Mesh Networking, MANET, Multihop Packet Radio… there are many names to this technology that originates from early DARPA project in the late 70’s, early 80’s. (Tobagi et al 1987). Many billions of dollar have been invested in several waves of research and development of mobile, multihop tactical networks. Of course there are many appealing properties – rapid deployment, infrastructureless systems, robust (no single point of failure) etc. However, research has repeatedly also shown severe drawbacks related to the scalability of these network – when they grow and the number of hops in a path grows, the per-node performance/capacity goes to zero (Frodigh 1991, Frodigh&Zander 1991, Kumar&Gupta 2000). In mobile networks you have the additional problem that you can never guarantee that the network remains connected.
So is multihop packet radio dead as now demands for data rates increase? Yes, and no. I think in the truly mobile applications like tactical communication, large single layer infratruceless MANETs are doomed. Instead some measures that keeps the hop counts at bay are required. Inserting numerous access points to a high capacity backbone (fixed or using some other spectrum), i.e. creating a mulitlayer or “infrastructure-assisted” network is certainly an interesting solution.
If the hop count can be kept low, and the topology controlled, we have a fixed/semi-fixed Mesh Network that can work reasonably well. An early pioneer was Apple co-founder Steve Allens Metricom and its Ricochet network launched in the Bay-area in the 1990’s with 51000 subscribers in some 15 larger cities in the US in 2001 when Metricom went bust.
Currently there are quite a few products using WiFi based solutions for Mesh Networks out there. Low cost equipment and targeting niche markets like rapid deployment networks at events, emergencies, surveillance as well as rural communities internet access where low cost rather than high data rates and mobility are key factors. Check out this Firetide slideshow describing Dallas PD surveillance
In high rate applications the multihopping “eats” capacity and cellular broadband solutions will prevail, although at significantly higher cost – at least with the current technology.