Elsevier

Ad Hoc Networks

Volume 52, 1 December 2016, Pages 173-182
Ad Hoc Networks

Performance evaluation of DTN protocols to deliver sms in dense mobile network : Empirical proofs

https://doi.org/10.1016/j.adhoc.2016.07.006Get rights and content

Abstract

In urban areas, the population density is still growing (the population density starts exceeding 20.000 inhabitants per km2), and so, the density of mobile users becomes very important. People are moving from home to work, from work to active places. One can take benefit of the mobility and the density to justify DTN (Delay Tolerant Network) approach protocol to convey sms (or alternative messaging services) traffic. Indeed, the mobility of users, especially during the day, create an ad hoc mobile network where the nodes are the smartphones hold by mobile clients. In this paper, their performance evaluations are based on a measurement and analysis of sms traces coming from a nationwide cellular telecommunication operator during a two month period, we propose several DTN like basic network protocols for delivering sms. We perform a temporal and spatial analysis of the Mexico City cellular network considering geolocalized sms to characterize the traffic. Such key characterization allows us to answer the question: is it possible to transmit sms using phones as relay in a large city such as Mexico City? We define four network protocols to transmit sms from a source to a destination. We study a mobile dataset including 8 Million users living in Mexico city. This gives us a precise estimation of the average transmission time and the global performance of our approach. Our analysis shows that after 30 min, half of the sms are delivered successfully to destination. On the contrary to the cellular networks, we explain how much the potentiality of the mobile users network can take benefit from complementary properties such as the locality of sms, the density of phones in Mexico City and the mobility of phone users. Moreover, we show that in a realistic scenario, our approach induces reasonable storage cost.

Introduction

The need of communicating in a dense city is always increasing. Every day, millions of sms are sent in a large city like Mexico City. Traditional sms is challenged by alternative messaging services such as Facebook Messenger, WhatsApp [1], Tango, Skype and Viber by using data connection and/or wireless hot spot [2]. One can also use P2P applications to enable the connection of smartphones via Bluetooth or Wi-Fi without an Internet connection. In this case, people have to constitute a dense network to insure the connectivity. Though it was not designed in purpose, FireChat was used as a communication tool in some civil protests. Nevertheless, sms is still a growing market and remains a very popular service over cellular networks. In our dataset, 82% of mobile users are sending sms. The sms is well-known and well-used in both developed and developing countries.

During rush hours, sms traffic may consume a non negligible part of the backbone network capacity, and sometimes saturates it. This saturation may come from the sms architecture itself which is totally centralized. Every sms is deliver to a unique SMS Center (SMSC) which acts as a centralized, store-and-forward server that is responsible for accepting, storing, retrieving subscriber information, and forwarding sms to the intended destination of the sms.

It is becoming a great challenge to increase the amount of traffic delivered to the users while keeping the infrastructure stable (i.e., same number of relays and backbone capacity). Apart from the typical rush hours, the mobile network can be globally challenged during special events such as natural disasters or locally saturated during concerts, conferences, riots or sport matches. Mobile networks are dimensioned to sustain the load 99% of the time, but for those specific events, the activity can be incredibly higher than the expected traffic during rush hours. In [3], authors revealed that voice calls and sms are still in use in these large scale events, despite frequent reports from users about the data network unavailability. It can thus be interesting to propose and test new protocols, less dependent on the cellular infrastructure and/or the backbone, that could carry a part of the traffic load.

In this study, we evaluate the benefits and the performances of a delay tolerant network (DTN) approach to transmit sms and more generally data from a source to a destination. Instead of using classical routing, we use relays close to the source and phone users that are connected to those local relays to reach the destination. The advantage of our approach is that we do not perform a routing algorithm, we do not need global knowledge, neither its associated mechanisms such as neighbor discovery, exchange of control messages, etc. and do not need to know where the destination is. Moreover, as we only use local relays that are close to the source, the bandwidth cost of a sms is smaller and the backbone infrastructure of the operator is not used. On top of that, our protocol works best when the capacity of the network is challenged during rush hours as the density of phones and the mobility of users are even higher.

It is important to notice that we clearly do not target an implementation of our approach in existing 3GPP standards or existing cellular networks protocols. Our goal is to demonstrate that a DTN approach could be validated, feasible and helpful. We replay traces containing millions of real sms traffic and it shows the reliability and the gain of our proposals. Note that such approach could be used in future cellular standards, and also be used as a key enabler for wireless P2P or future community applications.

The contributions of this paper are two-fold. We perform an analysis in time and space of sms traffic. It is based on a significant database describing sms sent during two months in Mexico city [4]. It represents 90 millions of phone users all over Mexico. The temporal analysis includes an event detection to check overloading periods and the evolution of the sms activity at different time scales. The spatial analysis is based on the distance between source and destination in kilometers and number of hops. The number of hops is computed according to the Voronoï diagram of the base stations. Thanks to the traffic characterization and more precisely its regularity and locality, the second contribution is the proposal of four protocols that aim to carry sms traffic, and to relieve the infrastructure network in terms of load. We use the same real sms traces to evaluate the efficiency of these less centralized protocols. Through the replay of these sms, we show that with a very simple and not optimized algorithm, it is possible to have a delivery ratio higher than 50%. It is obviously not perfect, but we consider that it is enough to prove the benefit of our approach. It is also worth noting that this deliver rate is underestimation as our date set is partial.

The paper is organized as follows. We first reference studies that are linked to our experiments and topic in Section 2. In Sections 3 and 4, we present our large trace that contains sms and localized calls and we perform a spatial and temporal analysis to point out some pertinent characteristics of the cellular networks. Two basic DTN protocols are presented in Section 5. In Section 6, we show the performance evaluation of these two protocols in terms of transmission success and delay, and point out how much higher activity and mobility can increase the success rate of our protocols. We deeper analyze the complementarity of our protocols by mixing and extending them in Section 7. In Section 8, we discuss the experimental choices and estimate the storage cost of protocols. We conclude in Section 9.

Section snippets

Related work

A study [5] of nationwide cellular network analyzes the temporal activity and peaks. It seems to be a great advantage to understand better the activity variations to limit the congestions. The overload during rush hours or specific events that induces high traffic is a great problem and diverse solutions have been proposed [6], [7]. It is possible to limit the size of the message, to include in mobile phone a software that reacts to congestions, or even to try to recognize and predict

Data source

We used traces extracted from mobile phones [4] representing 92 millions of clients in Mexico and 8 millions in Mexico City during a 2-month period: from March to April 2014. The city is covered by 775 base stations that are part of the telecommunication network of a cellular operator. This anonymized dataset contains 70 millions of sms and 170 millions of phone calls all over Mexico. Some calls are localized, i.e., we know the base station of the source or destination. From these mobile calls,

Data analysis

In this section, we make an analysis of the localized sms according to three parameters: the activity (number of sms sent), the distance in kilometers and the distance in hops (distance in hops from the source to the destination of the sms in the cellular graph, defined by Voronoï diagram, from source base station and destination base station). These parameters are computed according to the time (per hour) and the space (per cell). The temporal analysis consists in studying the number of sms

Protocols

We describe two protocols to deliver the sms in another way than the protocol used in the cellular network. Our approach relies on the density and mobility of phone users combined with the locality of sms. The protocols use the base stations that are close to the base station attached to the source and users that are connected to it. They are not totally decentralized but more centralized than the classic cellular protocol that routes sms in the infrastructure network. Fig. 5 depicts these

Results

We validate our concept by evaluating the feasibility of our approach. It is evaluated through the percentage of sms that are properly/successfully delivered. We are not competing with classical cellular routing but we show that our protocols can be an alternative for some applications to complement it by relieving the infrastructure network of a part of its load during some challenging time, for instance. For PP, we choose to limit the delay to 30 min which seems to be a reasonable upper limit

Mixed protocols

In Sections 5 and 6, we define and analyze two basic protocols. It allowed us to understand the impact of many parameters such as the distance of the sms, the number of packers around the source and their mobility for these basic mechanics. In this section, two main variations of these two basic protocols are defined and experimented. If some sms were delivered with PP and not LPk, it means that two protocols are complementary. Here, by defining mixing protocols, we quantify and analyze this

Discussions

Impact of choices. The spread of the sms relies on the geolocalization of users. In previous sections, each ego is geolocalized from his phone call activity. Thus, a 30 min is set to interpolate localization from call to sms. Yet, one can wonder how this choice is impacting our results. In practice, as one can notice on Fig. 12b, when the interpolation is precised and so the threshold to interpolate the geolocalization is small, the amount of geolocalized sms is decreasing. Contrarily, if the

Conclusions

In this study, we have first proposed two basic protocols based on DTN principles to evaluate the mobile users ad hoc network. They have been evaluated through an original method as we had the opportunity to replay a large trace of geolocalized sms in Mexico City. This evaluation has shown that the density of the network, users mobility and locality of sms play a major role concerning the efficiency of the protocols and so could unload efficiently the backbone network of the operator. Even if

References (17)

  • B.S. Fjeldsoe et al.

    Behavior change interventions delivered by mobile telephone short-message service

    Am. J. Preventive Med.

    (2009)
  • B.C. Csáji et al.

    Exploring the mobility of mobile phone users

    Physica A

    (2013)
  • K. Church et al.

    What’s up with whatsapp?: comparing mobile instant messaging behaviors with traditional sms

    Proceedings of the 15th International Conference on Human-computer Interaction with Mobile Devices and Services

    (2013)
  • C.K. Toh

    Ad Hoc Mobile Wireless Networks: Protocols and Systems

    (2001)
  • H.N. Curto et al.

    Using sms to transfer small data packets during periods of high workload on mobile data networks

    XXXIII Simpósio Brasileiro de Redes de Computadores e Sistemas Distribuídos

    (2015)
  • C. Sarraute et al.

    A study of age and gender seen through mobile phone usage patterns in mexico

    Advances in Social Networks Analysis and Mining (ASONAM), 2014 IEEE/ACM International Conference on

    (2014)
  • P. Zerfos et al.

    A study of the short message service of a nationwide cellular network

    Proceedings of the 6th ACM SIGCOMM Conference on Internet Measurement

    (2006)
  • E. Oliver

    Characterizing the transport behaviour of the short message service

    Proceedings of the 8th International Conference on Mobile Systems, Applications, and Services

    (2010)
There are more references available in the full text version of this article.

Cited by (0)

View full text