Deploying Internet Of Things in practice
Having chosen the scenario where it makes sense functional automation of information flow between devices according the needs of our business, we have to analyze what platforms will take as a base for deployment. In this post we will discuss the requirements for choosing the platforms, protocols to use, how to optimally structure data, security devices and enforceable minimum requirements for this deployment will become a success. In addition, article explain a case of deployment.
Currently IoT (Internet of Things) is an acronym related to devices connected.
Making history, Kevin Ashton (Professor, MIT) IoT defined as follows: “The IoT is the world in which each object has its own virtual identity and potential ability to integrate and interact independently in the network with any other individual either a machine (M2M) or human. “
When we read this phrase comes to mind the world of Matrix right?
IoT charges a leading role in the digital business from smart logistics, smart operation, active marketing …
For example imagine that our future car has internet connection with some of the following characteristics:
- Adaptation of the vehicle through our mobile application or sensors included therein.
- Wireless gateways traffic conditions in real time or temperature.
- Communication breakdown or actual impact on our insurance reporting a vehicle problem while.
Some of these new features may change the decisions we make in the future as modifying the route options take before or during a trip, making the right choose if we are in the middle of the city or choosing an alternative route to reach the destination? When clicking the wheel in the journey, the same vehicle can send the characteristics of the wheel by a message and the exact position of the vehicle to facilitate repair.
To bring to reality all these advantages, once you have defined the setting and functional purpose, we have to design the right technology architecture, considering that this is a technological world in constant boiling and movement.
Ask yourself the right questions before choosing the best option to decide the right platform
There are many questions that could help in decision-making:
- How many devices will take place in the new platform?
- What types of networks we use?
- What assurance can apply?
- What volumes should support the platform?
- What protocols will be used for the integration of devices and systems?
- Where standards evolve, which are consolidated or disappear?
- What is the human interaction in any of the processes?
The “state of the art” in the industry of electronic components ranging from processors, sensors, batteries, connectors or adapters network has reached a high degree of optimization to result in any imaginable device to include in our “smarter products” with degrees of miniaturization and connectivity growing and increasingly lower costs. Examples of devices such as Arduino, Spark, Intel Galileo, Waspmote among others.
The scenarios we want to implement IoT have different caseloads. Some projects can be internal within a facility that has a network connection quality while other devices because of their location or mobility may require the use of other types of private / public networks like GPRS, EDGE, 3G, 4G or combined.
Security and Privacy
Of all the topics of discussion it is the most important issue.
Can you imagine what would happen if our smart car was hacked and controlled remotely? For this issue we don’t pass have to think securize devices and platforms:
So that devices have the ability secure login by fingerprint or voice.
Encryption of communications between devices and platforms (gateways).
Records of devices and special authorizations included in the gateways.
Operational intelligence to know if a user is using a service with one device at a time in two different locations.
Knowledge of the various protocols will allow us to make better decisions in adapting communications between devices and gateways. Among the emerging we are:
MQTT / MQTTS: oriented devices that consume very little bandwidth and low reliability. It allows to implement a bidirectional model under publish / subscribe.
CoAP: Implements a base model equivalent to HTTP eliminating the need UDP session.
REST API: Used for both synchronous communications hubs and gateways.
XMPP: Very suitable for high volume and asynchronous models.
Volume, monitoring, data types, implementation and maintenance costs, are other important issues to consider.
For example, what data structures have to use? Do we use as standard XML structures, light as JSON or binary data?