The service-oriented architecture (SOA) of MCC consists of three layers [3], as shown in Figure 3.2.
SOA of MCC consists of the following components [3]: 1. Mobile network
2. Internet service 3. Cloud service
1. Mobile network: A mobile network contains mobile devices and network opera- tors. Mobile devices may be smartphones, PDA, satellite phones, laptops, and so on. They are connected to the network operator via the BTSs (base transceiver stations), access points, or satellites. They establish and control the connection between the functional interface between mobile device and network operator. A mobile device’s request and information, such as ID and location, are transmit- ted to the central processor and servers of the network providers. Here, opera- tors provides various services such as AAA (authentication, authorization, and accounting) based on the HA (home agent) and subscriber data stored in database. 2. Internet service: Internet service plays the role of a bridge between the mobile
network and cloud. Subscriber requests are delivered to the cloud via a high-speed Internet service. Using wired connections or advance 3G or 4G technologies such as HSPA, UMTS, WCDMA, LTE, and so on, the user can get seamless service from cloud. Cloud Application server Mobile network Server HA AAA Central processor Database BTS Access point
Mobile device (smartphone, laptop, PDA) Satellite Cloud controller Data center Internet service FIGURE 3.2
3. Cloud service: After getting all the requests from the users, the cloud controller pro- cesses the requests and provides service to them according to their requirements. Cloud has a few service providing layers, as shown in Figure 3.3. These service layers are discussed as follows:
a. Data center layer: Data center provides the hardware facilities and infrastruc- ture for the cloud. In a data center, there are several servers connected with high-speed networks and high power supply. Normally, they are built in less populated places with a low risk of disasters.
b. Infrastructure as a service: IaaS resides on the top of the data center layer. It provides storages, servers, networking components, and hardware to its cli- ents on a “pay as you use” basis. It has an elastic nature, so infrastructures can be expanded or shrunk dynamically according to user demands. Amazon EC2 and S3 are examples of IaaS.
c. Platform as a service: PaaS provides an integrated environment or platform for users to build, test, and deploy several applications. Any kind of platform such as Java, .NET, PHP, and so on, are available. Google App Engine, Microsoft Azure are examples of PAAS.
d. Software as a service: SaaS is a software delivery model provided by applica- tion service providers (ASPs). Software and the associated data are centrally hosted on the cloud. SaaS can provide numerous kinds of software solutions such as CRM, ERP, MIS, HRM, and so on, on demand without any dedicated installation in client site.
This way, in mobile cloud computing, data storage and computations are moved into the cloud, and the user gets seamless, on-demand service without having to worry about bat- tery life or the processing power of mobile devices.
Cloud clients (web browser, thin clients, terminal, emulator, etc.)
SaaS (CRM, e-mail, virtual desktop, communication, games, etc.) PaaS (execution runtime, database, web
server, development tools) IaaS (virtual machine, servers, storages,
local balancer, networks, etc.) Data center
FIGURE 3.3
3.3.2 Agent–Client Architecture
In this architecture, mobile devices are not connected to the cloud directly. They are con- nected to the cloud via some agents such as femtocell [4], cloudlet [5], or both, as shown in Figure 3.4. MCC is all about wide area network (WAN) and cloud. Usually, clouds are sit- uated at a long distance from the users, so there are chances of delay or cost inefficiency. These agents can fulfill the user demand with high bandwidth, low latency, and low cost. Only when the demands are not fulfilled by the agents will the request go to the cloud. The architecture is shown in Figure 3.4. Here, M1, M2, M3, and M4 are the mobile devices that are not directly connected to cloud but via agents such as femtocell, cloudlet, or both.
Cloudlet: Cloudlet [5] is a resource-rich computer or cluster of computers that is trusted and well connected to high-speed Internet and available to mobile devices. When a user does not want to offload any task directly to cloud due to delay or cost, he or she can offload it to the nearest available cloudlet. If the device cannot find any cloudlet available, then it will send its request to the cloud or, in the worst case, complete the task with its own resources. Thus, a user gets real-time response by low-latency, one-hop, high-bandwidth, and low-cost access to cloudlet.
Femtocell: Femtocell [4], well known as the “home base station,” is the smallest version of traditional macrocells. Femtocells are deployed indoors to provide good coverage [6–8]. Mobile devices are connected to femtocells, and femtocells
Agents Mobile devices M1 M2 M3 M4 Femtocell, cloudlet Cloud FIGURE 3.4 Client–agent architecture of MCC.
are connected to the mobile network with residential DSL, cable broadband, optical fibers, or wireless last-mile technologies [9–11]. Femtocell access points can implement cellular technologies such as UMTS/HSPA/LTE and mobile WiMax. So, they can provide seamless 3G and 4G service to the user and can be used to connect mobile devices with a cloudlet or cloud. As it gives higher bandwidth, the user will face very little latency to offload tasks to a cloud or cloudlet.
3.3.3 Collaborative Architecture
Nowadays, smartphones are operated independently using their local computing, sensing, networking, and storage capabilities. When data are shared with other devices through a centralized server or cloud, it requires expensive upload and download. It can be avoided by collaborative computing [3]. In this architecture, resources of a mobile device are used by considering the device as a part of a cloud. The cloud server may be the controller and scheduler for collaboration among the devices. By combining smartphone data and computing, a smartphone cloud can be generated. Mobile applications can utilize these resources of the smartphone cloud, so processing of mobile data in a smartphone cloud can remove the bottlenecks of global network and the limitation of offloading data to a remote server. In Figure 3.5, the collaborative architecture of MCC is shown, where M1, M2, M3, and M4 are the mobile devices, and M2, M3, and M4 have formed a smartphone cloud and M1 is using it.
Smartphones such as androids or i phones support cloud computing. Hyrax [12] is a platform based on Hadoop, and supports cloud computing on an Android smartphone. Hyrax allows client applications to conveniently utilize data and execute computing jobs on networks of smartphones and heterogeneous networks of phones and servers. By scal- ing with the number of devices and tolerating node departure, Hyrax allows applications to use distributed resources abstractly and exactly as the cloud. Using an x86 virtual machine, different applications can run from the BOINC server to an apple i-Phone, which integrates the grid computing framework. By this technique, an i-Phone can support the CC approach. M1 M4 M3 M2 Controller/ scheduler Smartphone cloud by collaborating resource with each other
FIGURE 3.5
3.4 Platform and Technologies