A Review on Mobile Cloud Computing D.Priyanka, Rajasekhar Boddu, B.Sunil Kumar, C.Srinivasulu Abstract--Cloud computing is an emerging technology that paved way for potential commoditization of computing resources. This technology is on top of virtualization that makes the cloud offerings affordable. With the advent of mobile and hand held devices and innovations with underlying mobile technologies and the ubiquitous nature of mobiles, cloud computing expands to mobile devices as well. This led to mobile cloud computing where mobile devices are associated with cloud computing and leverage benefits of cloud. As people of all walks of life are using mobile devices, the mobility feature of the devices can have tremendous impact on usage of cloud computing. There is steady growth rate projected with respect to mobile cloud computing in future. As mobile devices are energy and resource constrained, they are vulnerable to various security threats. Unless these threats are addressed, mobile cloud computing cannot be adapted easily. This paper throws light into the mobile cloud computing, its architecture, issues involved and solutions. The insights obtained through review of important papers can help in making well informed decisions with respect to mobile cloud computing and its applications in the real world. I. INTRODUCTION Cloud computing is the technology that realizes the dream of commoditizing computing resources in similar fashion to electricity and water. In fact cloud computing enables users to access huge computing resources. This new model of computing helps people and organizations to access computing resources in pay as you use fashion. Thus the model avoids the need for capital investment. It has got service models like Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS). Its deployment models include private cloud, public cloud, community cloud and hybrid cloud. Mobile Cloud Computing (MCC) is the cloud computing where mobile devices are involved. In fact MCC is nothing but the cloud computing that involves mobility. It involves mobile users whose storage and processing is done in cloud i.e. outside the mobile devices [1]. The architecture of MCC is as shown in Figure 1. Figure 1 Architecture of MCC [1] As can be seen in Figure 1, it is evident that the cloud computing facilities are being utilized by mobile devices. Mobile users get services from network operators. In turn the network operators are able to gain access to cloud services through Internet. There are many technologies that enable MCC. They are Web4.0, Cloudlets, Hypervisor, HTML5, 4G and CSS3. Mobile computing has plethora of advantages that include offloading of computations, executing applications remotely, remote processing, task migration, improved storage capacity and processing, improved availability and reliability, scalability, dynamic programming, ease of integration, multitenancy, mobile commerce, mobile learning, mobile healthcare, health cloud, telemedicine, mobile gaming and so on [1]. Businesses across the world grow faster with mobile computing. There are future projections that speak about the possible usage of MCC in future. Figure 2 presents the forecast pertaining to revenues on MCC in the period 2009 and 2015. ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2600
Figure 2 Forecast on revenues on MCC [1] As can be seen in Figure 2, it is well known that the overall revenues on MCC across the globe are increasing year by year in rapid pace. This is the indication that the mobile computing plays an important role to make mobile cloud computing a successful paradigm in future. According to Pragya and Sudha (2012), MCC has revolutionized the experience of mobile users as they can obtain plethora of cloud services on the go. In addition to this mobile computing technologies are growing rapidly making the cloud usage feasible and affordable. As many things are done outside mobile, the MCC causes increased bandwidth, storage and energy. Kovachev et al. (n.d) [3] provides comparison of MCC application models. The models help in understanding the true dynamics of MCC to bring effectiveness in cloud computing. There are certain challenges faced by MCC. They include limitations of mobile devices, communication quality, and the division of labor of application services. These limitations can be overcome by using technologies or methods like virtualization, task migration, upgrading bandwidth, and elastic application division [4]. Huang et al. [5] proposed a framework for MCC known as MobiCloud for enhancing communication quality in the MCC operations. Frenando et al. (2013) [6] provides a good survey of MCC and its related researches that have been carried out besides contrasting cloud computing and MCC. Dinh et al. [7] provides good architecture of MCC and its applications and approaches. They also provide insights into the issues in MCC and its solutions that are possible. Asrani [8] explored MCC platform for important applications like M-Commerce, M-Healthcare, and M-Gaming. It also throws light into the possible challenges of MCC and underlying solutions that can overcome the problems involved I MCC. Kumar and Lu [9] completely focused on energy efficiency of MCC as the mobile nodes outsource computing and storage. How mobile devices can save energy by offloading work was the main focus of their research. Miettinen and Nurminen [10] also did research on similar lines. They focused much on energy efficiency possibilities of MCC. This research is important as it finds the reasons why MCC is useful for mobile devices. As mobile devices are energy constrained and their network lifetime is affected by the storage and processing, offloading these activities certainly reduce the power consumption. Thus MCC is able to achieve energy efficiency. Mobile devices are vulnerable to various attacks. This might have its impact on MCC. Ko et al. [11] focused on security issues pertaining to MCC. This has very important aspect as the security issues cause problems in MCC. Security considerations need to be a continuous process as the overcoming of the issues can impact on the growth rate of MCC in the real world. Sahu et al. [12] also focused on issues with MCC and possible solutions that can be used to overcome problems and leverage the power of cloud and the ubiquitous nature of mobiles. The rest of this paper reviews important papers on mobile cloud computing that provide insights into the research area pertaining to the present state of the art besides knowing its implications on the real world. II. RELATED WORKS This section reviews some of the important papers which provide insights into the mobile cloud computing, its issues and solutions. The review also bestows various models in the mobile cloud computing besides other useful information. Energy efficiency of mobile clients in cloud computing Miettinen and Nurminen [10] focused on the mobile cloud computing research. Especially they considered energy efficiency problem. The reason behind this is that mobile devices are energy constrained and they lack sufficient resources. Optimum utilization of the resources is the fundamental aspect with respect to mobile cloud computing. Mobile devices consume less energy when they can offload storage and processing operations to cloud. However, the energy savings thus made should be more than the additional communication cost incurred for associating with a cloud. In [10] detailed analysis is made with respect to critical factors that affect the energy efficiency of the mobiles connected to a network. They also ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2601
presented some sort of measurements to measure the characteristics of mobile or hand held devices. The measures are with respect to the balance between the remote and local computing that can affect nodes consuming energy. Their paper also demonstrates a concrete example to describe the process of saving energy in mobile cloud computing. Miettinen and Nurminen made an energy trade-off analysis to find whether mobiles involving in mobile cloud computing can save energy. The energy required for local computations is represented as E local while the energy required to outsource storage and processing operations to cloud is represented as E cloud. Ideally the beneficial situation is achieved with the equation presented below [10]. E cloud < E local It does mean that the energy consumed when the storage and processing operations in the local device should be less that of the cloud. When this ideal situation is met, energy saving is possible. Considering D is the amount of data to be transferred and C is the computational power required in order to transfer the data. Provides these details the following equation is used to compute E cloud and E local. E cloud = D/D eff Nokia profiler is used to measure energy consumption. As per the energy efficiency formulae, the local performance of the devices is measured and recorded. Afterwards, the cloud performance of the devices will be recorded and compared. Experiments and Future Work When data is transferred, the energy consumption is depending on the bit rate. The more bit rate is the less energy consumption is. It does mean that energy efficiency is achieved by using high bit rate channels. Energy efficiency in case of cellular network actually depends on the data to be transferred and bit rate. Figure 3 illustrates the bit rate and energy efficiency dynamics. E local = C/C eff The computational characteristics for two kinds of mobile devices are presented in Table 1. The devices used for experiments are Nokia N810 and Nokia N900. Figure 3 Energy per bit for N95 WLAN and 3G As can be seen in Figure 3 it is evident that there is relation between bit rate, data transfer and energy efficiency. Energy efficiency is sensitive to bit rate in case of cellular networks when compared to WLAN. Table 1 Energy characteristics of local computing [10] As can be seen in Table 1, energy consumption details of two devices are presented with respect to local computing while the wireless computing results for the same is presented in Table 2. Figure 4 Illustrate traffic pattern effect for N95 WLAN ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2602
As can be seen in Figure 4, it is evident that dynamics are presented for bursty traffic sources and smooth traffic sources with respect to Nokia N95 device in WLAN. When bit rate is 1W it causes smooth communication while the bit rate 0.6W it causes bursty communication. Figure 5 Illustrates power consumption for viewing PDF As can be seen in Figure 5, example measured curves are presented that reflect the power required by the local viewer with respect to the device N900 WLAN. The remote cases needed higher average power. Interestingly the remote case causes less power consumption as the execution time is shorter. When compared with WLAN, 3G network causes more energy consumption. Future work includes studying end to end chain for mobile energy consumption with respect to mobile cloud computing. Securing Mobile Cloud Computing Communications Huang et al. [5] proposed an architecture for mobile cloud computing. Their proposed architectural framework is named MobiCloud. The MobiCloud is the combination of traditional services and modern services. The framework makes use of trust management, risk management and secure routing concepts to overcome corresponding issues. With this framework in place, it is possible to deploy and use a new class of applications to leverage the power of cloud computing through small hand held devices [5]. Figure 6 presents the architectural overview of MobiCloud. Figure 6 Architectural overview of MobiCloud [5] As can be seen in Figure 6, it is seen that cloud and mobile networks are integrated. MobiCloud service provisioning, MobiCloud trust management server, MobiCloud Resource Manager, and MobiCloud service or application store are the important components involved. The solid lines in the architecture include direct links while the dotted lines include indirect links. The secure communication is made possible with SSL connections. Secure Sockets Layer is the protocol that makes the communications in mobile cloud computing secure. The MobiCloud is using SSL connections to ensure that the communications are encrypted while sending and receiving. Mobile Ad Hoc Network is also involved as part of architecture [5]. Attribute Based Encryption The security to communication is made using a special encryption scheme known as attribute based encryption. The scheme has various things such as attributes from A 1 to A n and private key components, and secret sharing threshold gates besides a key for data encryption [5]. The encryption scheme is visualized as shown in Figure 7. Figure 7 Illustrates attribute based encryption [5] ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2603
As can be seen in Figure 7, the attribute based encryption scheme has encryption mechanism that makes use of attributes. The attribute based encryption is used for efficient key management scheme for access control. With this scheme and the MobiCloud framework it is possible to conceive many application scenarios. The scenarios include inter-operable communications, efficient communications, security scenarios, isolation scenarios, and delay tolerance communication scenarios. The MobiCloud can also be improved in future to support damage recovery, secure isolation of fine-grained resource isolation. Operation delay is the real time issue with respect to performance of MobiCloud. The applications that run in this cloud are to be designed carefully keeping the framework in mind and thus they can have the desired features. The main services of MobiCloud includes monitoring services that observe node status information while the on-demand services take care of many services that arise on the fly. Other important service is known as advising service which emulates MANET activities for analyzing post events [5]. Experiments and Future Work Experiments are made with MobiCloud with MANET involved in mobile cloud computing. The data transmission interval and routing overhead ratio are observed in MANET. The protocols used for experiments include DSR and AODV [5]. The data transmission intervals include 10, 100 and 1000. The experimental results are presented in Figure 8. III. APPLICATION MODELS FOR MOBILE CLOUD COMPUTING Dejan et al. [3] provide a collection of application models that can leverage the concept of mobile cloud computing. The models are designed systematically. They observed that decoupling of service delivery from the methods or techniques, reducing cost, processing cost, software cost over Internet, and delivery of services are the foundation of the cloud computing. However, it is challenging task to achieve all these benefits in case of mobile cloud computing. With respect to current status in mobile applications, there are two kinds such as Online and offline applications. The mobile experiences and environments are changing rapidly. Based on this the online and offline applications should get updated to resolve issues [3]. Novel Application Models for Mobile Cloud Computing Various models are described in [3] of applications models for mobile cloud computing. They include augmented execution model, elastic partitioned model, application mobility and ad-hoc mobile clouds. Figure 8 Routing overhead ratios vs. data transmission interval for DSR and AODV [5] As can be seen in Figure 8 it is evident that the routing overhead ratio is more with AODV protocol when compared with DSR protocol. The future work include damage recovery which does mean the loss of mobile devices, and fine grained resource and security isolation which brings about control and efficiency in mobile cloud computing [5]. Figure 9 Models for augmented execution [3] As can be seen in Figure 9, it is evident that there are many models for augmented execution with respect to mobile cloud computing. The models include outsourcing primary functionality, background augmentation, mainline which is in between primary and background, hardware which makes use of virtual machines, and multiplicity model is used for parallel execution of tasks [3]. ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2604
Computer Interaction (HCI), networks and systems. This will help in enhancing the utility of mobile cloud computing in future. Figure 10 Reference architecture for elastic applications [3] As can be seen in Figure 10, it is evident that the application in this architecture is split into three components. They are UI component, weblet component and manifest component. The UI component is related to user interface. The functional software entities that are invoked from UI are known as weblets. The cloud elasticity service has application manager, cloud sensing, cloud manager, and cloud fabric interface. The manifest component contains the details of the UI and the weblets available for the application. In [3], many existing and proposed application models with respect to mobile cloud applications. The summary of the findings is presented in Table 3. Table 3 Summary of comparison of existing and proposed mobile cloud models [3] As can be seen in Table 3, it is evident that there are architectures for mobile computing that have been proposed and some of them already implemented. The technologies used for the mobile cloud computing include Hadoop, web services, distributed file system, restful web services, C#, OSGi, Java and HTML 5 [3]. Future Work An important direction for future work is to continue with interdisciplinary research pertaining to Human IV. CONCLUSION This paper provides insights pertaining to mobile cloud computing. It covers issues pertaining to mobile clients in cloud computing and possible techniques to resolve issues. One such issue is energy efficiency. Since the mobile devices are energy constrained, they need to be utilized properly. It does mean that the integration with cloud should bring about energy efficiency as the work of the mobiles is offloaded to cloud. The operations such as storage and processing are outsourced to cloud. The energy saving due to cloud computing should be greater than the communication overhead. Towards this end many models came into existence. This paper focuses more on energy efficiency in mobile cloud computing, securing mobile cloud computing applications and various application models that are best used with mobile cloud computing. The paper also presents results of the work of those papers and provides possible future directions in the area of mobile cloud computing. REFERENCES [1] Anonymous. (n.d), Mobile Cloud Computing, p1-47. [2] Pragya Gupta and Sudha Gupta. (2012). Mobile Cloud Computing: The Future of Cloud. International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering. 1 (3), p1-12. [3] Dejan Kovachev, Yiwei Cao and Ralf Klamma. (n.d). Mobile Cloud Computing: A Comparison of Application Models., p1-8. [4] Han Qi and Abdullah Gani. (n.d). Research on Mobile Cloud Computing: Review, Trend and Perspectives., p1-8. [5] Dijiang Huang, Xinwen Zhang, Myong Kang and Jim Luo. (2010). MobiCloud: Building Secure Cloud Framework for Mobile Computing And Communication. IEEE, P1-8 [6] Niroshinie Fernando, Seng W. Loke and Wenny Rahayu. (2013). Mobile cloud computing: A survey. Future Generation Computer Systems, Elsevier. 29 (.), p84-106. [7] Hoang T. Dinh, Chonho Lee, Dusit Niyato, and Ping Wang. (n.d). A Survey of Mobile Cloud Computing: Architecture, Applications, and Approaches. Wireless Communications and Mobile Computing - Wiley, p1-38. ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2605
[8] Priyanka Asrani. (2013). Mobile Cloud Computing. International Journal of Engineering and Advanced Technology (IJEAT). 2 (4), p1-4. [9] Karthik Kumar and Yung-Hsiang Lu,. (2010). CLOUD COMPUTING FOR MOBILE USERS: CAN OFFLOADING COMPUTATION SAVE ENERGY?. IEEE, P51-57. [10] Antti P. Miettinen and Jukka K. Nurminen. (n.d). Energy efficiency of mobile clients in cloud computing, p1-7. [11] Soeung-Kon(Victor) Ko1, Jung-Hoon Lee and Sung Woo Kim. (2012). Mobile Cloud Computing Security Considerations. Journal of Security Engineering. 9 (2), p1-8. [12] Deepti Sahu, Shipra Sharma, Vandana Dubey, Alpika Tripathi. (2012). Cloud Computing in Mobile Applications. International Journal of Scientific and Research Publications. 2 (8), p1-9. AUTHORS Hyderabad, India. Presently working as Lecturer in Department of Software Engineering, College of Computing and Informatics(CCI), Haramaya University, Ethiopia. He is having 5 years of teaching experience. He is a member in UACEE, IACSIT, CSTA, IAENG, ACM. His research interests are Security, Data Mining and Cloud Computing. Mr. B. Sunil Kumar working as Assistant Professor in Department of Computer Science & Engineering in Jawaharlal Nehru Institute of Technology, Hyderabad, India. I am having 5 years of Teaching Experience. My interested subjects are Web Technologies, Linux Programming, Mobile computing, cloud computing, Computer Networks, Operating System, Computer Organisation, Java, C, and C++. Mrs. D. Priyanka completed M.Tech from Vignana Bharathi Institute of Technology Affiliated to JNTUH. I am Presently working as Assistant Professor in Department of Computer Science & Engineering in Vignana Bharathi Institute of Technology, I am having 6+ years of Teaching Experience. My interested subjects are Formal Languages and Automata Theory, Data Base Management Systems, Computer Networks, Mobile computing, Digital Logic Design and Computer Organization, Mr. C. Srinivasulu pursing M.Tech(cse) from VignanaBharathi Institute of Technology affiliated to JNTU-H. Completed B.Tech(cse) from Visvodaya Institute of Technology & Science, Kavali (formerly affiliated to JNTU-H). I am currently working as a Team Lead with NTT DATA. I am having 1.7 years of experience in Teaching and 8+ years of experience as a Java Developer. My interested subjects are Cloud computing, Java, C++, Web Technologies, Web Services, Computer Networks. Mr.Rajasekhar Boddu completed M.Tech and B.Tech from JNTU ISSN: 2278 7798 All Rights Reserved 2014 IJSETR 2606