Creative New

Exploiting social and mobile ad hoc networking to achieve ubiquitous connectivity.

Via developer.symbian

Via brockport.edu

This essay examines a particular example of what is known as a ‘Mobile Ad hoc Network’ (MANET) involving smartphones. This MANET would address the following problem:

‘A million and a half people annually ride the New York City subway system, which comprises 468 stations to form the largest subway complex in the world. The diverse human conglomerate of cultures and lifestyles that inhabit NYC’s subway though, share at least one common trend: a lack of mobile coverage – a rather serious, both economical and social, issue. Thus, a daily influx of 5,042,263 potential customers is practically lost for mobile companies. Not to mention the irritating circumstance that in a world of otherwise global communications one is forced to spend significant periods of time without a connection signal.’

One response to this problem is to wire the subway stations, using traditional methods.  But apparently this project is expected to cost up to $200 million, and will take ten years to complete.  That’s why the idea of using a MANET becomes more attractive.  In a MANET, communications can reach phones that are outside the direct reach of the cell towers, because other nearby phones act as intermediaries.  In this way, phones that are still within cell tower coverage pass on relevant data, via a series of local hops, to phones that are further and further away from cell tower coverage.

This avoids the need for any additional wiring or new cell towers, but relies on smart software installed on mobile phones that will convert an extended group of nearby phones into a fast-changing ad hoc network, with communications packets being passed down this network to reach the intended recipient phone:

‘Hence, true to the MANET paradigm, the individual mobile device would act both as a receiver and dispatcher of signal… Social phenomena and behaviors can be used in favor of disseminating network coverage. Thus the signal coverage will depend on social communities.

‘Collaborations of mobile society and devices can deliver services to the individual subscribers comprising it. To paraphrase President John F. Kennedy, ask not only what your signal can do for you, but what you can do for your signal.’

MANETs are interesting for another reason: they are likely to play larger roles in the future, with 4G networks in which connections are free to change dynamically between different bearers:

‘Within the context of 4G networks the connection between people and signal will be symbiotic; people will utilize the signal they get to fulfil their communication needs, and the signal will utilize the people to propagate itself across mobile ad hoc networks.’

I – Stable infrastructure network is provided by mobile phone operators.

II –Station gateway consists of nodes (in our case people) flowing in and out of the train station. The number of nodes in the gateways is constant and their density higher. Hence transmission range need not be large.

Handoff between MANET and mobile operator’s network happens at the left exit of gateway.

III – Station hall; MANET nodes here are people waiting for a train (arrivals) or ones about to leave the station (departures). People are more sparsely distributed and transmission ranges are larger.

IV – Train – moves nodes in and out of the network. Train schedules are predetermined and hence the approximate number of people in the station is known and uniform for predictable time intervals.

The essay reviews both technical and social factors which will impact the effectiveness of this particular MANET:

• Because there is a limit to how far a signal can hop from one phone to another, the network relies on a certain minimum density of users with suitably equipped phones (and for this reason, the MANET only works within the stations – not on the trains while they are between stations). Indeed, at given time instances, usually after the departure of a train or in late night hours, there may be too few people in the subway station, implying no formation of a MANET.
• Users will have to accept that power will be drawn from batteries in their own phones, as their phones convey messages between third parties, even though they don’t themselves benefit from that particular communication.
• Users are more likely to allow their phones to participate in this kind of network, so long as the MANET avoids draining batteries that are already low (nodes at critical energy level can be excluded from network’s topology altogether) and providing the network notices and discriminates against ‘free riders’ (thereby encouraging and rewarding users who take a pro-community stance).
• Cooperation of users in being willing to form a MANET depends on a higher order of trust between the people and devices constituting it. Trust is built upon security attributes such as confidentiality and service availability when needed in the network.
• Routing algorithms exist which can provide acceptable quality VoIP propagation in a wide variety of circumstances.

As the essay states,

‘In the perspective of a subway station, an ad hoc network can cover 1000 people and support about 20 calls in any instance. At present, a scenario of more than 20 simultaneous calls would cause an increase in end-to-end delay above 200 ms, which is sufficient to degrade a VoIP service…

‘But even better suited network schemes for real time applications exist. Some research from 2007 investigated properties of the Warning Energy Aware Clusterhead (WEAC) protocol specifically in the context of VoIP in MANET. In short, WEAC treats a mobile ad hoc network as a collection of clusters. Each cluster has a head node storing a membership list describing which other nodes are cluster members. Thus, to reach any node one only needs to discover a cluster head whose membership list contains a path to the destination. This reduces network request packet floods and thereby increases the ratio of useful data versus overhead. In addition to increasing the ratio of data packet transmitted to the control packet overhead, WEAC tends to minimize end-to-end path lengths and delays. Due to this favorable combination and coupled with G.729 audio codec, it manages to sustain a VoIP network of 100 nodes, all requesting service simultaneously. That is, WEAC supports 100 simultaneous calls…

‘… by using mobile ad hoc networking and implementing the appropriate routing protocols at the application layer of smartphones, one can obtain good network service on demand. Signal coverage, support of enough simultaneous calls and short service initiation delay are all within the grasp of current MANET technology. This is especially true when heterogeneous internetworks comprising relatively small mobile ad hoc and highly infrastructure-based networks are considered… Solutions would be facilitated by anticipated advances such as load balancing schemes and packet congestion control algorithms in MANET. The latter will increase number of sustainable source nodes (hence simultaneous phone conversations) and minimize end-to-end delay of packets: a leap towards better quality and reliability of real time, fault tolerant applications such as VoIP and even video communication.’

The main remaining concern is a social one, rather than something technical:

‘…the more people (up to a limit) within a MANET network, the better, that is, subway station coverage can be better if more people participate in the network; furthermore, the power consumption per device would be less, resulting in increased links stability and network lifetime. Therefore, the public interest would require more people willing to join the station’s MANET and share their technological resources more frequently to improve the network’s quality.

‘Chances are that the ability to remain connected or communicate in return for network participation may not be enough to attract as many people as needed. Different incentives and perks to join a smartphone MANET should be presented. For example, mobile networks operators can track and reimburse power consumed in transmitting alien packets with additional call minutes. Internet activity within MANETs can also be available for free to encourage and increase the influx of people opting into the networks. That is, individuals in a MANET-covered subway station can browse the web at no charge – a stimulus to wait a few additional minutes for the next train, or at least make their stay at the station more pleasant one!’

The essay has the following conclusion:

‘Exploiting a heterogeneous combination of large infrastructured network coupled with small MANET networks is feasible and beneficial both for communities and corporations. Forming MANET on demand, in areas of weak or non-existent signal, can tap potential capital currently outside the reach of mobile phone operators. The numerous people flowing through subway systems or buildings with no signal coverage would be able to connect to the, no longer, outside world. Novel competitive marketing models would be required and available to companies exploring an entirely new niche of communications.

‘Under 4G requirements of anytime-anywhere connectivity, research and implementation of MANET features (routing, security, energy awareness) now will provide a crucial competitive edge in 2012. Furthermore, as shown, even existing mobile ad hoc technologies are already sufficient for the establishment of comparatively small but well functioning pilot networks.

‘“Connecting people” would truly describe, both literally and semantically, next generation smartphones.’