In this chapter we introduced the MobTorrent software package. MobTorrent is the first BitTorrent client for feature phones with limited resources. MobTorrent is a full-featured BitTorrent client, it supports both download and upload as well.
During this research we discovered limitations related to feature phones which show that peer-to-peer solutions have different requirements than simple applications.
Despite the limitations we were able to implement MobTorrent which is based on a general mobile BitTorrent engine that can adapt to the capabilities of the target device. We demonstrated the portability of this engine by applying it on Android platform as well.
Then we have proposed a model to estimate energy consumption of MobTor-rent. We have shown measurements related to MobTorrent performance, memory requirements and energy consumption. The measurements proved that the per-formance is convincing and MobTorrent is able to run on phones with limited resources and the resource requirement remains on a relatively constant level.
After that we have introduced a hybrid peer-to-peer based architecture for content distribution which supports mobile devices and decreases the load of the server. The solution can be used by different type of service providers and it is also applicable in social networks. The main advantage of the architecture is that it also considers mobile phones and provides backend functionalities in case of low
share ratio as well. For the content distribution a modified BitTorrent protocol is used.
The proposed architecture had practical results, it was implemented at Nokia Siemens Networks in the Swarm project. Swarm architecture requires less storage capacity, energy consumption and processing power on the service provider side comparing to a client-server solution, thus it can be provided even on lower prices.
Swarm has hybrid architecture with a central unit, however the content distri-bution itself goes via BitTorrent. This architecture is suitable for social networks, because the central elements of Swarm can be stored on the social network servers.
Furthermore mobile clients are ideal for handling presence functions. By extending Swarm with social networking capabilities we can have functionalities like:
• Share contents only to selected people.
• Provide higher bandwidth to my friends.
The main advantage of this solution is the cost efficiency for service providers, because the produced traffic distributes in the network and significant part of the infrastructure and operation cost is handled by the Swarm client applications.
Swarm implements several features to enhance the user experience on mobile de-vices when it comes to content sharing. Among those users can find features such as easy search, browsing directories and "one click" content publish directly from the built-in camera and gallery applications. Naturally, these user interface ele-ments can be implemented in a traditional client-server based solution too. From user experience point of view, Swarm made a special effort in hiding all the com-plexity that comes with the hybrid architecture and BitTorrent technology. In practice, this means that user does not realize the complexity of the underlying system, what is more she or he gets the impression that the service is implemented with the traditional client-server approach.
Figure 6.8. MobTorrent engine class diagram
Chapter 7
Practical Application of the Scientific Results
The model of mobile related social networks and the similarity detecting and han-dling algorithm involved practical results. A mobile related social network was implemented by Nokia Siemens Networks in the Phonebookmark project. An at-tached statement to the thesis confirms my participation in Phonebookmark and other related research projects. The model of the network can also be used for other structural examination of mobile based social networks. The similarity de-tecting algorithm was used also in Phonebookmark, where more than 90% of the detected similarities were accepted by the users. Among others I have implemented the similarity handling in the system. The similarity handling algorithm and the learning methods with the similar term handling is applicable for other structure based similarity detection problems. The requirement of stability is also applicable in case of similar conditions.
The measured power law distribution of similarities and the analytical model for calculating the total number of identities in Chapter 5 is an important general measure of social networks. Power law distributions have infinite variance but in several cases the probability variable has a relevant upper bound. The proposed model, to calculate the variance in this case, is applicable in general probability theorems and practically in any case of power law distribution where the upper bound is considered reasonable.
We have implemented the MobTorrent software package presented in Chapter 6, which proves practically that feature phones are able to participate as full members in any BitTorrent network. MobTorrent was part of the peer-to-peer research project at Nokia Research Center. The 1.1 version was published as an open source project in April of 2009 and currently there are around 10.000 downloads.
Chapter 6 also examines a content sharing architecture that is applicable in social networks and decreases the load of the server. We took part in the im-plementation of this architecture at Nokia Research Center and Nokia Siemens Networks. Since mobile phones are key elements in mobile related social networks, in our research we focused on involving them in the content distribution as well.
This BitTorrent based hybrid solutions can be applied in several situations in social networks, e.g. distributing videos or whole photo albums between friends.
We would like to emphasize that, the practical results of this thesis, namely the similarity handling in Phonebookmark and the MobTorrent software package, were applied in practice and they are also unique proof of concepts in their fields.
Based on feedbacks, they operated well in real environment.