Search

Edge computing is the practice of placing computing resources at the edges of the Internet in close proximity to devices and information sources. This, much like a cache on a CPU, increases bandwidth and reduces latency for applications but at a potential cost of dependability and capacity.

Cellular networks have become an essential part of our lives. With increasing demands on its available bandwidth, we are seeing failures and performance degradations for data and voice traffic on the rise. In this paper, we propose the view that fog computing, integrated in the edge components of cellular networks, can partially alleviate this situation. In our vision, some data gathering and data analytics capability will be developed at the edge of the cellular network and client devices and the network using this edge capability will coordinate to reduce failures and performance degradations. We also envisage proactive management of disruptions including prediction of impending events of interest (such as, congestion or call drop) and deployment of appropriate mitigation actions. We show that a simple streaming media pre-caching service built using such device-fog cooperation significantly expands the number of streaming video users that can be supported in a nominal cellular network of today.

With the growing popularity of the Internet of Things (IoT), embedded devices continue to integrate more into our daily lives. For this reason, security for embedded devices is a vital issue to address. Attacks such as stack smashing, code injection, data corruption and Return Oriented Programming (ROP) are still a threat to embedded systems. As new methods are developed to defend embedded systems against such attacks, a benchmark to compare these methods is not present. In this work, a benchmark is presented that is aimed at testing the security of new techniques that defend against these common attacks. Two programs are developed that carry three key values needed for a benchmark: realistic embedded application, complex control flow, and being deterministic. The first application is a pin lock system and the second is a compression data logger. A complexity evaluation of the two applications revealed that the pin lock system contained 171 functions and 190 nodes with 252 edges in the control-flow graph, and the compression data logger contained 192 functions and 1,357 nodes with 2,123 edges in the control-flow graph. The current benchmark will be improved in the future by adding more applications with a wider range of complexity.

Wireless sensor networks (WSNs) are plagued by the possibility of bugs manifesting only at deployment. However, debugging deployed WSNs is challenging for several reasons—the remote location of deployed sensor nodes, the non-determinism of execution that can make it difficult to replicate a buggy run, and the limited hardware resources available on a node. In particular, existing solutions to record and replay debugging in WSNs fail to capture the complete code execution, thus negating the possibility of a faithful replay and causing a large class of bugs to go unnoticed. In short, record and replay logs a trace of predefined events while a deployed application is executing, enabling replaying of events later using debugging tools. Existing recording methods fail due to the many sources of non-determinism and the scarcity of resources on nodes. In this paper we introduce a software-only approach for deterministic record and replay of WSN nodes. The approach records all sources of non-determinism, based on the observation that such information is compressible using a combination of techniques specialized for respective sources. Despite their domain-specific nature, the techniques presented are applicable to the broader class of resource-constrained embedded systems.

Web traffic accounts for more than half of Internet traffic today. Camouflaging covert timing channels in Web traffic would be advantageous for concealment. In this paper, we investigate the possibility of disguising network covert timing channels as HTTP traffic to avoid detection. Extensive research has shown that Internet traffic, including HTTP traffic, exhibits self-similarity and long range persistence. Existing covert timing channels that mimic i.i.d. legitimate traffic cannot imitate HTTP traffic because these covert traffic patterns are not long range dependent. The goal of this work is to design a covert timing channel that can be camouflaged as HTTP traffic. To this end, we design a covert timing channel whose inter-arrival times are long range dependent and have the same marginal distribution as the interarrival times for new HTTP connection traffic. These inter-arrival times are constructed by combining a Fractional Auto-Regressive Integrated Moving Average (FARIMA) time series and an i.i.d. cryptographically secure random sequence. Experiments are conducted on PlanetLab, and the results are validated against recent real traffic trace data. Our experiments demonstrate that the traffic from this timing channel traffic is statistically indistinguishable from legitimate HTTP traffic and undetectable by all current detection schemes for timing channels.

In many reliable Wireless Sensor Network (WSN) applications, messages have different priorities depending on urgency or importance. For example, a message reporting the failure of all nodes in a region is more important than that for a single node. Moreover, traffic can be bursty in nature, such as when a correlated error is reported by multiple nodes running identical code. Current communication layers in WSNs lack efficient support for these two requirements. We present a priority-driven communication layer, called SeNDORComm, that schedules transmission of packets driven by application-specified priority, buffers and packs multiple messages in a packet, and honors latency guarantee for a message. We show that SeNDORComm improves energy efficiency, message reliability, network utilization and delays congestion collapse in a network. We extensively evaluate SeNDORComm using analysis, simulation and real experiments. We demonstrate the improvement in goodput of SeNDORComm over a default communication layer (134.78% for a network of 20 nodes), such as GenericComm in TinyOS.

Wireless Sensor Networks are increasingly being used for data monitoring in commercial, industrial, and military applications. Security is of great concern from many different viewpoints: ensuring that sensitive data does not fall into wrong hands; ensuring that the received data has not been doctored; and ensuring that the network is resilient to denial of service attacks. We study the fundamental problem of Secure Neighbor Discovery problem, which is critical to protecting the network against a number of different forms of attacks. Sensor networks, deployed in hazardous environment, are exposed to a variety of attacks like eavesdropping, message tampering, selective forwarding, wormhole and sybil attacks. Attacks against the data traffc can be addressed using cryptographic techniques. We frst present an effcient and scalable key-distribution protocol which is completely secure in the absence of colluding malicious nodes. Secure neighbor discovery can help to defend against a majority of the attacks against control traffc. We consider a static network and propose a secure one-hop neighbor discovery protocol. We show by analysis that this protocol effectively prevents two non-neighboring nodes from becoming neighbors even when both the nodes have been compromised by the adversary. We then extend this protocol so that it works even when nodes are incrementally deployed in the network. We also brie y study how this protocol could be modifed for mobile sensor networks. Finally, we compare our protocol with existing neighbor discovery protocols and analyze the advantages and disadvantages of using these protocols.

Sensor networks enable a wide range of applications in both military and civilian domains. However, the deployment scenarios, the functionality requirements, and the limited capabilities of these networks expose them to a wide-range of attacks against control traffic (such as wormholes, rushing, Sybil attacks, etc) and data traffic (such as selective forwarding). In this paper we propose a framework called DICAS that mitigates such attacks by detecting, diagnosing, and isolating the malicious nodes. DICAS uses as a fundamental building block the ability of a node to oversee its neighboring nodes’ communication. On top of DICAS, we build a secure routing protocol, LSR, that provides additional protection against malicious nodes by supporting multiple node-disjoint paths. We analyze the security guarantees of DICAS and use ns-2 simulations to show its effectiveness against representative control and data attacks. The overhead analysis we present shows that DICAS is a lightweight protocol appropriate for securing resource constrained sensor networks.

As sensor networks operate over long periods of deployment in difficult to reach places, their requirements may change or new code may need to be uploaded to them. The current state of the art protocols (Deluge and MNP) for network reprogramming perform the code dissemination in a multi-hop manner using a three way handshake whereby meta-data is exchanged prior to code exchange to suppress redundant transmissions. The code image is also pipelined through the network at the granularity of pages. In this paper we propose a protocol called Freshet for optimizing the energy for code upload and speeding up the dissemination if multiple sources of code are available. The energy optimization is achieved by equipping each node with limited non-local topology information, which it uses to determine the time when it can go to sleep since code is not being distributed in its vicinity. The protocol to handle multiple sources provides a loose coupling of nodes to a source and disseminates code in waves each originating at a source, with mechanism to handle collisions when the waves meet. The protocol’s performance with respect to reliability, delay, and energy consumed, is demonstrated through analysis, simulation, and implementation on the Berkeley mote platform.

required to diagnose the failure, i.e., to identify the source of the failure. Diagnosis is challenging because fast error propagation may occur in high throughput distributed applications. The diagnosis often needs to be probabilistic in nature due to imperfect observability of the payload system, inability to do white-box testing, constraints on the amount of state that can be maintained at the diagnostic process, and imperfect tests used to verify the system. In this paper, we extend an existing Monitor architecture, for probabilistic diagnosis of failures in large-scale network protocols. The Monitor only observes the message exchanges between the protocol entities (PEs) remotely and does not access internal protocol state. At runtime, it builds a causal & aggregate graph between the PEs based on their communication and uses this together with a rule base for diagnosing the failure. The Monitor computes for each suspected PE, a probability for the error having originated in that PE and propagated to the failure detection site. The framework is applied to a test-bed consisting of a reliable multicast protocol executing on the Purdue campus-wide network. Error injection experiments are performed to evaluate the accuracy and the performance overhead of the diagnostic process.

In today's world where distributed systems form many of our critical infrastructures, dependability outages are becoming increasingly common. In many situations, it is necessary to not just detect a failure, but also to diagnose the failure, i.e., to identify the source of the failure. Diagnosis is challenging since high throughput applications with frequent interactions between the different components allow fast error propagation. It is desirable to consider applications as black-boxes for the diagnosis process. In this paper, we propose a Monitor architecture for diagnosing failures in large-scale network protocols. The Monitor only observes the message exchanges between the protocol entities (PEs) remotely and does not access internal protocol state. At runtime, it builds a causal graph between the PEs based on their communication and uses this together with a rule base of allowed state transition paths to diagnose the failure. The tests used for the diagnosis are based on the rule base and are assumed to have imperfect coverage. The hierarchical Monitor framework allows distributed diagnosis handling Byzantine failures at individual Monitors. The framework is implemented and applied to a reliable multicast protocol executing on our campus-wide network. Fault injection experiments are carried out to evaluate the accuracy and latency of the diagnosis.