Sherif Abdelwahab

Enabling Smart Cloud Services Through Remote Sensing: An Internet of Everything Enabler

The recent emergence and success of cloud-based services has empowered remote sensing and made it very possible. Cloud-assisted remote sensing (CARS) enables distributed sensory data collection, global resource and data sharing, remote and real-time data access, elastic resource provisioning and scaling, and pay-as-you-go pricing models. CARS has great potentials for enabling the so-called Internet of Everything (IoE), thereby promoting smart cloud services. In this paper, we survey CARS. First, we describe its benefits and capabilities through real-world applications. Second, we present a multilayer architecture of CARS by describing each layers functionalities and responsibilities, as well as its interactions and interfaces with its upper and lower layers. Third, we discuss the sensing services models offered by CARS. Fourth, we discuss some popular commercial cloud platforms that have already been developed and deployed in recent years. Finally, we present and discuss major design requirements and challenges of CARS.

Replisom : Disciplined Tiny Memory Replication for Massive IoT Devices in LTE Edge Cloud

Augmenting the long-term evolution (LTE)-evolved NodeB (eNB) with cloud resources offers a low-latency, resilient, and LTE-aware environment for offloading the Internet of Things (IoT) services and applications. By means of devices memory replication, the IoT applications deployed at an LTE-integrated edge cloud can scale its computing and storage requirements to support different resource-intensive service offerings. Despite this potential, the massive number of IoT devices limits the LTE edge cloud responsiveness as the LTE radio interface becomes the major bottleneck given the unscalability of its uplink access and data transfer procedures to support a large number of devices that simultaneously replicate their memory objects with the LTE edge cloud. We propose Replisom; an LTE-aware edge cloud architecture and an LTE-optimized memory replication protocol which relaxes the LTE bottlenecks by a delay and radio resource-efficient memory replication protocol based on the device-to-device communication technology and the sparse recovery in the theory of compressed sampling. Replisom effectively schedules the memory replication occasions to resolve contentions for the radio resources as a large number of devices simultaneously transmit their memory replicas. Our analysis and numerical evaluation suggest that this system has significant potential in reducing the delay, energy consumption, and cost for cloud offloading of IoT applications given the massive number of devices with tiny memory sizes.

Efficient Virtual Network Embedding With Backtrack Avoidance for Dynamic Wireless Networks

We develop an efficient virtual network embedding (VNE) algorithm, termed Bird-VNE, for mobile wireless networks. Bird-VNE is an approximation algorithm that ensures a close to optimal virtual embedding profit and acceptance rate while minimizing the number of virtual network migrations resulting from the mobility of wireless nodes. Bird-VNE employs a constraint satisfaction framework by which we analyze the constraint propagation properties of the VNE problem and design constraint processing algorithms that efficiently narrow the solution space and avoid backtracking as much as possible without compromising the solution quality. Our evaluation results show that the likelihood that Bird-VNE results in backtracking is small, thus demonstrating its effectiveness in reducing the search space. We analytically and empirically verify that Bird-VNE outperforms existing VNE algorithms with respect to computational efficiency, closeness to optimality, and its ability to avoid potential migrations in mobile wireless networks.

Cloud of Things for Sensing as a Service: Sensing Resource Discovery and Virtualization

We propose Cloud of Things for Sensing as a Service: a global architecture that scales up cloud computing by exploiting the global sensing resources of the highly dynamic and growing Internet of Things (IoT) to enable remote sensing. The proposed architecture scales out by augmenting the role of edge computing platforms as cloud agents that discover and virtualize sensing resources of IoT devices. Cloud of Things enables performing in-network distributed processing of sensing data offered by the globally available IoT devices and provides a global platform for meaningful and responsive sensing data analysis and decision making. We design cloud agents algorithmic solutions bearing in mind the onerous to track dynamics of the IoT devices by centralized solutions. First, we propose a distributed sensing resource discovery algorithm based on a gossip policy that selects IoT devices with predefined sensing capabilities as fast as possible. We also propose RADV: a distributed virtualization algorithm that efficiently deploys virtual sensor networks on top of a subset of the selected IoT devices. We show, through analysis and simulations, the potential of the proposed algorithmic solutions to realize virtual sensor networks with minimal physical resources, reduced communication overhead, and low complexity.

Cloud of Things for Sensing-as-a-Service: Architecture, Algorithms, and Use Case

We propose Cloud of Things for sensing-as-a-service: a global architecture that scales up cloud computing by exploiting the global sensing resources of the Internet of Things (IoT) to enable remote sensing. Cloud of Things enables in-network distributed processing of sensors data offered by the globally available IoT devices and provides a global platform for meaningful and responsive data analysis and decision making. We propose a distributed sensing resource discovery and virtualization algorithms that efficiently deploy virtual sensor networks on top of a subset of the selected IoT devices. We show, through analysis and simulations, the potential of the proposed solutions to realize virtual sensor networks with minimal physical resources, reduced communication overhead, and low complexity. We also design an uncoordinated, distributed algorithm that relies on the selected sensors to estimate a set of parameters without requiring synchronization among the sensors. Our simulations show that the proposed estimation algorithm, when compared to conventional alternating direction method of multipliers (ADMMs), reduces communication overhead significantly without compromising the estimation error. In addition, the convergence time, though increases slightly, is still linear as in the case of conventional ADMM.

BIRD-VNE: Backtrack-avoidance virtual network embedding in polynomial time

The virtual network embedding (VNE) problem is known to be NP-hard, and as a result, several heuristic approaches have been proposed to solve it. These heuristics find sub-optimal solutions in polynomial time, but have practical limitations, low acceptance rates, and high embedding costs. In this paper, we first propose two heuristics that exploit the constraint propagation properties of the VNE problem to ensure both topological and capacity disjoint consistencies, thereby avoiding backtracking while increasing acceptance rates. Then, combining these two heuristics, we design a polynomial-time VNE algorithm (we term it BIRD-VNE) that, in addition to avoiding backtracking and increasing acceptance rates, incurs a low embedding cost when compared to existing approaches.

Flocking virtual machines in quest for responsive IoT cloud services

We propose Flock; a simple and scalable protocol that enables live migration of Virtual Machines (VMs) across heterogeneous edge and conventional cloud platforms to improve the responsiveness of cloud services. Flock is designed with properties that are suitable for the use cases of the Internet of Things (IoT). We describe the properties of regularized latency measurements that Flock can use for asynchronous and autonomous migration decisions. Such decisions allow communicating VMs to follow a flocking-like behavior that consists of three simple rules: separation, alignment, and cohesion. Using game theory, we derive analytical bounds on Flocks Price of Anarchy (PoA), and prove that flocking VMs converge to a Nash Equilibrium while settling in the best possible cloud platforms. We verify the effectiveness of Flock through simulations and discuss how its generic objective can simply be tweaked to achieve other objectives, such as cloud load balancing and energy consumption minimization.