Looking to the Future

3G handset software must be capable of managing and multiplexing multiple per-user
traffic streams, qualifying the radio bandwidth and network bandwidth requirements
by taking into account information provided by, for example, the MPEG-4 encoder.
The content itself may be capable of determining its bandwidth requirements (declarative
content, or content that can declare its bandwidth quantity and quality needs).
The software then has to be capable of negotiating with the network, which implies
an intimate relationship with network-based admission control procedures (we cover
these in detail in Part IV of this book on network software). This would involve in the
future the qualification of least-cost routing opportunities—but this is unlikely to be
very appealing to the network operator.
MPEG-4, MPEG-7, and MPEG-21 provide a relatively stable and well-documented
standards platform on which software added value can be built. MPEG-7-based image
search engines, as one example, will potentially revolutionize image surveillance as an
added value opportunity.
Given that much of the future value generation will be subscriber-based (subscribergenerated
added value), handset software becomes progressively more important. The
ability to develop session persistency and session complexity is a particularly important
prerequisite, as is the ability to manage and multiplex highly asynchronous traffic
(bursty bandwidth), including buffer management.
It seems to be generally assumed that there will be a multiplicity of hardware and
software form factor in the future. This is not a good idea. Hardware needs to talk to
hardware, and software needs to talk to software. What is needed is a de facto dominant
hardware and software form factor for the handset and a dominant network hardware
and software form factor. Ideally (from a technical perspective), this would all be supplied
by one vendor, but this might prove rather expensive. To use multiple vendors but
avoid device diversity is probably the best technical/commercial compromise.