Summary: Through the convergence of smartphones and cloud computing, mobile services are being provided with richer communication and higher flexibility. But there are some serious security threats to this mobile cloud infrastructure as malwares can be run on virtual mobile instances targeting on the virtualization. Although signature based vaccine applications can detect such malwares but it makes an additional overhead on instances.
Past researches have been made in the concerned domain, where some have focused on the malware detection by monitoring behavior in mobile devices using several features grouped together to define the behavior. While others targeted on intrusion detection for cloud computing infrastructure. This paper is focused on the behavioral based abnormal detection in mobile cloud infrastructure. Here, mobile cloud computing is defined as processing jobs for mobile devices in cloud computing infrastructure and delivering job results to mobile devices.
The authors also discussed some possible scenarios to explain how the mobile cloud service can be used by individual users and office workers. When a mobile application is executed on the mobile cloud infrastructure it should use some virtual resources e.g. CPU or memory, which in turn changes the value of these resources. So based on this assumption, it is deduced that each mobile application and each user has a unique behavioral pattern.
The proposed methodology for monitoring and detecting abnormal behavior observes activities of both host data on virtual mobile instances and network data of mobile cloud infrastructure. To monitor the host data on each virtual mobile instance, the proposed architecture installs an agent mobile application into virtual mobile instance. The host data which comprises of the detailed information about CPU usage, Memory, Process, OS & Network data, is then transported to the analyzer. Network data is monitored through port-mirroring that is provided by virtual routers.
The analyzer performs Random Forest (RF) machine learning algorithm to train abnormal behavior with the collected data set. In the test bed, GingerBread 2.3.6 is used with Linux kernel version 2.6.39. Open vSwitch is installed to use the port-mirroring functionality in each physical node.
To validate the methodology, some malicious programs have to be injected into the mobile cloud test bed alongwith the normal applications, so the test environment is consisted of total ten mobile instances in which two are malicious applications by ‘GoldMiner’. The proposed methodology is finally tested by deploying the mobile cloud test bed and it detects the abnormal behavior successfully.
Cloud computing: Issues and Challenges Tharam Dilon, Chen Wu and Elizabeth Chang Digital EcoSystems and Business Intelligence Institute Curtin University of Technology Perth, Australia Summary: Cloud Computing is a model for enabling convenient, on demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction. The five essential characteristics of Cloud computing are
1. On-demand self service where computing resources are provided in an automatic fashion without resorting to human interaction. 2. Broad network access 3. Resource pooling 4. Rapid elasticity 5. Measured service With these characterstics, the cloud community has extensively used the following service models for categorization: 1. Software as a Service(SaaS) 2. Platform as a Service(PaaS) 3. Infrastructure as a Service(IaaS) 4. Data storage as a Service(DaaS) In addition, four Cloud deployment models have also been defined. 1. Public Cloud 2. Private Cloud 3. Community Cloud 4. Hybrid Cloud Some relationships are identified while comparing the Cloud computing with Service-Oriented Computing and Grid Computing. The numerous challenges that are preventing the adoption of cloud computing are security, it’s costing and charging model, the definition of SLA and the decision of what to migrate. Interoperability is essential for Cloud Computing. The scope of interoperability here refers both to the links amongst different clouds and the connection between a cloud and an organization’s local systems. There are a number of solutions for different cloud service deployment models. 1. Intermediary Layer
2. Standard 3. Open API 4. SaaS and PaaS interoperability
Software Architecture and Mobility: A Roadmap Nenad Medvidovic, George Edwards Computer Science Department, University of Southern California, Los Angeles, CA 90089, USA
Summary: The key software architectural abstractions are components, connectors, their interfaces, configurations, and constraints on system structure, behavior, composition, and interaction. Traditional styles provide many design guidelines that can prove to be useful in mobile systems, which include component decoupling, avoiding shared memory, stateless components and interactions, and implicit invocation etc.
However all of them are partially suited for mobility and that suitability will be context dependent. Mobility paradigms or styles divide mobile systems into three categories: remote evaluation, code on demand, and mobile agent. Other proposed ideas were in terms of entities exchanged by sites, chemical abstract machine (CHAM), software connectors in mobile systems’ architectures, collection of potentially mobile modules (complets) with specific relationships.
Providing architecture-level support for migrating active components with active state, and understanding and quantifying the explicit software connectors’ cost remains an open issue. The objective of implementing architecture is simple: realize the principal design decisions and achieve the key intended non-functional system qualities as a result, hence support for effective implementation of mobile systems is critical. As with any large, distributed, long-lived systems, mobile systems are most effectively developed with the help of middleware facilities.
Mobile middleware platforms are grouped based on their particular foci, which include Traditional, Context-aware, Data sharing, tuple spaces middleware and architectural middleware. Deployment involves planning, modeling, analysis, and implementation. Extensible architecture description languages (ADLs), such as xADL provides the structural core of an architecture modeling and analysis approach that natively incorporates system deployment characteristics, called XTEAM.
Novel software architectures that are specialized for the domain of mobile systems are required because most of the assumptions, constraints, and goals of software development are radically shifted in the mobile setting. In this paper the authors highlighted three particularly compelling areas of research: mobile device management, context-aware mobile applications, and mobile robotics.
New experimental mobile device management systems simplify the provisioning, maintenance, and evolution of mobile services and integrate mobile technology with business processes. Software architectures for mobile device management are centered on the requirement for integration of heterogeneous platforms and services.
The software architecture of a mobile device management system also commonly includes components and user interfaces for defining, storing and executing policies and processes. Context-aware applications are made possible because of the integration of multiple context-sensing devices on mobile platforms like smartphones. Software architectures for context-aware mobile applications are commonly based on event-based, publish- subscribe architectural styles and middleware infrastructures.
Autonomous mobile robots carry out high-level goals by sensing their environment, planning responsive actions, and executing those plans. Software architectures for adaptive mobile robotics generally adopt a layered approach. At each layer, components are responsible for creating and executing plans that are beyond the capability of components at the layers below. Cloud computing allow service clients to have a single, location- and machine independent view of their applications and data, while freeing them from having to understand or manage the technology that implements those services.
References: 1. Kim,T., Choi, Y., Han, S., Chung, J.Y., Hyun, J., Li, J., & Hong, J.W.-K. (2012). Monitoring and Detecting Abnormal Behavior in Mobile Cloud Infrastructure. IEEE Network Operations and Management Symposium (NOMS), Maui, HI, April 2012, pp. 1303 – 1310.
2. Dilon, T., Wu, C., & Chang, E. (2010). Cloud computing: Issues and Challenges. The 24th IEEE International Conference on Advanced Information Networking and Applications (AINA), Perth, WA, April 2010, pp.27-33. 3. Medvidovic, N., & Edwards, G. (2010). Software Architecture and Mobility: A Roadmap. Journal of Systems and Software, Vol. 83, Issue 6, June 2010, pp.885-898. 1.