1 Introduction

A wireless sensor network (WSN) is a collection of tiny sensor devices not connected by wires, that can sense, process, and propagate environmental factors like temperature, humidity, pressure, and movements. Every sensor device, referred to as a node, has four basic units: a processor, communication unit, memory and power supply. There are multiple vendors manufacturing different types of sensors and sensor subunits for assembly as sensors. Consequently, there are multiple implementation choices for the hardware and drivers used for communication and to manage the sensors. Some vendors provide an end-to-end solution including sensors, testbeds, drivers and application interface modules; these solutions often come at huge cost, are limited by minimal customization options and have only fixed protocol support. The critical factor in setting up an application of a WSN is to limit node energy as the nodes must be able to execute with limited battery power [1].

We need an extensible design of a wireless sensor node and a service panel to consolidate all the hardware/software options so that any desired application with the required WSN profile can easily be set up. The service panel should assist not only in setting up an implementation testbed, but also in managing the WSN activities, especially energy engineering, thereby ensuring extended lifetime of the nodes. Since the demise of a node causes a break in the coverage area, this should be monitored, predicted in advance and avoided. Thus, we need preventative action to identify energy depletion in nodes and predict the loss of a node in the coverage enabling this to be addressed ahead of time as a self healing activity [2]. In this paper, we discuss the energy management framework Aatral, which acts as an application neutral service panel to set up, functionally manage activities, and support the self-healing energy engineering processes in a WSN.

2 Challenges in setting up a WSN

2.1 Node design

Sensor implementations are created for various purposes such as wild fire detection, intruder detection, industry automation, intelligent traffic systems and precision agriculture, using temperature, motion, and humidity sensors. When considering setting up these implementations, the initial considerations revolve around the traditional quality of service (QOS) parameters, namely, baud rate, latency delays, channel utilization, collision avoidance schemes and shortest path routing protocols. However, the key fact to consider is that a sensor network is not just an extension of a traditional network; it exists with limited battery power supply in the nodes, which determines the lifetime of the network [2, 3]. It is difficult to charge remotely distributed nodes and enormous extra effort and cost is spent on energy engineering to reduce the power cost in the name of power budgeting.

2.2 Power aware decisions

Based on the discussion above, it is necessary to consider power-aware hardware, drivers, applications, energy optimized protocols, aggregation schemes, scheduling mechanisms and time synchronization methods when designing an implementation. If multimedia data transmission is needed, the graphical ports should be enhanced. Thereafter, there may be a need for multiple integrated sensors rather than a single sensor. In the case of a traffic sensor, temperature sensors for fog detection and movement sensors for understanding the traffic are incorporated. If there is a need for location awareness, geosensors must also be taken into account [3]. Multiple protocol support and multiple communication unit support with configurable bandwidths constitute the next decisions. A single testbed supporting multiple purposes has many features that can lead to complexity in the future. The critical issues are how the hardware and drivers can be configured and integrated end-to-end with the application and how the data are processed and reported [4] (Table 1).

Table 1 Different considerations in setting up WSN implementations
Full size table

2.3 Custom hardware design with custom software

In the current market, there are no ready-made hardware designs that support all these considerations. In most cases, either custom hardware is designed or standard devices are customized. Adopting the former option involves drivers and application design. In the latter case, it is almost impossible to customize the system, since in doing so, we invariably loose the technical support and guarantees of the vendor [4]. Hardware and drivers are almost integrated with fewer choices.

Software applications are associated with the hardware and consolidate the programming abstraction of the selected application logic as the needs of the next level [4, 5]. Nevertheless, developing a framework for a configurable hardware neutral application is near impossible.

3 Related work

Most simulator and implementation hardware vendors consider a WSN as an extension of a traditional network, with the exception that data and control communication is wireless. It is a fact that the limited power supply in a wireless network is hardly considered. The difficulty in remote charging and the risk of dead nodes and broken links for communication due to a lack of energy in the critical implementation are seldom considered during the design and deployment, nor in the selection of an application, that is, the algorithmic components of the WSN [6]. Recently, researchers have begun incorporating and spending more effort on energy engineering in WSNs.

3.1 Energy efficient protocol design

Low energy adaptive clustering hierarchy (LEACH) protocol design creates distributed cluster heads and uses energy efficient techniques like rotating cluster heads and local processing to reduce global communication and compress data for minimal transmission frames. LCA2 and HEED are two variations in the cluster based communication protocol.

Data centric protocols (SPIN) are designed in such a way that the metadata are published as advertisement metadata (ADV) packets rather than publishing the actual data. Nodes that are interested in the advertised message send a request for the data (REQ) to the publisher node, which, in turn, sends the data only to those nodes having expressed interest, rather than to all the nodes. Bluetooth smart protocols are available as single and dual mode low energy versions, in which all communication is low energy and builds on the service-based architecture.

3.2 Energy efficient software component design

In existing energy-balanced data aggregation algorithms, there is a trade-off between delay and energy consumption. Thus, data can be prioritized and lower-priority data can be delayed to save energy. PEGASIS is an energy greedy chain-based aggregation algorithm where the energy consumption during the aggregation process is also considered in decision-making. In the case of PREMON data aggregation algorithms, the network of sensors is considered as images and data packets are transmitted based on behavioral analysis, receiver tolerance limits and predicted data rates, rather than using immediate aggregation or transmission.

Location-aware approaches and suppression-based selection, aggregation and multipath aggregation are some of the variations of the generic aggregation methods in the more recent protocols for energy efficiency improvement.

Tiny sync, mini sync, reference broadcast synchronization, flooding time synchronization and lightweight time synchronization for sensor networks are different time synchronization methods that consider energy consumption [7].

The time division multiple access (TDMA) based MAC protocol employs a periodic duty cycle to change the state from sleeping to the awake. Many variations such as distributed adaptive scheduling like TRAMA, optimized scheduling, location-aware priority-based scheduling and contention-based protocols like PAMAS have been introduced in the quest for energy efficiency [8].

3.3 Energy engineering in WSNs

Energy engineering initiatives in WSNs can be broadly classified as energy auditing, energy optimization, energy harvesting and energy scavenging as in the Fig. 1.

Fig. 1
figure 1

Energy engineering in a WSN

Full size image

Lord Kelvin once stated: “If you cannot measure it, you cannot improve it”. Thus, the first step in energy engineering is to measure and monitor energy usage. An energy audit is an inspection, survey and analysis of energy flows for energy conservation in a process or system to reduce the amount of energy input into the system without negatively affecting the output(s). This has the effect of making the design and operation energy-aware, thereby reducing energy consumption. This is referred to as energy optimization. The energy reclaimed from the unused portion of energy that is dumped is referred to as energy scavenging. Energy harvesting is the systematic process by which energy is derived from the external sources captured and stored [9]. As the energy reclaimed from scavenging is minimal, here we consider only the first three engineering aspects (Table 2).

Table 2 Consolidated survey of energy engineering in a WSN
Full size table

3.4 Existing gaps and actual needs

  1. (i)

    Each researcher typically focuses on only one aspect of energy improvement and very often such endeavors are not even validated to show whether they reduce overall network energy consumption, thus ignoring the holistic approach for energy improvement initiatives.

  2. (ii)

    No quantification of the costs involved in implementing a proposed energy improvement in a WSN compared with the savings derived from greater energy efficiency is given.

  3. (iii)

    No ready-made hardware or combination of hardware components supporting energy studies with a simple configuration and setup is available from commercial vendors.

  4. (iv)

    Few proposals in this field have been offered as a service that can be fine-tuned or configured to any domain-specific real-time need of a WSN. Thus, redundant effort and costs have been invested in energy engineering in the different WSN applications and WSN platforms.

  5. (v)

    No clear holistic standard policies and procedures on energy management, nor accurate energy models with a mechanism for finding the energy dissipation rate are readily available.

4 Our proposed solution

4.1 Aatral-an integrated energy management framework for WSNs

Aatral is an integrated energy management framework for WSNs developed and tested with specific WSN profiles at the Madurai Kamaraj University. The framework has been designed as a device and platform neutral framework with guided setup to port it to any WSN implementation. This portability makes it an inevitable energy management backbone structure and helps to manage the energy consumption of any WSN network.

4.2 Motivation for and objectives of the Aatral framework

  1. (i)

    A software service that can assist in monitoring and auditing a specific WSN profile or setup.

  2. (ii)

    A consolidated service panel that identifies energy leakages in the WSN communication at various levels, suggests fixes and provides energy efficient recommendations for these.

  3. (iii)

    An online service that suggests the best way to heal an energy issue and attempts to perform this healing on its own when the user consciously selects the Fix option provided in the suggestions.

  4. (iv)

    Alternate energy resources can be availed as a service based on the WSN administrator option.

  5. (v)

    A master database of historic data of the WSN is provided for comparison. An audit checklist of measurable factors and an energy baseline data matrix of recorded values of different parameters is available for the selected energy portfolio. Based on historical data, we can derive the energy benchmark for different WSN profiles. We have developed a mechanism to find the energy peak index (EPI), energy use index (EUI) and energy dissipation rate (EDR) together with trends and prediction of the lifetime of an individual sensor network node. The EPI defines the maximum energy the unit has consumed over a given period of time, while the EUI measures overall energy consumption. EUI can be calculated using the sum of energy at different activity levels and energy consumption at different unit levels. This index is calculated to validate whether the energy reduction techniques have altered the final result of the index and remain constant after different energy consumption scheme changes.

  6. (vi)

    A WSN specific energy efficiency plan is prepared from the energy potential scan and using clear graphs, this is compared with the benchmark data for the WSN profile. Outputs are the energy graph, energy auditing checklist, methodology, principles and policies.

  7. (vii)

    We have designed an energy economics calculator (EEC) that calculates the operational cost of energy consumption for a WSN profile over a period of time together with a simple comparison of the energy cost saved by optimization. This is validated against the cost of implementing the optimization.

4.3 Aatral: an accurate energy model

An energy model is the collection of factors considered in a WSN system to measure energy. In traditional energy measurement methods, only the residual energy is considered; activity wise, unit wise and application component wise energy is not recorded [10]. Therefore, we need a more accurate energy model for micro energy management. The energy model must be able to capture both activity level energy and operational level energy. It is depicted in Fig. 2.

Fig. 2
figure 2

Energy model for more accurate energy reading in a WSN with Aatral Special registers are provided to store the details of energy spent on programming aspects like data aggregation, time synchronization and scheduling, and client applications

Full size image

4.3.1 Node wise energy consumption (component wise energy)

$$\begin{aligned} & {\text{EUIcom}} = {\text{energy}}\,{\text{consumed}}\,{\text{by}}\,{\text{micro}}\,{\text{controller}} \\ & \quad + {\text{energy}}\,{\text{consumed}}\,{\text{by}}\,{\text{internal}}\,{\text{memory}} \\ & \quad + {\text{energy}}\,{\text{consumed}}\,{\text{by}}\,{\text{sensing}}\,{\text{unit}} \\ & \quad + {\text{energy}}\,{\text{consumed}}\,{\text{by}}\,{\text{sensing}}\,{\text{unit}} \\ & \quad - {\text{energy}}\,{\text{depletion}} \\ \end{aligned}$$
(1)

4.3.2 Activity wise energy consumption

$$\begin{aligned} & {\text{EUIAct}} = {\text{energy}}\,{\text{for}}\,{\text{sensing}} + {\text{energy}}\,{\text{for}}\,{\text{localization}} \\ & \quad + {\text{energy}}\,{\text{for}}\,{\text{encryption}} + {\text{energy}}\,{\text{for}}\,{\text{data}}\,{\text{processing}} \\ & \quad + {\text{energy}}\,{\text{for}}\,{\text{communication}}\;\left( {{\text{transmission}},\,{\text{receiving}}} \right) \\ & \quad - {\text{energy}}\,{\text{depletion}} \\ \end{aligned}$$
(2)

4.4 Node design in Aatral

In setting up our application testbed, we tried to address the above mentioned challenges with clear decisions on node design as in the Fig. 3. Aatral project nodes include an enhanced graphical port, support for multiple sensors, power measuring circuits, attached solar panel, an LED display, the latest ARM processor with multiple protocols, communication units in a master/slave mode and a power supply supporting the test mode.

Fig. 3
figure 3

Node design for Aatral framework

Full size image

The application design abstracts the hardware and software components in easily configurable options for choosing the profile setup, auditing and monitoring the system, acquiring sensor readings and performing comparisons [11–13].

4.5 Tested WSN profile for Aatral

See Table 3.

Table 3 One of the WSN testbed profiles considered for Aatral
Full size table

4.6 Instrumentation

The complete Aatral implementation testbed consists of three sensor nodes, one coordinator node, an energy harvesting unit, energy optimization switches, an energy measurement unit and Zena data packet analyzer circuits. This will be scaled up in the OMNETPP simulator to include 50–100 nodes together with the incorporation of a realistic energy model. Realistic energy measurements have been considered in the OMNETPP simulator for scaling up the 3-node actual testbed to a 100-node simulation that measures the energy variances. The instrumentation and OMNETPP scaled up simulations are explained in the Figs. 4, 5.

Fig. 4
figure 4

Energy engineering framework for Aatral testbed

Full size image
Fig. 5
figure 5

Scaled-up OMNETPP simulator with 100 nodes

Full size image

The challenge is that any improvement in energy consumption should not affect the other QOS parameters like baud rate, latency, delay, dropped packet rate and number of re-requests, because communication activities require more energy than the other activities in the network [15]. Thus, we have verified in both real and scaled-up simulated environments that the QOS parameters are not negatively affected.

5 Architecture and design of Aatral

Aatral has been designed as a layered service-oriented architecture as in the Fig. 6, with portability, extensibility and easy maintenance as the primary goals. MySQL is used as the backend database, while the operating system is Windows. The application server was designed with Matlab Server Pages in sync with a Java virtual machine, which in turn uses a Java Native Interface through Matlab services. The JSP(Java Server Pages) web clients were designed and tested with the Apache Tomcat web server to consume the application server services, while a JDBC(Java Database Connectivity) driver is used for database connectivity. Silicon Laboratory USB-based device drivers are used to connect the base station with the system for ARM Cortex, Arduino, and Raspberry Pi base stations. This paper discusses only the ARM Cortex based WSN profile.

Fig. 6
figure 6

Architecture of the Aatral framework

Full size image

A unified data model of the data and metadata is provided by the JSON document format (key-value pair) as different hardware implementations may support different types of data; in this way, these are interpreted in the same way. Pre-built Word and Excel templates are used by the report engine which relies on the graph and chart utility of Matlab. LPCExpresso is used to write, edit and test the embedded code to implement connections between the sink and base nodes, and for node level operations of scheduling, synchronization, sensing, cleaning the sensing data, EEPROM writes/reads, potential energy scans of channels to choose a higher energy channel, setting baud rates, state changes in nodes, aggregation of data, publishing interest cache, propagation using the transreceiver and processing the received data.

5.1 Pseudo-code for Aatral

The high level pseudo-code for Aatral is given below in Fig. 7. Management framework for WSNs.

Fig. 7
figure 7figure 7

Pseudo-code for Aatral energy management framework for WSNs

Full size image

5.2 Energy auditing system

An energy measuring circuit is incorporated, which helps monitor the energy of the node to which it is attached [14, 15]. The energy auditing system reports the current processor mode (active or sleeping) and the voltage, amp and watt values to the PC. From these we are able to ascertain the current battery voltage level and the current energy consumption at any point in time. The desired number of readings are grouped and mapped with a timestamp for later reference. The unit is designed in this way, so that at a later stage we can compare a reading set with timestamp 1 with other readings sets with timestamp 2.

5.3 Energy optimization system

Two switches are provided to transition the microprocessor from normal mode to optimized mode. Once optimized mode has been selected, the scheduler takes care of the state change mechanism of the microprocessor and moves it from a sleeping to an active mode in the desired time interval. The sample energy correction codes are given in the Fig. 8.

Fig. 8
figure 8

Code snippet for energy optimization

Full size image

In data aggregation and routing schemes, decision making also considers the energy factor. Thus, the overall network energy is improved and is accurately measured by this unit [16].

In addition, a module interface service has been designed in such a way as to identify the protocol stack, currently (802.15.4, Zigbee, MIWI) in the WSN profile that is ported to Aatral, list the existing options or algorithmic approaches for aggregation, routing, cluster head formation, propagation and the data format, so that the auditing system can identify possible optimization fixes and make recommendations [17, 18]. Once the user has selected the fix option, the module replacement mechanism is activated. Embedded code corrections are suggested, together with an analysis of a sample correction; however, these suggestions are not replaced in the software components. If there is a conflict between changes in the programming components and the embedded coding, the code is not replaced, but retained with a possible hint at a later stage for correction.

5.4 Energy harvesting system

The energy harvesting unit finds the external energy source as in Fig. 9. The rechargeable battery, which obtains electrical power from a solar unit, supplies power to the node instead of from the battery [19]. You can cross check the data packets in the communication using the ZENA tool as in the Fig. 10.

Fig. 9
figure 9

Energy harvesting unit that supplies power to the sensor node

Full size image
Fig. 10
figure 10

Snapshot of the data packets observed in WSN communication

Full size image

5.5 Energy economics calculator

Business intelligence reports, which are currently very popular, are nothing more than an extension of decision support systems, where historical data over a period of time are used as benchmarks and thresholds and to provide hints for possible improvement areas [20]. One such initiative among the energy engineering initiatives of WSNs is the energy economics calculator. Typically, vast costs are incurred for energy improvements. Here, we check whether the invested amount is recovered in terms of energy saved over a certain period of time. If we were to invest 100 million dollars in energy saving initiatives, but only saved 10,000 dollars, the return on investment (ROI) and internal rate of return (IRR) would be much less and the payback period would be longer [21]. Interest on the investment is also taken into account when deciding on what economies should be addressed.

Thus, all opportunities for energy improvement are first cross checked with this business intelligence factor of energy economics before decisions are made. The development framework of Aatral is called a self-deterministic energy engineering framework, because the cross validation automatically implements the energy corrections for saving energy [20, 21]. Hence, the system addresses the coverage issue by extending the lifetime of a node. This feature justifies the use of the adjective “self-healing” in connection with the Aatral framework.

6 Results and discussion

6.1 Consideration of energy economics

After every 30 readings in a second, there is an option in the Aatral framework to find the average power. We recorded and compared the power after 10,000 readings (5.5 min, i.e., approximately 6 min) and 100,000 readings (55.5 min, i.e., approximately an hour) in the normal and energy optimized modes (involving microprocessor state change scheduler optimization, aggregation scheme optimization, energy aware routing, and data delay energy correction of lower baud rate) (Table 4).

Table 4 Energy optimization techniques applied in Aatral
Full size table

The cost of electricity is calculated using the following formula:

$${\text{Wattage}} \times {\text{hours}}\,{\text{used}} \div 1000 \times {\text{price}}\,{\text{per}}\,{\text{kWh}} = {\text{cost}}\,{\text{of}}\,{\text{electricity}}$$
(3)

If the usage duration is less than an hour, the calculated cost is negligible, and therefore, it is reported as being not applicable. The price per kWh basically depends on the associated state and the source of the energy. An average of $2 per kWh is considered here as the basic cost of electricity. A capital investment of $20 is used to set up the WSN system to consider energy optimization (Tables 5, 6, 7, 8).

Table 5 Average energy variance of 30 readings in a second
Full size table
Table 6 Average energy variance for 10,000 readings in 6 min
Full size table
Table 7 Average energy variance for 100,000 readings in an hour
Full size table
Table 8 Summary of cost savings
Full size table

For four nodes, the energy optimization capital investment is $20, which results in an average saving of $43.20 per month. When the amount saved through energy optimization per month is twice that of the capital investment of the energy optimization initiative, a threshold validation of the values of the other QOS parameters is considered, and only then is the “good to go” decision made. The payback period is ultimately less than a month.

6.2 Quality of service parameter verification

Figure 11 shows the auto-generated output of OMNETPP for a given WSN profile with the real-time energy model scaled to 100 nodes. Here, the average baud rate of 256 KBS is verified before and after energy optimization for 10,000,000 nodes. In the same way, the number of dropped control packets and queue length are cross validated after the energy optimization to ensure they remain the same.

Fig. 11
figure 11

Before and after energy optimization, values of QOS parameters are verified

Full size image

6.3 Sensor and energy readings

Temperature sensor measurements obtained by using a solder iron for heating and an ice bag for cooling, were observed in the Aatral application testbed. These were plotted as graphs depicting energy readings in the normal and optimized modes, and compared to ascertain an inferred energy variance.

The MCP9800 temperature sensor is used in the testbed. Output generated by the report generator, depicted in Fig. 12, shows how the measurements of room temperature from three sensors are tracked. When one of the sensors is heated with a solder iron, the temperature reading varies only for that sensor, and similarly, when one of the sensors is cooled with the ice bag, the change is also reflected and tracked by the report generator. The aggregated sink node value is also depicted in Fig. 12.

Fig. 12
figure 12

Temperature measurements obtained from temperature sensors in Aatral

Full size image

Energy can be measured in terms of current, voltage, or watts. Figure 13 shows the current energy consumption of a single node in the default mode, after selecting the auditing option. Figure 14 shows how after optimization, the energy level is reduced, depicting the actual energy difference. Here, the system transitions from active to sleep state, and energy-aware aggregation, energy-aware routing and delayed data communication occur. Energy corrections are introduced and the appropriate energy readings are recorded. Notable energy saving is demonstrated, although this varies for different WSN profiles: 20 % saving in the worst case and around 42 % in the best case.

Fig. 13
figure 13

Energy measurements in normal mode in Aatral

Full size image
Fig. 14
figure 14

Energy measurement for the optimized mode of Aatral

Full size image

6.4 Performance monitoring

The main performance factors, namely, delay, throughput, and baud rate are verified in conjunction with the relevant energy measures (Table 9).

Table 9 Performance monitoring in Aatral
Full size table

We observe both the frequency of communication for a fixed baud rate and the QOS parameters for the configured WSN profile. Figure 15 indicates the percentage of dropped packets per second, that is, the chance of retransmission, which is directly proportional to energy consumption.

Fig. 15
figure 15

Packet loss % per second—chance of retransmission

Full size image

Figure 16 shows the maximum and overall energy consumption indices calculated for the observed second. These values give an indication of energy efficiency and the energy load of a particular node.

Fig. 16
figure 16

Energy use index and energy peak index of Aatral nodes

Full size image

The current energy level of a node, shown in Fig. 17, provides information of the battery state and the energy and health status of the node as well as channel utilization as a percentage. The values of these are verified for consistency before and after energy optimization.

Fig. 17
figure 17

Channel utilization and current energy level of Aatral

Full size image

Figure 18 shows the latency delay of the end-to-end propagation of each node. Delayed communication and avoiding duplicate data packet communication improve energy consumption and are incorporated in Aatral. However, these should not affect the average latency delay in communication, which is one of the critical QOS factors that cannot be compromised in WSN communication.

Fig. 18
figure 18

Average latency of the end-to-end delay of the propagation of nodes

Full size image

7 Summary of findings and benefits

When energy optimization is considered proactively in the design of the application testbed and in the WSN itself, power cost can be reduced substantially with a minimum capital investment in optimization. However, validation of the investment in energy saving, the cost saved from the energy optimization, and the consistency of the QOS parameters before and after the energy saving must be performed. The benefits of Aatral are given below.

  1. (i)

    It is possible to configure popularly available WSN profiles using the guided setup.

  2. (ii)

    At any point in time it is possible to trace the activity wise and component wise energy of a node or the entire WSN to ascertain the energy and health status of the WSN and store it for later reference in the database.

  3. (iii)

    It is promising to investigate energy consumption and find possible energy optimizations at the programming component level and also to be able to fix some of the main optimization options.

  4. (iv)

    It is possible to find the dissipation rate of energy for a node based on the average number of events, average distance from the sink and predicting the lifetime of the node and proactively instituting replacement orders. It is also possible to raise a service request to alert and switch to the harvesting unit energy source, if available.

  5. (v)

    Minimal values of energy measures can be set as benchmarks for a profile and these can be compared with actual values to obtain the energy variance.

  6. (vi)

    The energy economics calculator helps track the energy operating cost. With this basic information we can derive a power budget and for energy relevant investments find the ROI, IRR, and payback period.

The Aatral framework helps manage the energy of a WSN network with full coverage and zero downtime. This is crucial for health care, intruder detection or industry automation sensor systems. All the operations are energy aware to reduce energy consumption. The best feature is the ability to track the cost involved in providing energy to help derive the power budget and returns on energy relevant investments.

8 Conclusion and future directions

At present, since node hardware is connected to a MATLAB application, the processing of huge volumes of data using traditional methods is tedious. Incorporating the embedded code, drivers, application programming and end-to-end connectivity in one platform to provide configurable hardware features is desirable in big data platforms [22]. Big data platforms for the Internet of Things (IOT) are available for establishing better end-to-end integration and voluminous big data processing. These inclusive platforms simplify the connectivity of devices in the IOT, protocol stacks, data processing and data reporting. ARM’s Mbed and the ThingSpeak platforms serve this purpose, while Sitewhere, Glass beam, and Zetta are some of the other IOT big data platforms. Efficient data handling of data in a WSN with a big data platform invariably improves energy consumption and gives us an insight into the business intelligence of power budgeting. In the foreseeable future, a big data platform for IOT will be considered for a smarter WSN setup.