In last week’s post we primarily dealt with the security issues of Internet of Things (IoT). This week, we will take a look at some issues with making IoT a reality.
1. How is data going to be protected? Is there going to be encryption? What about security at the hardware level?
While it is cheap to create hardware and to connect devices to the network, who is taking care of security? Many manufacturers often want to dismiss these questions. Others simply don’t want to carry the costs of added security. Yet others have already established their minimum security guidelines that satisfies both the clients and their customers. As of today, there are few manufacturers who even think about security at the hardware level – which would increase the costs of manufacturing IOT devices.
In short, security is a huge concern for the Internet of things. It is most likely going to be answered after consumers put considerable pressure on IOT hardware and software makers.
2. Will the devices be interoperable? How can that become a reality?
Imagine the nightmare if all ‘smart’ devices could not work together. (Kind of reminiscent of devices that couldn’t get Bluetooth right or phones having different chargers…)
While the precise interoperability characteristics remain to be decided, it is safe to say they will be based on the technologies in use today. Most will use the standards already in place – USB connectors, Bluetooth and Wi-Fi. Although these technologies will likely evolve. In this respect, consumers will affect what becomes the new normal based on their preferences for devices compatible with the devices and services they already use.
As for different devices “talking” to each other, that will depend on security settings and guidelines that manufacturers and users implement. In general it will take some time to “standardize” all devices, but interoperability of the Internet of Things is crucial if we are to reap the benefits from this new network of objects.
3. What about international standards? Are international standards for collecting data possible?
It took years to get internet (somewhat) standardized, so what’s in store for data analysis?
Complete standardization is something unlikely to happen in the data collection/analysis realm. There is just too much variation. Each device will be created with a specific set of data it wants to collect. Moreover, when collecting their own data, each company decides what data it needs and what it doesn’t and which format will best fit their analysis. Standardization may be forcing unnecessary inefficiency.
Most likely an unspoken minimum standards for data collection will develop. We can see it today already: storage or export in a number of different formats, collecting certain type of information etc.
4. Will there be a different approach to designing objects that utilize real time data vs. historical data?
Absolutely. Applications that rely on real time data will need more specific designs as to what they need to analyze and what conclusions people using these technologies need to draw in order to take action. A medical device that is designed to only mine historical data is unlikely to accurately analyze real-time data needed by doctors to identify a heart attack.
On the other hand, companies that are interested in general trends, or those that don’t know which correlations they are looking for, can take the boarder “collect all, analyze later” approach. They will instead design applications and devices that collect and store data for analysis in the future.
5. Will software needs change? What will the software engineering teams look like?
Every device will now require some sort of software to get it to run, connect, or collect data. Higher demand for custom software alongside globalization and (often) lack of local talent will lead to more distributed teams being created for software development, systems integration and data processing.
The distributed teams will be created either through a local partner elsewhere or by going into new locations. For smaller IOT makers finding a local partner might be more efficient than launching operations in a new location themselves.
For many companies, finding a mature, reliable partner will become vital, since the success of distributed teams will depend on constant communication, developed processes and successful implementation (and understanding) of requirements.
More on borderless software engineering? Download this case study.
