Pubblicazioni Aperte DIgitali Sapienza > Informatica > INFORMATICA >

Please use this identifier to cite or link to this item:

Tutor: Mei, Alessandro
Mancini, Luigi V
Piperno, Adolfo
Issue Date: 18-Feb-2013
Abstract: In my first year of my Computer Science degree, if somebody had told me that the few years ahead of me could have been the last ones of the so-called PC-era, I would have hardly believed him. Sure, I could imagine computers becoming smaller, faster and cheaper, but I could have never imagined that in such a short time the focus of the market would have so dramatically shifted from PCs to personal devices. Today, smartphones and tablets have become our inseparable companions, changing for the better numerous aspects of our daily life. The way we plan our days, we communicate with people, we listen to music, we search for information, we take pictures, we spend our free time and the way we note our ideas has been totally revolutionized thanks to them. At the same time, thanks also to the rapid growth of the Cloud Computing based services, most of our data and of the Internet services that we use every day are just a login-distance away from any device connected to the Internet that we can find around us. We can edit our documents, look our and our friends’ pictures and videos, share our thoughts, access our bank account, pay our taxes using a familiar interface independently from where we are. What is the most fascinating thing is that all these new possibilities are not anymore at the hand of technically-savvy geeks only, but they are available to newer and older generations alike thanks to the efforts that recently have been put into building user interfaces that feel more natural and intuitive even to totally unexperienced users. Despite of that, we are still far from an ideal world. Service providers, software engineers, hardware manufacturers and security experts are having a hard time in trying to satisfy the always growing expectations of a number of users that is steadily increasing every day. People are always longing for faster mobile connectivity at lower prices, for longer lasting batteries and for more powerful devices. On top of that, users are more and more exposed to new security threats, either because they tend to ignore even the most basic security-practices, or because virus writers have found new ways to exploit the now world-sized market of mobile devices. For instance, more people accessing the Internet from their mobile devices forces the existing network infrastructure to be continuously updated in order to cope with the constantly increase in data consumption. As a consequence, AT&T’s subscribers in the United States were getting extremely slow or no service at all because of the mobile network straining to meet iPhone users’ demand [5]. The company switched from unlimited traffic plans to tiered pricing for mobile data users in summer 2010. Similarly, Dutch T-Mobile’s infrastructure has not been able to cope with intense data traffic, thus forcing the company to issue refunds for affected users [6]. Another important aspect is that of mobile security. Around a billion of people today have their personal information on Facebook and half of them access Facebook from their mobile phone [7]; the size of the online-banking in America has almost doubled since 2004, with 16% of the American mobile users conducting financial-related activities from their mobile device [8]; on 2010, customers spent one billion of dollars buying products on Amazon via mobile devices [9]. These numbers give an idea of the amount of people that today could find themselves in trouble by not giving enough care into protecting their mobile device from unauthorized access. A distracted user who loses his phone, or just forgets it in a public place, even if for a short time only, could allow someone else to get unrestrained access to his online identity. By copying the contents of the phone, including passwords and access keys, an attacker could steal money from the user’s bank account, read the user’s emails, steal the user’s personal files stored on the cloud, use the user’s personal information to conduct scams, frauds, and other crimes using his name and so on. But identity theft is not the only security problem affecting mobile users. Between 2011 and 2012, the number of unique viruses and malwares targeting mobile devices has increased more than six times, according to a recent report [10]. Typically, these try to get installed in the target device by convincing the user to download an infected app, or by making them follow a link to a malicious web site. The problems just exposed are major issues affecting user’s experience nowadays. We believe that finding effective, yet simple and widely adoptable solutions may require a new point of view, a shift in the way these problems are tackled. For these reasons, we evaluated the possibility of using a hybrid approach, that is, one where different technologies are brought together to create new, previously unexplored solutions. We started by considering the issues affecting the mobile network infrastructure. While it is true that the usage of mobile connectivity has significantly increased over the past few years, it is also true that socially close users tend to be interested in the same content, like, the same Youtube videos, the same application updates, the same news and so on. By knowing that, operators, instead of spending billions [11] to update their mobile network, could try an orthogonal approach and leverage an ad-hoc wireless network between the mobile devices, referred to in literature as Pocket Switched Networks [12]. Indeed, most of the smartphones on the market today are equipped with short-ranged radio interfaces (i.e., Bluetooth, WiFi) that allow them to exchange data whenever they are close enough to each other. Popular data could be then stored and transferred directly between devices in the same social context in an ad-hoc fashion instead of being downloaded multiple times from the mobile network. We therefore studied the possibility of channeling traffic to a few, socially important users in the network called VIP delegates, that can help distributing contents to the rest of the network. We evaluated VIP selection strategies that are based on the properties of the social network between mobile devices users. In Chapter 2, through extensive evaluations with real and synthetic traces, we show the effectiveness of VIP delegation both in terms of coverage and required number of VIPs – down to 7% in average of VIPs are needed in campus-like scenarios to offload about 90% of the traffic. These results have also been presented in [1]. Next we moved to the security issues. On of the highest threats to the security of mobile users is that of an identity theft performed using the data stored on the device. The problem highlighted by this kind of attacks is that the most commonly used authentication mechanisms completely fail to distinguish the honest user from somebody who just happens to know the user’s login credentials or private keys. To be resistant to identity theft attacks, an authentication mechanism should, instead, be built to leverage some intrinsic and difficult to replicate characteristic of each user. We proposed the Personal Marks and Community Certificates systems with this aim in mind. They constitute an authentication mechanism that uses the social context sensed by the smartphone by means of Bluetooth or WiFi radios as a biometric way to identify the owner of a device. Personal Marks is a simple cryptographic protocol that works well when the attacker tries to use the stolen credentials in the social community of the victim. Community Certificates works well when the adversary has the goal of using the stolen credentials when interacting with entities that are far from the social network of the victim. When combined, these mechanisms provide an excellent protection against identity theft attacks. In Chapter 3 we prove our ideas and solutions with extensive simulations in both simulated and real world scenarios—with mobility traces collected in a real life experiment. This study appeared in [2]. Another way of accessing the private data of a user, other than getting physical access to his device, could be by means of a malware. An emerging trend in the way people are fooled into installing malware-infected apps is that of exploiting existing trust relationships between socially close users, like those between Facebook friends. In this way, the malware can rapidly expand through social links from a small set of infected devices towards the rest of the network. In our quest for hybrid solutions to the problem of malware spreading in social networks of mobile users we developed a novel approach based on the Mobile Cloud Computing paradigm. In this new paradigm, a mobile device can alleviate the burden of computationally intensive tasks by offloading them to a software clone running on the cloud. Also, the clones associated to devices of users in the same community are connected in a social peer-to-peer network, thus allowing lightweight content sharing between friends. CloudShield is a suite of protocols that provides an efficient way stop the malware spread by sending a small set of patches from the clones to the infected devices. Our experiments on different datasets show that CloudShield is able to better and more efficiently contain malware spreading in mobile wireless networks than the state-of-the-art solutions presented in literature. These findings (which are not included in this dissertation) appeared in [3] and are the result of a joint work with P.h.D student S. Kosta from Sapienza University. My main contribution to this work was in the simulation of both the malware spreading and of the patching protocol schemes on the different social networks datasets. The Mobile Cloud Computing paradigm seems to be an excellent resource for mobile systems. It alleviates battery consumption on smartphones, it helps backing up user’s data on-the-fly and, as CloudShield proves, it can also be used to find new, effective, solutions to existing problems. However, the communication between the mobile devices and their clones needed by such paradigm certainly does not come for free. It costs both in terms of bandwidth (the traffic overhead to communicate with the cloud) and in terms of energy (computation and use of network interfaces on the device). Being aware of the issues that heavy computation or communication can cause to both the battery life of the devices [13], and to the mobile infrastructure, we decided to study the actual feasibility of both mobile computation offloading and mobile software/data backups in real-life scenarios. In our study we considered two types of clones: The off-clone, whose purpose is to support computation offloading, and the back-clone, which comes to use when a restore of user’s data and apps is needed. In Chapter 5 we give a precise evaluation of the feasibility and costs of both off-clones and back-clones in terms of bandwidth and energy consumption on the real device. We achieved this by means measurements done on a real testbed of 11 Android smartphones and on their relative clones running on the Amazon EC2 public cloud. The smartphones have been used as the primary mobile by the participants for the whole experiment duration. This study has been presented in [4] and is the result of a collaboration with P.h.D. Student S. Kosta from Sapienza University. S. Kosta mainly contributed to the experimental setup, deployment of the testbed and data collection.
Research interests: Computer Networks, Network Security, Cloud Computing, Network Neutrality
Personal skills keywords: Programming Languages (C, C++, Java, Python)
Android development
Unix System Administration

Files in This Item:

File Description SizeFormat
thesis.pdfTesi di Dottorato1.18 MBAdobe PDF

File del Curriculum Vitae:

CurriculumVitae.pdf 188.84 kBAdobe PDF

This item is protected by original copyright

Recommend this item

Items in PADIS are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - Feedback Sviluppo e manutenzione a cura del CINECA