2013 IEEE International Conference on Cloud Computing Technology and Science
Providing Desirable Data to Users when Integrating Wireless Sensor Networks with Mobile Cloud Chunsheng Zhu∗ , Victor C. M. Leung∗ , Hai Wang† , Wei Chen‡§ , Xiulong Liu¶
∗ Department
of Electrical and Computer Engineering, The University of British Columbia, Canada Email: {cszhu, vleung}@ece.ubc.ca † Department of Computer Science and Engineering, Pohang University of Science and Technology, South Korea Email:
[email protected] ‡ College of Computer Science and Technology, China University of Mining and Technology, China § State Key Laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, China Email:
[email protected] ¶ School of Computer Science and Technology, Dalian University of Technology, China Email:
[email protected] figurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction [4] [5]. Inheriting the advantages of CC, mobile cloud computing (MCC) is a novel technology to enable the mobile users to utilize the powerful storage and processing ability of CC to alleviate the burden (e.g., battery lifetime, data storage capacity, data processing power) of the mobile devices. This could further enable a lot of new mobile applications (e.g., mobile cloud commerce, mobile cloud learning, mobile cloud healthcare, mobile cloud gaming) [6] [7]. For example, traditional mobile learning based on electronic learning and the mobility of mobile devices may encounter issues like high storage cost, low processing speed and limited education resources on mobile devices. With the cloud which could store the data and process client requests via wireless network on a smart phone, MCC could offer much richer services as well as faster processing speed to both learners and teachers. Motivated by the potential applications when integrating WSNs and MCC, the integration of WSNs and MCC becomes a hot research topic [8] [9] [10] [11] [12] [13]. Particularly, the integration of WSNs and MCC could enable the mobile users to obtain the sensory data information of WSNs. For instance, as shown in Fig. 1, the sensory data (e.g., traffic, temperature, humidity) collected by the deployed WSNs can be continuously transmitted to the cloud. Then the mobile users can access these real-time traffic, temperature or humidity information by issuing data requests to the cloud. However, after integrating WSNs and MCC, the following challenges arise. • From the mobile user side, there are the following two issues. First, sometimes the data required by the mobile user may have a preference. For example, the mobile users may be particularly interested in the sensory data from one area of the WSN and want to obtain more data from that area. However, the WSNs are not aware of this and still transmit the same amount of sensory data as before. Second, sometimes the data required by the
Abstract—Wireless sensor networks (WSNs) receive a lot of attention because of their great potential in monitoring the physical or environmental conditions of military, industry, and civilian. Moreover, mobile cloud computing (MCC) is widely focused, as they can greatly alleviate the hardware limit of mobile devices as well as enable a lot of new mobile applications. All these make the integration of WSNs and MCC a very hot research topic. In this paper, we first observe a context nonawareness issue between mobile user and WSNs, which affects the mobile user obtaining the desirable data when integrating WSNs and MCC. Then focusing on solving the context nonawareness issue to provide desirable data to mobile users, we propose a novel framework for integrating WSNs and MCC. The proposed framework performs data recommendation, data prediction as well as data traffic monitoring in the cloud to obtain the data feature information required by the mobile users and potential status of WSNs. Then these user data feature information and potential WSNs status information are utilized to optimize the deployment of WSNs and check the status of WSNs. This could in turn offer the desirable data to the mobile users. Extensive evaluations also validate the effectiveness of the proposed framework. Keywords—wireless sensor networks; mobile cloud computing; integration; data; deployment
I. I NTRODUCTION Wireless sensor networks (WSNs) are distributed sensor networks, consisting of spatially distributed autonomous sensors which could sense the physical or environmental conditions (e.g., temperature, sound, vibration, pressure, motion, etc.). Since the proposal of WSNs, they are widely focused as they have great potential in areas of military, industry, and civilian (e.g., battlefield surveillance, battle damage assessment, industrial process and environmental monitoring, etc.) [1] [2] [3]. For instance, by deploying sensor nodes in a battlefield, the condition of critical terrains, approach routes, paths and straits can be continuously monitored by these sensor nodes thus the activities of the opposing forces can be closely watched by surveillance center without the involvement of physical scouts. Furthermore, cloud computing (CC) is a model for enabling convenient, on-demand network access to a shared pool of con978-0-7695-5095-4/13 $31.00 © 2013 IEEE DOI 10.1109/CloudCom.2013.86
607
Mobile device 1
Mobile device 2
Mobile device 3
Sen d da t Rep ly da a reque st ta re ques s ts
Data tranmission
Cloud Computing Platform
Wireless sensor networks (WSNs) Fig. 1.
•
Application of integrating wireless sensor networks (WSNs) and mobile cloud computing (MCC)
mobile user may have a trend to been changed but the WSNs still gather the same data scope. For instance, the mobile users may tend to get data from the area with higher temperature or higher humidity after a certain time. Furthermore, from the WSNs side, there are the following two problems. First, sometimes the data required by the mobile user may have a trend to excess the threshold value to be considered safe indicating that the WSNs may be likely to collect the dangerous data. Second, sometimes the data required by the mobile user may be missing or there are too much sensory data available for the mobile user due to various factors (e.g., sensor is compromised).
by the WSNs is not the data mobile users require and WSNs may already are compromised or under destruction. Targeted to solve the context non-awareness problems mentioned above, we propose the following framework. In this framework, we utilize the cloud to perform data recommendation, data prediction as well as data traffic monitoring, since mobile users may have some preferences about the data from mobile user side and WSNs may stay in an undesirable status from the WSNs side. Particularly, data recommendation targets to solve the mobile user data preference problem. Data prediction is utilized to mitigate the mobile user data trend and WSNs collecting dangerous data issues. Data traffic monitoring is performed to alleviate WSNs data missing or too much WSNs data problem. With that, we could use these information to optimize the deployment of WSNs and check the status of WSNs. For instance, we can add or remove the nodes in WSNs and check whether the WSNs are in a dangerous or missing data state. We aim at providing desirable data to mobile users when integrating WSNs with MCC.
In this paper, we define these problems as the context nonawareness issue between the mobile user and WSNs during the integration of WSNs and MCC. And it is obvious that the context non-awareness issue can strongly prevent the mobile users from obtaining the data they desire. This context nonawareness issue also degrades the performance of WSNs which act as the data source for MCC, as the data transmitted
The main contributions of this paper are summarized as follows.
608
First, this paper observes the context non-awareness issue between the mobile users and WSNs during the integration of WSNs and MCC. This context non-awareness issue is ignored by current frameworks integrating WSNs and MCC. • Second, this paper proposes a framework to mitigate the context non-awareness problems identified in this paper for integrating WSNs and MCC. Data recommendation, data prediction and data traffic monitoring are performed by the cloud to achieve the data preferences and data feature information of mobile users as well as the potential status information of WSNs. • Last, a simplified prototype of our proposed framework is implemented to verify the effectiveness. Evaluation and analytic results show that the proposed framework can greatly favor the mobile users to obtain the data they desire regardless of the context non-awareness issue. The rest part of this paper is organized as follows. Section II reviews the related work. The proposed architecture to integrate WSNs and MCC is shown and analyzed in Section III. The evaluation is conducted in Section IV and this paper is concluded in Section V.
large-scale sensory data processing with cloud. It utilizes the Hadoop Distributed File System (HDFS) and HBase to store the sensory data. Hadoop MapReduce is used for parallelly processing the sensory data. Finally, [18] extends WSNs into the cloud using Amazon web services. It mainly argues that the WSNs could offload some resource-intensive tasks to the cloud. And it demonstrates that Amazon web service can meet the dynamic computational needs of environmental applications of WSNs. The sensory data required by the mobile users is not considered. To the best of our knowledge, our work is the first to analyze the data required by mobile users and collected by WSNs first and then utilize these data to optimize the deployment of WSNs and check the status of WSNs.
•
III. P ROPOSED FRAMEWORK Fig. 2 shows the proposed framework to integrate WSNs and MCC and Fig. 3 shows the flowchart of the proposed framework for the integration of WSNs and MCC. There are three entities (i.e., mobile users, the cloud and the WSNs) and the working rules of this framework are described as follows. • First, the WSNs transmit sensory data to the cloud and the cloud performs data traffic monitoring of the sensory data transmitted to the cloud. • Second, when mobile users issue data requests to the cloud, the cloud returns the required data to the mobile user and performs data recommendation as well as data prediction. • Third, after the cloud obtains the data feature or data preference information which may indicate that the mobile users may have some features or preferences about the required data and WSNs may be in an undesirable status. The cloud informs the WSNs manager to utilize these information to optimize the deployment of WSNs and check the status of WSNs. In the following section, we introduce the detailed information with respect to data recommendation, data prediction, data traffic monitoring as well as WSN deployment and optimization.
II. R ELATED W ORK There are quite a few research work about integrating WSN and mobile cloud. However, these frameworks mainly focus on utilizing the sensory data after the sensory data is transmitted to the cloud. They do not consider the situation that the mobile users may have a preference about the sensory data. Or they ignore that the mobile users may have a trend of changing their required data. Also, the WSNs may tend to gather the dangerous data and miss some data or collect too much data. All these result from the context non-awareness problems between the mobile users and the WSNs as illustrated in the previous section. Particularly, an architecture based on cloud computing is proposed in [14] to improve the performance of WSN. Specifically, it assumes that the cloud acts as a virtual sink with many sink points that gather the sensing data from sensors. And each sink point is in charge of collecting data from the sensors within a zone. The sensory data of sensors is finally stored and processed in a distributed manner in cloud. [15] puts forward a dynamic proxy-based approach to connect sensors to the cloud. At the sensor tier, the aggregated sensor of a sink are then sent to the local proxy via a local communication mechanism. Then at the gateway tier, the local proxy parses the received messages and generates dynamic components over the event bus for each new sensor detected. The gathered sensor data is modelled and further relayed to the cloud. What’s more, another sensor-cloud integration framework is proposed in [16]. This framework includes the data processing unit (DPU), Pub/Sub Broker, request subscriber (RS), identity and access management unit (IAMU) and data Repository (DR). Sensory data collected from the WSNs is transmitted via a gateway to the DPU. Then DPU processes the data into a storage format and sends the data to the DR. In addition, [17] focuses on
A. Data recommendation Here, the Apriori algorithm [19] is utilized as it is a classic algorithm for frequent item set mining and association rule learning [20]. Specifically, Apriori is a level-wise complete search algorithm using the anti-monotonicity of itemsets. The Apriori Property is that any superset of an itemset will not be frequent if an itemset itself is not frequent. Let the set of frequent itemsets of size k be Fk and the candidates of the corresponding itemset be Ck . The main idea of Apriori is as follows. • First, the database is scanned and searched for frequent itemsets of size 1 by accumulating the count for each item. Those that satisfy the minimum support requirement are collected. • Second, iterating with the following three steps to find frequent itemsets with cardinality from 1 to k (k-itemset).
609
Mobile device 1
Mobile device 2
Mobile device 3
Data recommendation Data prediction Data traffic monitoring
Cloud Computing Platform
Wireless sensor networks (WSNs) Proposed framework to integrate wireless sensor networks (WSNs) and mobile cloud computing (MCC)
sm an ag er
WSNs
R ta
SN
da
In
ts es
fo rm
W
qu re
Cloud platform
ly ep
Send data requests
Mobile device
Transmit data
Fig. 2.
Cloud platform
Cloud platform
Data recommendation
Cloud platform
Data traffic monitoring
Data prediction Fig. 3.
Flowchart of proposed framework to integrate wireless sensor networks (WSNs) and mobile cloud computing (MCC)
610
•
support count(I) ≥ min conf support count(s)
1) From the frequent itemsets of size k, Ck+1 that are the candidates of frequent itemsets of size k + 1 is generated. 2) The database is scanned and the support of each candidate of frequent itemsets is calculated. 3) Those itemsets satisfying the minimum support requirement are added to Fk+1 . Third, the frequent itemsets are utilized to generate association rules if necessary with minimum support count & minimum confidence.
(1)
min conf is the minimum confidence threshold. B. Data prediction Here, since the sensory data collected by the WSN are time series data, we utilize the secondary exponential smoothing model (SESM) data prediction technique [21] [22] [23]. SESM is a widely used technique that can be applied to time series data, either to produce smoothed data for presentation or to make forecasts [24] [25] [26]. From [26], the prediction process of SESM is as follows. Suppose the time sequence consists of n times and the primary data Y = yl , y2 , . . . , yn are the data with original time (1) (2) sequence. Let St and St be the first and second smoothing (1) (2) value at the time t. St−1 and St−1 are the first and second smoothing value at time t − 1. With α which is the smoothing (1) (2) parameter, St and St are expressed as follows.
Pseudocode of Apriori algorithm 1. F1 ={Frequent itemsets of cardinality 1} 2. for (k = 1; Fk = ∅; k + +) do begin 3. Ck+1 =apriori-gen(Fk ); //New candidates 4. for all transactions t ∈ Database, do begin 5. Ct = subset(Ck+1 , t); //Candidates contained in t 6. for all candidates c ∈ Ct do 7. c.count++; 8. end 9. Fk+1 = {C ∈ Ck+1 |c.count ≥ minimum support} 10. end 11. end 12. return k Fk
⎧ (1) (1) St = αyt + (1 − α)St−1 ⎪ ⎪ ⎪ ⎨S (2) = αS (1) + (1 − α)S (2) t t t−1 (1) (2) ⎪ S = S = y ⎪ 1 1 1 ⎪ ⎩ t = 2, 3, · · · , n
The pseudocode of Apriori algorithm is shown as follows, with Ck denoting the candidate itemset of size k and Fk representing the frequent itemset of size k. Specifically, the following two steps (i.e., Join step and Prune step) are utilized to generate Ck+1 from Fk using function apriori-gen in line 3. • Join step: RK+1 which are the initial candidates of frequent itemsets of size k + 1 are generated, through taking the union of the two frequent itemsets of size k, Pk and Qk that own the first k − 1 elements in common. Specifically, RK+1 = Pk Qk = {item1 , . . . , itemk−1 , itemk , itemk }, in which Pk = {item1 , item2 , . . . , itemk−1 , itemk } and Qk = {item1 , item2 , . . . , itemk−1 , itemk }, where item1 < item2 < · · · < itemk < itemk . • Prune step: Check if all the itemsets of size k in Rk+1 are frequent and Ck+1 is generated by removing those candidates that do not pass this requirement from Rk+1 . This is based on the the Apriori Property that any subset of size k of Ck+1 which is not frequent cannot be a subset of a frequent itemset of size k + 1. Moreover, all the candidates of the frequent itemsets included in transaction t is found with function subset in line 5. Then only the frequency of those candidates generated in this way are calculated by scanning the database. Apriori scans the database at most kmax+1 times if the maximum size of frequent itemsets is set at kmax . After the set of frequent itemsets F is achieved, the association rules are generated from the frequent itemsets with the following procedure. • For each frequent itemset I in F , all nonempty subsets of I are obtained. • For every nonempty subset s of I, the rule s → (I − s) is selected if the following equation holds.
(2)
With (1), we can further deduce that Ft+m = at + bt m. Here, Ft+m is prediction value at the time t + m, in which m is the prediction steps and the maximum prediction step is M. ⎧ (1) (2) at = 2St − St ⎪ ⎪ ⎪ (1) (2) ⎨ α bt = 1−α (St − St ) ⎪ t = 1, 2, 3, · · · , n ⎪ ⎪ ⎩ m = 0, 1, 2, · · · , M
(3)
In addition, the prediction error P Eα is M
α |yn+m − yn+m |
(4)
m=1 α is yn+m is actual data value at the time n + m and yn+m the predicted data value at the time n + m.
C. Data traffic monitoring Based on the fact that the sensors are normally collecting the data with a set frequency (e.g., every 30 seconds), we can observe the data records to see whether there is too much or too less data for a specific time interval to perform data traffic monitoring here. If there is too much or too less data during a specific time, then this may indicate that some nodes in the WSNs are probably compromised and the WSNs manager should check whether the situation is true to avoid further harms.
611
D. WSN deployment and optimization
TABLE I DATA RECOMMENDATION PERFORMANCE BEFORE ATTRIBUTE COMBINATION
Generally, the WSN is deployed with a specific purpose (e.g., forest fire monitoring, etc.). Considering the specific WSN purpose, the sensor nodes are deployed considering coverage, energy consumption or delay with the node deployment methods in [27]. For the WSN, the base station records all data transmission route and data value. For instance, the humidity value is 50% at time 12:00 am and transmitted from node A to node B, etc. Thus, if we get the required data feature information, we can know where to deploy new sensor nodes in the WSNs. In addition, there is a WSN manager who could check the status of WSNs (e.g., in the forest). The WSN deployment is optimized as follows after the cloud manager informs the WSNs manager, with the data feature information as well as the potential WSNs status information. First, if the mobile users are particularly interested in the data from a specific area of WSN and want to achieve more data from that area, the WSN manager deploys more nodes in that specific area. Moreover, if the mobile users tend to achieve data with higher value (e.g., higher temperature or higher humidity), then more sensors are generally deployed for collecting more related data. Third, if the mobile users achieve some data that tend to be dangerous, then the WSN is checked to avoid the situation that it is in a warning state. Last, if sometimes the sensory data required by the mobile users may be missing or there are too much data transmitted from the WSNs to the cloud due to various factors, then the WSNs should also be checked to see whether it is compromised. If the WSNs are compromised, then the WSNs manager takes corresponding measures to solve the compromise issue.
Attribute Temperature Humidity Light Voltage
Confidence 0.48 0.52 0.52 0.51
TABLE II DATA RECOMMENDATION PERFORMANCE AFTER ATTRIBUTE COMBINATION
Attribute Combination Temperature& Humidity Temperature& Light Temperature& Voltage Humidity& Light Humidity& Voltage Light& Voltage
Confidence 0.51 0.43 0.56 0.45 0.52 0.46
B. Evaluation results 1) Data recommendation performance: Table I and Table II show the data recommendation performance with Apriori with respect to confidence before and after attribute combination. Here, we randomly select 1000 groups sensory data records and preset the attributes (i.e., temperature, humidity, light, voltage) that the mobile user will utilize. Then the support confidence for each attribute is determined and the corresponding confidence of two attribute combination is obtained. From this table, we can know that the proposed framework can accomplish the following goal. When the mobile user chooses the temperature data, by setting a confidence threshold, we can get some data combinations (e.g., temperature and humidity) whose confidence is larger than then threshold. Then these data is the recommended data that indicates stronger relationship than other data patterns. These recommended data also will help the WSNs to deploy sensors to the corresponding area that result in these strong data combinations. 2) Data prediction performance: Table III and Fig. 4(a) to Fig. 4(d) show the data prediction performance with the secondary exponential smoothing model (SESM) data prediction technique. Table III summarizes the prediction error with
IV. E VALUATION A. Evaluation setup To perform evaluation, we implement the following simplified prototype of the proposed framework. We assume that there is only one mobile user and the mobile user requests four types sensory data (i.e., temperature, humidity, light, voltage) for the Jupiter cloud platform of Dalian University of Technology in China. The Jupiter cloud platform utilizes 4 virtual CPU cores (2.6GHz) and 64GB memory. Then 8 real Mica2Dot sensor nodes built on the TinyOS platform are utilized to collect four types of sensory data (i.e., temperature, humidity, light, voltage) for 14 days in the Software School Campus of Dalian University of Technology in China. Specifically, the temperature is in degrees Celsius. And humidity is the temperature corrected relative humidity, which ranges from 0 to 100%. The Light is in Lux. Voltage is expressed in volts. The sensory data collected by sensor nodes are further processed with the Jupiter cloud for data analysis. The evaluated performance are data recommendation performance, data prediction performance as well as data traffic monitoring performance. Moreover, the mobile user quality of service is discussed.
TABLE III DATA PREDICTION PERFORMANCE Attribute Temperature Humidity Light Voltage
Prediction error 0.068328 0.08121 3.6759 0.005665
TABLE IV DATA TRAFFIC MONITORING PERFORMANCE Attribute Temperature Humidity Light Voltage
612
16 16 16 16
Data traffic records/minute records/minute records/minute records/minute
as well as WSNs potential status information. With these information about mobile user required data and WSNs data, we conduct the optimized deployment of WSNs and check the status of WSNs. The effectiveness of the proposed framework is proofed by extensive evaluations.
SESM for these 4 attributes utilizing 2000 groups of data. Fig. 4(a) to Fig. 4(d) show the data prediction sample with respect to temperature, humidity, light and voltage, respectively. From Table III and Fig. 4(a) to Fig. 4(d), we can observe that the our predicted data value is almost the same as the actual data value. 3) Data traffic monitoring performance: Table IV shows the data traffic monitoring performance. We monitor the data traffic for a specific time (i.e., every one minute) regarding the four types of sensory data (i.e., temperature, humidity, light, voltage). From this table, we can obtain that the data traffic is normal. 4) Quality of service of mobile user: We analyze whether the mobile user’s data requests can be continually satisfied with and without our proposed framework. If the mobile users are specially interested in one specific area data of WSN and want to get more data from that area, with data recommendation in the cloud, our framework already deploys more nodes in that specific area thus the desirable data could be transmitted to the mobile user. In contrary, although the data requests of mobile users could also be answered without our framework, the data is unlikely to be what mobile user desires. Similarly, if the mobile users tend to achieve higher data value (e.g., higher temperature or higher humidity), our proposed framework could transmit these desirable data to the mobile user directly with data prediction by the cloud. While it is unlikely to happen without our framework. In addition, if the mobile users achieve some data that tend to be dangerous, our framework already checks the WSN state utilizing cloud side data prediction. Also, the WSN state is checked with our framework if the sensory data required by the mobile user are missing or abnormal resulting from various factors by data traffic monitoring on the cloud side. In these two cases, without our framework, the WSN may already be compromised or the cloud service has to be stopped (e.g., WSN is compromised to not send sensory data any more) or natural disaster (e.g., forest fire happens) because that WSN detects the dangerous data too slow. In summary, from the view of offering required data to the mobile users, we can observe that our proposed framework is effective in promptly offering the desirable data service for the requests of mobile user, while the mobile user’s requests cannot be always timely responded without our framework. Our proposed framework also can enhance the performance of WSNs.
ACKNOWLEDGEMENT This work is supported by a Four Year Doctoral Fellowship from The University of British Columbia and by funding from the Natural Sciences and Engineering Research Council of Canada, TELUS and other industry partners. This work was supported by the National Natural Science Foundation of China (Grant No. 51104157), the Ph.D. Programs Foundation of Ministry of Education of China (Grant No. 20110095120008), the China Postdoctoral Science Foundation (Grant No.20100481181), the Fundamental Research Funds for the Central Universities (Grant No. 2011QNA30), and Jiangsu Overseas Research & Training Program for University Prominent Young & Middle-aged Teachers and Presidents. R EFERENCES [1] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, “Wireless sensor networks: a survey,” Computer Networks, vol. 38, pp. 393–422, 2002. [2] C. Zhu, L. Shu, T. Hara, L. Wang, S. Nishio, and L. T. Yang, “A survey on communication and data management issues in mobile sensor networks,” Wireless Communications and Mobile Computing, 2011. [3] M. Li and Y. Liu, “Underground coal mine monitoring with wireless sensor networks,” ACM Transactions on Sensor Networks, vol. 5, 2009. [4] Q. Zhang, L. Cheng, and R. Boutaba, “Cloud computing: state-of-the-art and research challenges,” Journal of Ineternet Services and Applications, vol. 1, pp. 7–18, 2010. [5] C. Zhu, V. C. M. Leung, X. Hu, L. Shu, and L. T. Yang, “A review of key issues that concern the feasibility of mobile cloud computing,” in Proc. IEEE International Conference on Cyber, Physical and Social Computing (CPSCom), 2013, pp. 769–776. [6] H. T. Dinh, C. Lee, D. Niyato, and P. Wang, “A survey of mobile cloud computing: Architecture, applications, and approaches,” Wireless Communications and Mobile Computing, 2011. [7] S. Wang and S. Dey, “Adaptive mobile cloud computing to enable rich mobile multimedia applications,” IEEE Transactions on Multimedia, vol. 15, pp. 870–883, 2013. [8] W. Lumpkins, “The internet of things meets cloud computing,” IEEE Consumer Electronics Magazine, vol. 2, pp. 47–51, 2013. [9] W. Kurschl and W. Beer, “Combining cloud computing and wireless sensor networks,” in Proc. 11th International Conference on Information Integration and Web-based Applications & Services (iiWAS), 2009, pp. 512–518. [10] J. Jayashree and J. Vijayashree, “Ubiquitous life care integrates wireless sensor network and cloud computing with security,” Global Journal of Computer Science and Technology, vol. 11, 2011. [11] A. Kapadia, S. Myers, X. Wang, and G. Fox, “Secure cloud computing with brokered trusted sensor networks,” in Proc. International Symposium on Collaborative Technologies and Systems (CTS), 2010, pp. 581– 592. [12] S. K. Dash, J. P. Sahoo, S. Mohapatra, and S. P. Pati, “Sensorcloud: Assimilation of wireless sensor network and the cloud,” in Proc. Second International Conference on Computer Science and Information Technology (CCSIT), 2012, pp. 455–464. [13] X. H. Le, S. Lee, P. T. H. Truc, L. T. Vinh, A. M. Khattak, M. Han, D. V. Hung, M. M. Hassan, M. Kim, K.-H. Koo, Y.-K. Lee, and E.-N. Huh, “Secured wsn-integrated cloud computing for u-life care,” in Proc. IEEE Consumer Communications and Networking Conference (CCNC), 2010. [14] P. Zhang, Z. Yan, and H. Sun, “A novel architecture based on cloud computing for wireless sensor network,” in Proc. International Conference on Civil, Structural and Earthquake Engineering (ICCSEE), 2013, pp. 472–475.
V. C ONCLUSION The integration of wireless sensor networks (WSNs) and mobile cloud computing (MCC) is widely focused. In this paper, focusing on solving the context non-awareness issue that we observe between the mobile user and the WSNs, we provide a novel framework to provide desirable data to mobile users when integrating WSNs and MCC. The proposed framework takes account of the data recommendation, data prediction as well data traffic monitoring so as to achieve the mobile user data preferences and data feature information
613
(a)
(b)
Fig. 4.
(c)
(d)
Data prediction performance sample regarding temperature (a), humidity (b), light (c) and voltage (d)
bridge, Massachusetts, Arthur D. Little Inc., 1956. [22] C. C. Holt, “Forecasting trends and seasonal by exponentially weighted averages,” International Journal of Forecasting, vol. 20, no. 1, pp. 5–10, 1957. [23] S. K. Prajakta, “Time series forecasting using holt-winters exponential smoothing,” Technical report of Kanwal Rekhi School of Information Technology, pp. 1–13, 2004. [24] R. J. Hyndman, A. B. Koehler, R. D. Snyder, and S. Grose, “A state space framework for automatic forecasting using exponential smoothing methods,” International Journal of Forecasting, vol. 18, no. 3, pp. 439– 454, 2002. [25] J. W. Taylor, “Exponential smoothing with a damped multiplicative trend,” International Journal of Forecasting, vol. 19, no. 4, pp. 715– 725, 2003. [26] F. Hao and Z. Pei, “A method of determining the secondary exponential smoothing parameter based on owa,” in Proc. International Symposium on Communications and Information Technologies (ISCIT), 2006, pp. 459–462. [27] W. Y. Poe and J. B. Schmitt, “Node deployment in large wireless sensor networks: coverage, energy consumption, and worst-case delay,” in Proc. Asian Internet Engineering Conference (AINTEC), 2009, pp. 77–84.
[15] W. Wang, K. Lee, and D. Murray, “Integrating sensors with the cloud using dynamic proxies,” in Proc. IEEE 23rd International Symposium on Personal Indoor and Mobile Radio Communications (PIMRC), 2012, pp. 1466–1471. [16] K. Ahmed and M. Gregory, “Integrating wireless sensor networks with cloud computing,” in Proc. Seventh International Conference on Mobile Ad-hoc and Sensor Networks (MSN), 2011, pp. 364–366. [17] B. Tang and Y. Wang, “Design of large-scale sensory data processing system based on cloud computing,” Research Journal of Applied Sciences, Engineering and Technology, vol. 4, no. 8, pp. 1004–1009, 2012. [18] K. Lee, D. Murray, D. Hughes, and W. Joosen, “Extending sensor networks into the cloud using amazon web services,” in Proc. IEEE International Conference on Networked Embedded Systems for Enterprise Applications (NESEA), 2010, pp. 1–7. [19] R. Agrawal and R. Srikant, “Fast algorithms for mining association rules,” in Proc. 20th International Conference on Very Large Data Bases (VLDB), 1994, pp. 487–499. [20] X. Wu, V. Kumar, J. R. Quinlan, J. Ghosh, Q. Yang, H. Motoda, G. J. McLachlan, A. Ng, B. Liu, P. S. Yu, Z.-H. Zhou, M. Steinbach, D. J. Hand, and D. Steinberg, “Top 10 algorithms in data mining,” Knowledge and Information Systems, vol. 14, pp. 1–37, 2008. [21] R. G. Brown, “Exponential smoothing for predicting demand,” Cam-
614