Determining Basic Network Architecture

I have already mentioned the basic network architectures: point-to-point, point-to-consecutivepoint,
point-to-multipoint, and mesh. While hybrids consisting of one or more topologies are
certainly possible, generally a single architecture is chosen for the entire locale to be served,
and the choice reflects not just technical considerations but the nature and composition of the
local market.
The band or bands in which the network operator is transmitting limits the choice of network
architecture. For example, transmissions in the highest band now commonly in use,
59GHz to 64GHz, can never be point-to-multipoint because spreading the beam to accommodate
such an architecture would result in unacceptably short range and also a failure to reach
many potential customers because of strict line-of-sight considerations. Similarly, a mesh
architecture, while technically feasible in the millimeter wave frequencies, is apt to be quite
expensive because each node essentially becomes a base station (one U.K. manufacturer,
Radiant Networks, now out of business, did make a 28GHz mesh product, however). Finally,
point-to-point connections are rarely encountered below 3GHz though they certainly are possible,
because the bandwidth lends itself to point-to-multipoint deployments and the profit
potential of the latter is generally much greater than that associated with the sale of the total
bandwidth to a single customer via a narrow-beam bridge link.