that can be drawn from all this?
When we consider VoIP applications in general, they will probably become more widely used as time evolves. Currently, the main problem for such applications is the lack of QoS guarantees. When QoS supporting protocols like RSVP are used on a larger scale, this will certainly make VoIP more popular since people can then communicate with the quality that they desire. On LANs, where there is normally plenty of bandwidth, VoIP applications can already be used with little or no problems. However, on a larger scale, like the Internet, such QoS providing protocols will be necessary to make VoIP applications perform adequately.
In the computer industry, everything evolves very rapidly. Therefore, I assume that the available bandwidth on networks will keep getting larger. This will also be helpful for the spreading of VoIP applications. When the available capacity is sufficiently large, even high quality sound will be possible, which will certainly be a stimulus for the use of VoIP programs. Furthermore, since compression techniques are still improving, such high-quality communications will be available even sooner.
Standards like H.323 and SIP make interoperability between applications of different developers possible. This way, people can choose from a variety of VoIP applications and use the ones they like the most. In turn, this will stimulate the use of VoIP. Also, since more applications will be developed, the possibilities of these applications will keep growing and improving.
The use of VoIP as a telephony alternative can save quite some costs. Since voice and data traffic can be integrated, the necessary infrastructure to provide both services is reduced. This integration will also make better use of the available bandwidth: first of all, bandwidth on a network is rarely entirely filled with data traffic. Second, classic telephone calls waste a lot of bandwidth since this bandwidth is reserved for the two parties even when someone is not speaking. Making long distance telephone calls over the Internet or another IP network will also be cheaper than using the telephone network for this purpose.
For VoIP in networked virtual environments, there are certainly a lot of possible applications. Currently, there are not many programs which provide this functionality, but I definitely believe that this will change. When better quality can be guaranteed, such applications can be an attractive alternative to chat environments like IRC. As CPU power keeps growing and more dedicated hardware becomes available, better sound localisation will improve the realism of the virtual environment. This in turn will make VoIP in virtual environments even more attractive.
Some observations can also be made about my own development. The RTP library which I developed seems to be very useful in several applications. This is indicated by the many positive reactions which I received. The library proved to be both fast and simple to use. Because of its clear structure, new features can easily be added to it.
The VoIP framework also proved to be very useful. It easily allowed me to test several components and made the VoIP part of an application portable to several platforms. This is because normally only the digitisation and reconstruction components need to be rewritten for a new platform. The framework also allowed the rather fast creation of several VoIP test programs.
Among these programs, the Internet telephony application and the 3D environment are quite useful. When enough bandwidth is available, they allow good quality conversations. These programs also made me realise that VoIP has a lot of potential for future development.