PDF Download
PubReader
Export Citation
Email the Author
Share Facebook
Share Twitter
Cited by 0 Article
Metrics (View:4,307)
I. INTRODUCTION
Cloud Computing consists a technology of internet services, providing remote use of hardware and software. As a result, the users of Cloud Computing could have access to information and data from any place at any time. In recent years, giant companies of the IT and software sectors investigate in survey the services of Cloud Computing. Furthermore, another technology which generated relaying on Cloud Computing is Mobile Cloud Computing. Mobile Cloud Computing based on the concept of the “Cloud” and provides any type of information and data by no matter where and when, through mobile devices. Particularly, Mobile Cloud Computing is defined as “the integration of Cloud Computing and Mobile technology in order to make any type of mobile devices resourceful in terms such as computational power, memory, storage and energy”. Regarding the usage of Cloud services in Mobile devices many types of services could be processed through it. Thus, high quality media could be transmitted through Cloud environment progressed in applications which were installed and operated in Cloud. Considering this Mobile Cloud Computing, and also Cloud Computing in general, could be settled as a base technology to operate other technologies, such as Internet of Things, and consequently to be accomplished an integration of Cloud and IoT [1] [2] [3] [4] [5].
Internet of Things (IoT) is “a system of interrelated computing devices, mechanical and digital machines, objects that are provided with unique identifiers and the ability to transfer data over a network without requiring human-to-human or human-to-computer interaction”. Based on this definition, in IoT environments could be established transmission networks for information and data produced by applications running on it. Some examples of the possible application scenarios in Internet of Things would be the domotics, e-health, assisted living, and enhanced learning. Thus, all these could lead us to the conclusion that IoT can be used in order to operate applications that provide digital media [6] [7] [8] [9].
Digital media could be produced by surveillance video systems. In most cases, surveillance is used by people in order to influence, manage, direct, or protect them, by using sensors and cameras or other compatible devices. Ordinarily, digital media format could be defined as the high quality video format produced by surveillance cameras. Those three aforementioned technologies could be combined and operate consolidated with the aim to have a more efficient network. This network would be based on Cloud and IoT environment, for transmitting high quality data, such as digital media [10] [11] [12] [13].
The rest of the paper is organized as follows. In section 2 there is a review of the related research in the area, which deals with the integration of Cloud and IoT, as these are used as a basis for data transmission. Section 3 presents and illustrates the simulation method of this work. In Section 4 demonstrated the simulation and the experimental results recorded. Finally, section 5 provides the conclusions of the current work and offers new possibilities for the development of future work.
II. RELATED REVIEW
In this section we present related works to our research. By studying the areas of collection, delivery, management, and analysis of large-scale data (Big Data), it is concluded that data centers are responsible for everyone since everything that happens to them will affect us all. So in [14] is presented, through several open source platforms (e.g. Arduino), the implemented data center environmental monitoring system. The system’s architecture design is the implementation key of success. With the implemented design and through the Internet we can identify in real-time the system logs and status. As an extension of that system is proposed the monitoring of real-time Big Data through HTML5 charts.
A region based approach is presented in [15], where Jun-Ho Huh and Kyungryong Seo discuss about efficient power consumption through the technologies and techniques of Smart Grid. The main focus area of this research is the Programmable Logic Controller (PLC) technology in conjunction with power lines for the transmission of data in a network since it is an efficient and low-cost solution for efficient metering. The results from the analysis of the implemented PLC-based power-aware home network system design, using OPNET Modeler 14.5 PL8, were analyzed and compared to those of IEEE 802.11 WLAN MAC.
In another region based approach [16], a novel power-aware routing protocol is proposed. With this protocol and a mechanism which controls the delays, researchers maximize the lifetime of every node in an Ad-hoc network system. NS-2 was the simulator used for the verification of the network.
As is known, with the blooming of Big Data, the Cloud Computing (CC) also blossomed. However, there are open issues and challenges in this technology, for some of which are provided solutions by several researchers. In [17] there is an attempt to solve the problems of ignoring the content of multimedia and the difficulty in implementing solutions for the cloud platform. So, researchers proposed a new distributed multimedia programming model for its implementation on different service platforms and different multimedia applications. Also, an algorithm for decision making by users, based on local information, is also proposed.
One of the most challenging fields of Multi-clouds is the efficient workflow scheduling. So, in [18] researchers proposed an algorithm (Multi-Clouds Partial Critical Paths, MCPCPP) for Big data scheduling in Multi-clouds. This algorithm reduces the workflows’ execution costs. At the same time, the algorithm indulges the determined restriction deadline. From the results it is concluded that the proposed algorithm is promising.
Moreover, in [19] researchers talk about the networking perspectives of three popular applications. These are YouTube, Facebook, and WhatchUp. Researchers analyzed the traffic and the network infrastructure which hosts these data flows. The DBStream platform was used to analyze the large amounts of data. Solutions for traffic monitoring, analysis, and services of cellular networks have also been proposed and discussed.
The Big Data are usually transmitted from the data production center to the remote environment so that it can be provided the analysis of these large amounts of data. The multiple bandwidth reservation requests issue is discussed with the use of a High-Performance Network (HPN) in which succeeded with the best average transmission. So, in [20] have been proposed two efficient and high-speed algorithms with polynomial time complexity. The algorithms were compared with two others and from the experimental results were both verified for their advanced performance.
The pervasive network services outstretch into ubiquitous computing environment. The users to get the services they need, they have to share personal and private information. To avoid the exposure to various attacks (eg. eavesdropping) researchers proposed in [21] a security scheme to secure the communications. The authentication scheme guarantees reliability and availability by securing the remote access in Pervasive Computing Environment (PCE). The scheme provides security and convenience to the users.
III. SIMULATION METHOD
Based on previous works, with the aim to succeed a new type of network, which could provide more efficient data transmission a simulation tool was used. The simulation platform that used in this work is CloudSim. This simulation platform operates in the Eclipse environment, in java programming language. In CloudSim using the logic of a virtual system, and thus virtual management, we create Virtual Machines (VMs) [22] [23] [24] [25].
Figure 1 demonstrates how a user could interact through a Global Manager (application software) to a number of Cloud Virtual Machines. More specifically, each Cloud Virtual Machine consists of a Local Manager which interacts with a Virtual Machine Manager (VMM), and through the VMM established a communication path with the various individual VM devices. Each VM is connected to four sensors, from which it receives the data it then transmits to VMM. For each Cloud Virtual Machine there is a Physical Node which connects it to the network.
Table 1 lists the two types of Virtual Server Configuration for the Cloud which have been used for the simulation. In this work we used 400 Virtual Servers of the model Dell PowerEdge T110 and 400 Virtual Servers of the model HP ProLiant ML 110 G5.
Table 2 depicts the rate of Watt Power Consumption from the information produced and transmitted from each type of Cloud Server, either Dell PowerEdge T110 or HP ProLiant G5. As we can observe, when the rate of consumption of watts increases, both the transmission of information increases.
| Consumption in % | Dell PowerEdge T110 | HP ProLiant ML110 G5 |
|---|---|---|
| 0% | 86 | 93.7 |
| 10% | 89.4 | 97 |
| 20% | 92.6 | 101 |
| 30% | 96 | 105 |
| 40% | 99.5 | 110 |
| 50% | 102 | 116 |
| 60% | 106 | 121 |
| 70% | 108 | 125 |
| 80% | 112 | 129 |
| 90% | 114 | 133 |
| 100% | 117 | 135 |
Table 3 shows the four types of VM that created and used for the simulation method. Each type had differentiated characteristics in order to be studied a wide range of results.
Table 4 demonstrates the Cloudlet Parameters which represent the volume of data used in a network in association with IoT technology. With the aim to proceed at a better simulate of high quality data, referring to digital data, we used large sizes in MB.
| Length (MB) | File size (MB) | Output size (MB) |
|---|---|---|
| 5000 | 5000 | 5000 |
Subsequently, for further simulation procedure, we used Cooja Contiki simulator with the purpose of personalizing and extracting our network data in an environment with a defined topology.
IV. EXPERIMENTAL RESULTS
Having already tested the performance of the network we created in CloudSim, we perform a simulation in Cooja Contiki, where we tried to map the same network scenario, but also studied more aspects of this network.
We implement a network topology of a small part of the previous scenario where examining a single network segment. Namely, we examined the communication and the efficiency of data transmission in a VMM, which includes in its range five VMs.
Figure 2a and Figure 2b show the topology of our proposed network. As we already mentioned, each separate part of our network consists of one VMM and five VMs. In each VM are connected four sensors. The range of the VMM contains only the five VMs, and the range of every VM contains only its four sensors. According to this, we observe that each VM contains only four sensors in its range, so as not to be inserted from the range of other VMs.
Figure 3 (a) and Figure 3 (b) demonstrates the Power Consumption of the Network. Figure 3 (a) demonstrates the average Power Consumption, where we can observe that LPM’s power (red color) remains almost constant over time, as well as CPU’s power (blue color). In contrast, the Radio listen’s power (green color) and a little less the Radio transmit’s power (yellow color) where there is a greater variation in Power Consumption. Figure 3 (b) demonstrates the Instantaneous Power Consumption, where, same as before, we can observe that LPM’s power (red color) remains almost constant over time, as well as CPU’s power (blue color). And also, in contrast again, the Radio listen’s power (green color) and a little less the Radio transmit’s power (yellow color) where there is a greater variation in Power Consumption. In Instantaneous Power Consumption we observe that there is a big difference as regards the variation of the Radio transmit’s power, compared with Average Power Consumption, as there are momentary fluctuations in the change in energy consumption during transmission.
Figure 4 shows the transmitted packets which have been received per node. As we can observe, in most cases and almost every time all the nodes received the same number of packets. In addition to this, we can conclude that there are no duplicated packets received.
Figure 5 demonstrates the packet transmission procedure through all the motes we used for the simulation in Contiki. With the word ‘mote’ is defined the node in ‘Contiki language’. Through Figure 5, we can see that during the simulation process transmission, there are array packages with large size (large-scale data).
V. CONCLUSION
Due to the blooming of IoT in CC which takes part in the last years, the there is a need of discovering new methods of technological support in many sciences by the researchers. As part of these researches, in this work, with the aim to achieve a type of network that will provide more intelligent media-data transfer, we have studied new technologies, and the use of various open source tools, such as CC analyzers and simulators. Tool like these are useful for studying the collection, the storage, the management, the processing, and the analysis of large volumes of data. Furthermore, the simulation platform used is CloudSim and operates on Eclipse environment. Thus, after measuring the network performance with CloudSim, we use the Cooja emulator of the Contiki OS in order to confirm and access more metrics and options. As a result, we implemented a network topology from a small section of the script of CloudSim with Cooja, so that we can simulate a single network segment. The results of the experiment show that there are not duplicated packets received.
Finally, as future research, we suggest a further examination of the simulation analysis of the network performance in CloudSim simulator, and other simulation platforms, with the aim to have a better and improved contribution of the technology of Internet of Things with the additional ‘help’ of the Cloud Computing technology for the purpose of better transmission of high quality data. This research could be a start point for better and more efficient media data transmission.
