How the synergy between Cloud & Edge computing is helping IoT
Cloud computing along with Edge computing is shaping the future of Internet of Things (IoT). This combination brings stability to the devices connected in the IoT network along with addressing the latency issues by processing the data closer to its source.
Cloud computing has clearly changed the shape of data processing, particularly for Big Data. Today, we capture, store and process data without worrying about provisioning the compute resources and managing them. Leveraging the compute power of cloud, IoT has grown leaps and bounds. IoT has been installing billions of smart devices per year and some estimates suggest that more than 20 billion smart devices will be installed by the year 2020. With numerous devices installed and connected to IoT, the amount of data to be processed has been increasing astronomically. Data analysts and data scientists are facing challenges to process and analyze this data, especially in situations where these need to be processed in near real time. Cloud computing alone will not be able to help in processing such massive data sets and provide resolutions in real time. So, this leads us to the solution to this problem, which is Edge computing.
Decoding Edge Computing
According to International Data Corporation (IDC), Edge computing is a mesh network of micro data centers that process or store critical data locally and push all received data to a central data center or cloud storage repository, in a footprint of less than 100 square feet.
In simpler words, Edge computing enables processing and analyzing data closer to the source where the data is generated.
A smart device installed and connected in an Edge computing environment has the capability to process the mission critical data and respond back in real time, instead of sending all of its data over the internet to the cloud and wait for the cloud to respond back. The device itself acts like a mini data center and since the basic analysis is happening on the device, there is near zero latency. With this new added capability, the processing of data becomes decentralized and the traffic over the network is vastly reduced. Cloud can later collect this data for second round of evaluation, processing and deeper analysis.
Benefits of Edge Computing for IoT
There are a number of benefits to leverage Edge computing for IoT devices, such as near zero latency, lesser network load, increased resiliency, reduced data exposure and lesser costs for data management. Let’s take a look at these one by one:
Near Zero Latency:
Near zero latency is the biggest advantage of Edge computing. The time lapse between data collection, processing and taking an action is near real time. This is a significant requirement for IoT devices in mission critical situations. A very good example for this would be the driverless autonomous cars.
Google estimates that their self-driving cars produce about 1GB of data every second! A lot of this data needs to be processed quickly so that the car can keep itself on the correct course and avoid collisions. Imagine if this data is collected, transmitted to cloud, cloud processes it and then sends it back to the car. Though the entire process is done in a few seconds, it would prove to be too late and the car might have already met with a collision. The best solution in this scenario is to analyze the data at the sensor itself using Edge computing and then send it to cloud for subsequent analysis. Latency sensitivity is even more critical in health care industry where devices are connected to heart rate monitors or pacemakers and minor delays could result in a life or death situation of the patient.
Lesser Network Load:
Cisco estimates that the amount of data processed by IoT devices will reach nearly 7.5 zettabytes by 2020, which is 7.5 followed by 21 zeros! That is a lot of data on the internet highway and can result in increased network congestion, especially in areas where connectivity is weaker. With Edge computing, most of this traffic load will be reduced by processing the data at the source and not sending all of it over the internet.
With the decentralized architecture that Edge computing provides, it is easier for other connected devices in the network to become more resilient. Compare this to a single virtual machine failure on the cloud which would affect thousands if not millions of IoT devices connected to the network. Even if one of the devices fail, it does not affect the other devices and they remain active and operational.
Reduced Data Exposure:
We have seen that Edge computing reduces the amount of data that it sends over the network. In doing so, it also helps in reducing data exposure in transit. In some cases, the sensitive and critical data like Payment Card Industry (PCI) and Personally Identifiable Information (PII) that the smart devices collect need not be transmitted at all. This helps in situations where each country has different regulations for this data and processing the data closer to its source helps in avoiding many privacy, legal and security complications. By further encrypting the data and controlling the access, we can make it more secure against known threats.
Lesser Costs for Data Management:
Storage costs on the cloud is brought down significantly with Edge computing as we are not storing everything on the cloud. This also helps in managing the data efficiently because of relatively lower volumes. Only the aggregate data that requires deeper analysis is sent to cloud, which is subsequently analyzed and extrapolated.
Edge and Cloud working in tandem for IoT
Now that we have seen how edge computing benefits the IoT, it is still important to understand that it is not a replacement for Cloud. To address all the requirement and demands for IoT devices, Cloud and Edge will need to work in tandem. All the data from smart devices and sensors still need to be aggregated at cloud, which would need deeper analysis so that meaningful insights can be drawn out of it. Cloud still plays the key role in making the IoT devices smarter and better.
Let’s go back and take a look at the autonomous cars example.
After collecting the data across all of the vehicles and using Cloud to analyze this, Google can come up with best practices and driving algorithms that will improve their navigation and make the vehicles behave optimally for locations visited for the first time.
Major trucking companies in the USA and Europe are already using this approach to benefit from technology and save major costs. They place sensors in their fleet of vehicles and collect individual data ranging from engine’s performance, tires, fuel levels, transmission and batteries. Processing this data at the edge would be of no use, instead all this is sent over to cloud. After deeper analysis, companies can publish out alerts on picking up best routes to travel, when to replace old parts, vehicles low on fuel that needs refueling, replacing faulty transmission and so on, which improves and saves maintenance, repair and operations costs.
With the enormous compute power that the cloud provides, it makes sense to let it do the heavy lifting on enormous and heavy data sets. Most of the times, the centralized nature of cloud beats the decentralized nature of Edge in terms of speed, cost and scalability. So, to completely address the major needs of IoT, that is latency as well as big data processing, we see that Edge and Cloud needs to work in harmony. Edge takes care of the real time analysis and response while Cloud does the heavy lifting and processing of data sets to improve the functions of these smart devices.
In conclusion, IoT is evolving at a rapid pace and is estimated to do so in the coming years. Though Cloud computing has fueled the growth of IoT, its tandem with more powerful insights into the data collected. This trend is going to continue into the future and with advancement in technology, it will help you better manage and significantly improve your IoT efforts. What is your take on the introduction of Edge computing with Cloud computing in IoT?Edge computing is providing the ability to not just analyze data closer to the source with near zero latency but also provide IOT with deeper and