Back to index
Cellular Packet Communications
David J. Goodman
One-line summary:
Third-generation cellular packet networks will have decentralized
handoff control, use microcells, and employ PRMA as a multiplexing
method.
Overview/Main Points
- Third-generation networks: cellular networks will evolve
towards low-power transmitters, microcells, decentralized call control
and call management, and frequent handoffs and call set-up control
messages.
- Architecture: A conventional MAN (wireless of wireline)
will link together the core components: base-station interface units
(BIUs), trunk interface units (TIUs), controller interface units
(CIUs), and wireless terminal interface units (WIUs). TIUs are the
gateway between the cellular MAN and the long-haul network. WIUs are
the interface between the user's terminal or voice information source
and the the terminal's radio transmitter. BIUs relay information
between the MAN and the terminal radios. CIUs provide centralized
network control such as virtual circuit setup and authentication, but
are not involved with handoff control.
- Packet Reservation Multiple Access: PRMA is a combination
of slotted ALOHA and TDMA. The bitstream consists of frames,
subdivided into slots. Terminals contend for available slots using
slotted ALOHA with self-imposed transmit probability p; if one
in successfully transmits in an available slot, it gains a reservation
on that slot. Reservations are revoked if a holder neglects to
transmit during a frame.
Relevance
The architectural predictions made by this paper are accurate, but
quite broad and generic. The PRMA description is more interesting.
A good point made is that handoff control needs to be decentralized.
Flaws
- The architecture described is generic and somewhat unsupported
within the paper - it is stated as though it is fact, not possibility.
- The paper focuses on voice transmission as the dominant
application for cellular packet communications - this is nearsighted,
IMHO.
- There is not enough detail about any of the components described
in the paper; it is very terse and high level.
Back to index