Sriram Sankaran

Energy modeling for mobile devices using performance counters

The increasing complexity of mobile applications coupled with growing user demands lead to rapid battery drain in mobile devices. However, battery technology cannot keep up with these trends thus making power management one of the foremost concerns. While system-level approaches to power management exist, the energy impact of applications on individual system components needs to be better understood for energy efficient system design. In this work, we develop energy models for mobile devices using performance counters and estimate the power consumption of system components for numerous embedded applications. Our models provide enhancements in I/O towards estimating I/O energy and cache to incorporate energy consumed during cache refill and write-back in the energy estimation process. We further compare our power estimates with existing models and demonstrate the uniqueness of our model.

Predictive modeling based power estimation for embedded multicore systems

The increasing number of cores in embedded devices results in improved performance compared to single-core systems. Further, the unique characteristics of these systems provide numerous opportunities for power management which require models for power estimation. In this work, a statistical approach that models the impact of the individual cores and memory hierarchy on overall power consumed by Chip Multiprocessors is developed using Performance Counters. In particular, we construct a per-core based power model using SPLASH2 benchmarks by leveraging concurrency for multicore systems. Our model is simple and technology independent and as a result executes faster incurring lesser overhead. Evaluation of the model shows a strong correlation between core-level activity and power consumption and that the model predicts power consumption for newer observations with minimal errors. In addition, we discuss a few applications where the model can be utilized towards estimating power consumption.

User-adaptive energy-aware security for mobile devices

Energy Management is of primary importance in mobile devices due to increasing functionality coupled with rapid battery drain. Our analysis reveals that users differ in their context and resource usage patterns which can be profiled towards developing predictive models for energy savings. A key challenge lies in providing user-adaptive security in an energy-aware manner due to increasing sensitivity of user data and analyzing the energy-security trade-offs which we address in this work. Towards this goal, we develop a statistical user model to predict available energy at a given time instant using historical user data and further describe a generic multi-level security model for mobile devices. The available energy from the user model in conjunction with the energy estimates from the security model can be used for energy-aware security adaptation in mobile devices.

Pattern Matching Based Sensor Identification Layer for an Android Platform

As sensor-related technologies have been developed, smartphones obtain more information from internal and external sensors. This interaction accelerates the development of applications in the Internet of Things environment. Due to many attributes that may vary the quality of the IoT system, sensor manufacturers provide their own data format and application even if there is a well-defined standard, such as ISO/IEEE 11073 for personal health devices. In this paper, we propose a client-server-based sensor adaptation layer for an Android platform to improve interoperability among nonstandard sensors. Interoperability is an important quality aspect for the IoT that may have a strong impact on the system especially when the sensors are coming from different sources. Here, the server compares profiles that have clues to identify the sensor device with a data packet stream based on a modified Boyer-Moore-Horspool algorithm. Our matching model considers features of the sensor data packet. To verify the operability, we have implemented a prototype of this proposed system. The evaluation results show that the start and end pattern of the data packet are more efficient when the length of the data packet is longer.

Sybil attack in IOT: Modelling and defenses

Internet of Things (IoT) is an emerging paradigm in information technology (IT) that integrates advancements in sensing, computing and communication to offer enhanced services in everyday life. IoTs are vulnerable to sybil attacks wherein an adversary fabricates fictitious identities or steals the identities of legitimate nodes. In this paper, we model sybil attacks in IoT and evaluate its impact on performance. We also develop a defense mechanism based on behavioural profiling of nodes. We develop an enhanced AODV (EAODV) protocol by using the behaviour approach to obtain the optimal routes. In EAODV, the routes are selected based on the trust value and hop count. Sybil nodes are identified and discarded based on the feedback from neighbouring nodes. Evaluation of our protocol in ns-2 simulator demonstrates the effectiveness of our approach in identifying and detecting sybil nodes in IoT network.

Modeling the performance of IoT networks

Internet of Things (IoTs) is gaining increasing significance due to real-time communication and decision making capabilities of sensors integrated into everyday objects. Predicting performance in IoTs is critical for detecting performance bottlenecks, designing optimal sleep/wake-up schedules and application-aware performance tuning. However, performance prediction becomes a significant challenge in IoTs due to varying needs of applications coupled with the resource constrained nature of sensors. In this work, we analyze the impact of factors affecting performance in IoT networks using simulation based models. Further, an analytical framework is developed to model the impact of individual node behavior on overall performance using Markov chains. In particular, we derive steady state transition probabilities of transmit and receive states using protocol execution traces and further utilize them towards predicting per-flow throughput. Our proposed model is generic in that it can be applied across domains. Accuracy of the model is evaluated by comparing the predictions with the actual estimates obtained using simulations.