Question 1: lack of understanding of the data
Your hands hold a lot of data does not mean that you can understand and utilize them. Because over a wide area in the case of networking equipment, multi-species, a supplier is difficult to give the best solution for the entire warehouse. Even if someone came out, his program is also likely to be overthrown again, this alone had cost far too high up.
Want to build a warehouse for the efficient and secure networking environment, it is necessary to establish a P2P network, each device can provide a business exchange and communication equipment and other business information. However, in so many suppliers and warehouses also retained some traditional equipment situation. The program is almost unrealistic. If you can not solve this problem better, then Internet of Things warehouse designers face is simply a modern Tower of Babel.
To ensure that all data communication can understand each other and maximize the effectiveness of the case, one possible solution is to establish a hub model. We can build a lot of Internet of Things via the central server is responsible for the entire period of receiving data from various devices and sensors. The rules engine responsible for analyzing the data, then the hub of the correct instruction is sent to the controller to accept, for example, due to higher sunlight southeastern external wall temperature, need to immediately lower the temperature of refrigerated area 2.
This hub must be able to translate the translation of different kinds of data and units, for example, from Celsius to Fahrenheit. It must also be equipped with a common data model, so as to compare and integrate data from different information from unfair suppliers of equipment, so as to ensure the system can “understand the data.”
Question 2: the amount of information is too great
In some cases, since the amount of data the system is too great, to move data to a central server via a network is simply impossible. For example, only one sensor is mounted on the wall of the warehouse, the data it had collected to temperature, humidity, hardware version, software version, remaining power, position changes, etc., if you want to hear I have a long list of I can speak.
This information may be updated every 30 seconds, or may be due to environmental requirements, have a few seconds to update again. However, due to the amount of information is too great, all these data are sent to the central server is not possible. And the whole warehouse is also more than this one sensor, all add up to about many thousands; they may not even models are not the same.
This time we need information integration solutions, to ensure the system can filter out the necessary information, is converted into a common data model, then release reports, maintenance and other instructions. For example, we can only be the case in the warehouse area of â€‹â€‹the sensor 3 through 50 on the outer wall of the entire region and beyond, it is determined that the temperature equilibrium 3.
Question 3: Security
Before we say, P2P model Internet of Things for us in the warehouse or in the case of any large area networking use cases are essential. But this program will bring a huge security risk.
Security of the entire system depends on the system-wide security worst that device. If one device security provider is poor, then the equipment other vendors and then security is useless; a device error, might cause unexpected butterfly effect. For chestnuts, a security vulnerability device to device may report errors outdoor temperature, resulting in equipment issued thermostat instruction error, temperature error of the entire region, so the region where the food is also all bad.
Want to solve this problem, the entire warehouse P2P model must take Internet of Things a certain way, so that the system confirm the sensor data through a sensor close to double-check the sensor to the data. For chestnuts, if one outdoor sensor measured temperature is particularly high, and it is near the temperature sensor to it but had generally lower than when the system should not only immediately issued instructions for adjusting the temperature of the sensor to the data. The system should send an alert, verify and confirm the reliability of the sensor and its peripheral sensors to compare the data again.
By confirming the perimeter sensor data to verify data reliability is a very practical way. In addition, we can also let the system by reviewing historical anomalies in the data readings to determine whether these anomalies and climatic conditions, inventories, year month, day time and other factors.
Question 4: Equipment problems
Internet of Things also a drawback, I called “neuropathy devices”, refers to the networking device or sensor failure suddenly and inexplicably began to send false readings to the system. “Neurological” symptoms of this device has many potential reasons; most likely have a software bug, level is low, or the device itself is flawed, and so on. Relatively rare reason may also be no small renovation work to paint spilled on the device, a piece of plywood covering the sensor.
Although these “neurological devices” does not cause system external security threats, but their destructive power is immeasurable. For example, in the case of our warehouse, if the administrator does not do the insurance policy in accordance with our previous proposal, then this failure might make an area full of food spoilage. As we said before, beware security approach these “neurological devices” and improve system can be generic. Comparative History abnormal readings confirm perimeter sensor readings, these programs can effectively prevent the system to make devastating decisions or instructions issued danger.
Many companies believe that Internet of Things are a business, they can put Internet of Things apply to existing products. But before these ambitious plans come true, Internet of Things designers must clearly Internet of Things drawbacks and risks. You know what the easiest and most important way is it? First, be careful to choose your supplier. Remember to choose can integrate and extend the platform to develop a common data model. In short every step sure to be careful, to anticipate worst-case scenario, to create the most perfect solution. With these, you can succeed.