Research in M.S.E. Course

Mobile Ad hoc Computational Cloud

This research aims to develop a mobile ad hoc distributed computing infrastructure which will allow mobile nodes to share computing resources to form an integrated mobile cloud computing infrastructure through a mobile ad hoc networking system. In mobile ad hoc computational cloud, multiple nearby mobile devices interconnected through a mobile ad hoc network are combined to form a single virtual, unified powerful computing system. The main purpose is to empower mobile devices’ capabilities for supporting the execution of computationally and data-intensive (i.e. image and video processing) applications on mobile devices in mobile ad hoc environment. The system architecture of mobile ad hoc computational cloud is illustrated in Figure 1.

A mobile ad hoc computational cloud may include several types of mobile computing devices such as laptops, tablets, smart phones, wearable computing devices, miniature autonomous mobile robots and vehicles. According to their roles in MAC, mobile nodes are divided into three following categories: service consumer node, service provider node, and service master node.

The service consumer node (SCN) refers to nodes that request/consume value added abstract computing resources provided by MAC service provider nodes (SPN). SCNs submit computation and data-intensive application tasks to mobile ad hoc computational cloud, whereas SPNs are nodes contributing with computing resources used to perform users’ submitted tasks by SCNs. Service master node (SMN) is the maestro, which coordinates activities and resources sharing within MAC system, and is responsible for discovery process, resources management, task allocation, task migration, failure management, etc.

A mobile ad hoc computational cloud should include at least one service consumer node, one service provider node, and one service master node. Mobile ad hoc cloud member nodes refer to nodes contributing with computing resources to MAC, and may be whether service provider or master nodes. SCNs need only a minimal middleware configuration to access and/or consume provided resources or services. Based on the context, a single node can perform simultaneously multiple roles. For example, a node may act as service master node and/or service provider node and consuming services from other node members.

The mobile ad hoc computational cloud system framework architecture is shown in Figure 2. It comprises essentially four layers described below:

• Application layer includes end-users’ applications running on mobile ad hoc computational cloud system.

• Mobile Ad hoc Computational Cloud (MAC) layer is a distributed middleware that enables autonomous mobile computing nodes to coordinate their activities and to share computing resources. Thus, it hides all the complexities and provides a single system image to users and applications running on the MAC system.

In order to design and implement a mobile ad hoc computational cloud middleware the following keys components and services are required:

· Resource pool contains all information about available computing resources in the MAC, associated with their status as well.

· Resource management and task migration: resource management service is in charge of managing the pool of available computing resource in MAC and users’ submitted tasks allocation, whereas task migration service migrates tasks from one node to another node across the network. It is also used to avoid task failure due to hardware and software problems.

The main functions or services of a MAC resource management system include:

- Resource description service uses a resource description language (RDL) to describe computing resources/services available within a coverage range.

- Resource discovery service is used to find available computing resources within the coverage area.

- Resource allocation service finds suitable computing resources available in the resource pool to fulfill the tasks execution requirements and assigns these resources to tasks for execution.

- Resource monitoring service monitors continuously the tasks execution progress and mobile device resource utilization status.

- Task migration service migrates tasks from one node to another across the network to avoid task failure and/or computing load imbalance.

· Application task management provides task submission interface and manages end users’ submitted tasks to mobile ad hoc computational cloud.

· Task queue contains users’ applications submitted tasks to the MAC, waiting for their allocation and execution.

· Mobility management is responsible for node mobility tracking and periodically reporting the node mobility status for effective resource allocation decision.

· Security and privacy management provides to MAC infrastructure and users’ applications security and privacy preservation features, including user/system authentication, resource access control, data integrity, trust and service level agreement (SLA).

· QoS management system is used to ensure that user application tasks’ requirements such as processing power, available memory, available energy, bandwidth guarantees are fulfilled.

• Mobile Ad hoc Network (MANET) layer provides networking and communication services including data transfer, routing and medium access control services to mobile nodes, which communicate with each other through a mobile ad hoc network.

• Mobile node resource layer includes mobile devices’ resources such as processing resource, memory, storage and networking resource.

※ Benefits of Mobile Ad Hoc Computational Cloud Systems

Mobile ad hoc computational cloud is still attractive even when infrastructure-based communication network systems are available. In the sense that it enables multiple computing mobile devices to collaborate and share computing resources to form a virtual supercomputing node, without relying on any pre-existing communication network infrastructure. Mobile devices communicate through mobile ad hoc network which is cheaper, easy and fast to deploy, especially in the context of security defense and disaster relief emergency response operations. Where communication infrastructures are not always guaranteed to be available. Even if they are available their communication performance is much lower than short range wireless communication technologies such as Wi-Fi Direct. For example, compared to 4^th generation networks that provide 100 megabits per second of data rate, wireless local area networks provide up to 600 megabits per second of data rate with less transmission energy consumption. Mobile ad hoc computational cloud systems offer many opportunities to design and develop next generation of smart and digital convergence solutions.

※ Application Scenarios

Mobile ad hoc computational cloud has several applications in numerous domains such as mobile games, mobile healthcare and bio-surveillance systems, mobile learning and natural language processing, patterns recognition and augmented cognitive assistance, robotics, security defense and military applications. Cyber-physical systems, big data analysis, Internet of Things, and smart city applications can also benefit from MAC systems as well.

Mobile ad hoc computational cloud can play a key role in success of security defense and disaster relief emergency response missions. A MAC application scenario for security defense and disaster relief emergency response illustrated in Figure 3 is described below.

Hybrid Mobile Ad Hoc Cloud Computing