The current architecture supporting data services to mobile devices is built below the network layer (IP) and users receive the payload at the application layer. Between them is the transport layer that can cause data consumption inflation due to the retransmission mechanism that provides reliable delivery. In this paper, we examine the accounting policies of five large cellular ISPs in the U.S. and South Korea. We look at their policies regarding the transport layer reliability mechanism with TCP's retransmission and show that the current implementation of accounting policies either fails to meet the billing fairness or is vulnerable to charge evasions. Three of the ISPs surveyed charge for all IP packets regardless of retransmission, allowing attackers to inflate a victim's bill by intentionally retransmitting packets. The other two ISPs deduct the retransmitted amount from the user's bill thus allowing tunneling through TCP retransmissions. We show that a "free-riding" attack is viable with these ISPs and discuss some of the mitigation techniques.

Packet retransmission is a fundamental TCP mechanism that ensures reliable data transfer between two end nodes. Interestingly, when it comes to cellular data accounting, TCP retransmissions create an important policy issue giving rise to a tension between ISPs accounting for network resource consumption, and users only being aware of the application layer data. Regardless of the policies, we find that TCP retransmissions can be easily abused to manipulate the current practice of cellular traffic accounting. In this work, we investigate the TCP retransmission accounting policies of 12 cellular ISPs in 6 countries and report the accounting vulnerabilities with TCP retransmission attacks. First, we find that cellular data accounting policies vary between ISPs. While the majority of cellular ISPs blindly account for every IP packet, some ISPs intentionally remove the retransmission packets from the user bill for fairness. Second, we show that it is easy to launch the "usage-inflation" attack on the ISPs that blindly account for every IP packet. In our experiments, we could inflate the usage up to the monthly limit with an attack invisible to the subscriber and lasting only 9 minutes. For those ISPs that do not account for retransmission, we successfully launch the "free-riding" attack by tunneling the payload over fake TCP headers that look like retransmissions. To counter the attacks, we implement and evaluate Abacus, a lightweight , scalable accounting system that reliably detects "free-riding" attacks even in the 10 Gbps links. Permission to freely reproduce all or part of this paper for noncommercial purposes is granted provided that copies bear this notice and the full citation on the first page. Reproduction for commercial purposes is strictly prohibited without the prior written consent of the Internet Society, the first-named author (for reproduction of an entire paper only), and the author's employer if the paper was prepared within the scope of employment.

The explosive popularity of smartphones and mobile devices drives massive growth in the wide-area mobile data communication. Unfortunately, the current or near-future 3G/4G networks are deemed insufficient to meet the increasing data transfer demand. While opportunistic offloading of mobile data through Wi-Fi is an attractive option, the existing transport layer would experience frequent disconnections due to mobility, making it hard to support seamlessly reliable data delivery. As a result, many mobile applications either depend on ad-hoc downloading resumption mechanisms or redundantly re-transfer the same content when disruptions happen.In this paper, we present DTP, a disruption-tolerant, reliable transport layer protocol that masks the failures of the preferred network. Unlike previous disruption/delay-tolerant protocols, DTP provides the same semantics as TCP on an IP packet level when the mobile device is connected to a network while providing the illusion of continued connection even if the underlying physical network becomes unavailable. This would help the mobile application developers to focus on the application core rather than addressing the frequent network disruptions. It would also greatly reduce the phone network costs both to ISPs and end users. Our current implementation in UDP shows a comparable performance to that of TCP in network, and it greatly reduces the delay and power consumption when the mobile devices frequently switch from one network to another.

Developers of cloud-connected mobile apps need to ensure the consistency of application and user data across multiple devices. Mobile apps demand different choices of distributed data consistency under a variety of usage scenarios. The apps also need to gracefully handle intermittent connectivity and disconnections, limited bandwidth, and client and server failures. The data model of the apps can also be complex, spanning inter-dependent structured and unstructured data, and needs to be atomically stored and updated locally, on the cloud, and on other mobile devices.In this paper we study several popular apps and find that many exhibit undesirable behavior under concurrent use due to inadequate treatment of data consistency. Motivated by the shortcomings, we propose a novel data abstraction, called a sTable, that unifies a tabular and object data model, and allows apps to choose from a set of distributed consistency schemes; mobile apps written to this abstraction can effortlessly sync data with the cloud and other mobile devices while benefiting from end-to-end data consistency. We build Simba, a data-sync service, to demonstrate the utility and practicality of our proposed abstraction, and evaluate it both by writing new apps and porting existing inconsistent apps to make them consistent. Experimental results show that Simba performs well with respect to sync latency, bandwidth consumption, server throughput, and scales for both the number of users and the amount of data.