Geo-distributed Situation Awareness applications are large in scale and are characterized by 24/7 data generation from mobile and stationary sensors (such as cameras and GPS devices); latency-sensitivity for converting sensed data to actionable knowledge; and elastic and bursty needs for computational resources. Fog computing [7] envisions providing computational resources close to the edge of the network, consequently reducing the latency for the sense-process-actuate cycle that exists in these applications. We propose Foglets, a programming infrastructure for the geo-distributed computational continuum represented by fog nodes and the cloud. Foglets provides APIs for a spatio-temporal data abstraction for storing and retrieving application generated data on the local nodes, and primitives for communication among the resources in the computational continuum. Foglets manages the application components on the Fog nodes. Algorithms are presented for launching application components and handling the migration of these components between Fog nodes, based on the mobility pattern of the sensors and the dynamic computational needs of the application. Evaluation results are presented for a Fog network consisting of 16 nodes using a simulated vehicular network as the workload. We show that the discovery and deployment protocol can be executed in 0.93 secs, and joining an already deployed application can be as quick as 65 ms. Also, QoS-sensitive proactive migration can be accomplished in 6 ms.

No abstract

Up to now, correlations in complex event processing (CEP) systems are detected by a well defined set of operators, whose configuration is determined ahead of deployment time. Although CEP operators involve location specific attributes, state of the art systems are heavily constrained in detecting situations where the interest in a situation changes depending on the consumer's location, e.g., with the movement of mobile devices.This paper adopts the concept of range queries to CEP systems. We propose a mobility-aware event delivery semantics and present a corresponding execution model, which accounts for mobility driven selection of primary event streams to the CEP system. By utilizing the properties of this execution model, we derive an algorithm that establishes low cost and coordinated reconfiguration of CEP operators in a distributed system. The algorithm minimizes the amount of information that needs to be streamed between operators and avoids additional delays as a result of a reconfiguration of CEP operators. We present an analysis of the algorithm's properties and evaluate the efficiency of the proposed reconfiguration algorithm.

A recent trend in communication networks -sometimes referred to as fog computing -offers to execute computational tasks close to the access points of the networks. This enables real-time applications, like mobile Complex Event Processing (CEP), to significantly reduce end-to-end latencies and bandwidth usage. Most work studying the placement of operators in such an environment completely disregards the migration costs. However, the mobility of users requires frequent migration of operators, together with possibly large state information, to meet latency restrictions and save bandwidth in the infrastructure.This paper presents a placement and migration method for providers of infrastructures that incorporate cloud and fog resources. It ensures application-defined end-to-end latency restrictions and reduces the network utilization by planning the migration ahead of time. Furthermore, we present how the application knowledge of the CEP system can be used to improve current live migration techniques for Virtual Machines (VMs) to reduce the required bandwidth during the migration. Our evaluations show that we safe up to 49% of the network utilization with perfect knowledge about a users mobility pattern and up to 27% of the network utilization when considering the uncertainty of those patterns.

With the proliferation of mobile devices and sensors, mobile situation awareness is becoming an important class of applications. The key requirement of this class of applications is low-latency processing of events stemming from sensor data in order to provide timely situational information to mobile users. To satisfy the latency requirement, we propose an opportunistic spatio-temporal event processing system that uses prediction-based continuous query handling. Our system predicts future query regions for moving consumers and starts processing events early so that the live situational information is available when the consumer reaches the future location. In contrast to existing systems, our system provides timely information about a consumer's current position by hiding computation latency for processing recent events. To evaluate our system, we measure the quality of results and timeliness of live situational information with various query parameters. Our evaluation shows that we can achieve highly meaningful query results with near-zero latency in most cases.