Cloud and Internet of things: complementing each other, changing each other

With the passage of time, the concept of IoT involves more and more things. Apart from the built-in internal sensors and processors, these things are connected directly to the network and are transmitted online. Although home automation may be the "main" area of the concept, such as the milk in the fridge is gone, the fridge automatically orders milk from the grocery store. But the application of the Internet of things is actually getting bigger. We're going to have a lot of things to interact with and be independent of each other, and the Office will automatically order office items when needed, without our interference, and even our clothes and body sensors will instantly transmit our health data to our doctor. This kind of machine-to-machine (machine to machine) communication is the key point of the Internet of things.

To give full play to the full potential of the Internet, cloud computing must be used as the foundation of the Internet of things. The idea behind the interconnection is that most of the data collected should be transmitted online, and only in this way can the data be pooled, analyzed and utilized effectively. Now let's look back at the fridge example. In this case, instead of ordering the milk from the grocery store itself, the fridge transmits all of its data to the application, including current food stocks and user consumption, and then reads and analyzes the data by application. Then, consider other factors, such as the user's current food budget and how long the milk will be sent to such factors, and then decide whether to buy, and cloud is the ideal destination for these applications.

If all our daily necessities are installed with this sensor, the amount of data generated will be very large. Therefore, the Internet of Things must consider how to store and analyze these generated data. It's not just a matter of data volume, it's also about the speed at which these data are generated. Sensors are generating more and more data, and the data is being generated faster than most business applications.

Cloud based solutions are the basis for the number and speed of data generation. The cloud can automatically and dynamically provide reserve storage resources according to our requirements without human intervention. The cloud also gives us the ability to access virtual storage and access to large pools of storage resources, which are not available locally, through a cloud database cluster or a virtualized physical storage that adjusts capacity without downtime.

The second question about these data is how to deal with them. This problem has two difficulties. The first difficulty is how to deal with all the data points obtained from each of the different objects in real time. The second difficulty is to extract useful information from all the available data points collected, and to correlate the information obtained from different objects to add real value to the stored data.

While real-time processing may seem simple--receive, analyze, and then use--this is not the case in real time. Let's take a look at the fridge, and imagine that every time someone opens the fridge, the fridge sends a packet that includes what's moved and what's in it. We estimate that there are about 2 billion refrigerators in the world, switch the fridge door 4 times a day, then a day will produce 8 billion packets, on average, about 100,000 packets per second, this amount is very alarming. To make things worse, these data points may be focused on the characteristic periods of the day (mainly mornings and evenings). If we are prepared to handle capacity according to the maximum load, then a large amount of infrastructure will be wasted.

Once processed in real time, we will encounter a second difficulty, that is, how to extract useful information from these stored data and make them a step further than personal affairs. If the fridge can automatically place an order for you to the grocery store, this is pretty good for you personally, but if the manufacturer knows that refrigerators from certain areas are overheating, or that the refrigerator that stores certain items is consuming too much, it will be more meaningful for the manufacturer. To extract such information from stored data, we need to leverage existing large data solutions (and some upcoming solutions).

Cloud computing is ideal for dealing with these issues. In the first difficulty, cloud computing allows dynamic allocation (and recycling) of processing resources, allowing applications that require real-time analysis of refrigerator data to cope with these massive amounts of data and to optimize infrastructure costs. In the second difficulty, cloud computing can collaborate with large data solutions.

To sum up, the Internet of things may change the overall architecture of cloud computing, but at the same time cloud computing is critical to achieving this change. In terms of virtualization computing resources, although the application can dynamically allocate these resources without human intervention, there is no development in cloud computing. Because the internet of things is the only driving force for their development.