This post is a part of course From beginner to expert in IoT, so if you’re interested, more information you will find inside the link.
I’m describing Internet of Things (IoT) as the Internet of Data. Mostly, each device connected to the internet is a part of IoT. It’s more a concept that extracting “things” that are useful for collecting and predicting new approaches from data.
IoT is most common for a business that creating projects for the smart city, smart home, manufacturing, healthcare or transportation. Devices are fundamentals of IoT and according to statistics more than 30 billion objects will be connected to the Internet by 2020 and will add $14.2 trillion to the economy.
The IoT would encode 50 to 100 trillion objects and it’s mean that each human in an urban environment would be surrounded by 1000 to 5000 trackable objects.
We can describe the device as “thing” or object that doesn’t need any interaction with the human and works independently from them. It’s mean that we need to exclude smartphones or PC’s from IoT even if has sensors or network technologies that normally IoT objects offer.
Devices don’t have one standard to exchange data or communicate together. We can say that they use protocols, such as MQTT to do that.
Most important topics about IoT is security. Without any “defense” against hacker’s attacks, we cannot send sensitive data to the cloud. To protect our data we are using symmetric or asymmetric encryption. Another approach is to use certification, such as X.509 certification.
Nowadays, cloud and IoT are highly correlated, because small devices do not have enough power to extract useful information from data that creating. Therefore the cloud is a useful thing to do that, but when we are adding more and more objects and gathering more data, we need an approach that will help store and use this information.
For business, the cloud has major importance, as a company, we don’t need to build own data centers in physical locations. Microsoft, Google or Amazon creating solutions in many ways, but for IoT are useful PAAS and SAAS approach. For companies that want to use IoT without digging into complex fields by creating a system from pre-built solutions, such as Azure IoT Suite is a good approach.
The industry has created a concept called Big Data, for data sets that are large and complex for traditional software applications to deal with them. Azure provides Streams Analytics to extract information that we need in a scalable way. Sending too many results to cloud causes large expenses. For example, using edge computing is better and less expensive than cloud computing.
Even if Big Data is not enough for us and we want to predict what happens in the next 40 minutes or days, probably we will use Machine Learning to do that. Machine Learning is a concept that provides a sufficient approach to find patterns from data when learning and remember them. More data, the better the chance of finding useful correlations between them.
Machine learning has many ways to learn from data. We can extract supervised learning, unsupervised learning, reinforcement learning or semi-supervised learning. These popular approaches based on different conventions, such as supervised learning algorithm has examples to learn from them.
The mythical blockchain is a new approach to collect and store data. It’s useful when we have a lot of data in decentralized locations and be sure that they aren’t corrupted by anyone. Data are publicly available for everyone and only the owner can “modifying” this information. I marked “modifying” in quotes because once we sent data to the system we cannot change them again. It’s mean that you can add an only new block with data that rewrite information that have the old one.
Another useful feature is a smart contract. If you want to exchange data for funds smart contract will make sure that both sides will get what was in the contract. For me, Tangle and IOTA are the appropriate way to think about IoT and blockchain. They are growing up from blockchain and we can scale new transactions without the drawbacks that plain blockchain has.
Internet of Things is a big subject, but worth to consider. On this blog, I will describe each subject independently digging into more advanced topics for the better understanding whole concept. Over time I will add more topics to expand that are useful in IoT. I recommend starting a reading blog from