Interview with Gabriel Montenegro
What are the main challenges to implementing the Internet of Things (IoT) in Latin America and the Caribbean? Is the region ready? What should organizations do? What are the privacy guarantees on the Internet of Things? Are security issues on the rise? Is there a limit to the number of sensors that can be connected to the World Wide Web? Will these sensors be interoperable when “things” are connected to multiple devices?
During LACNIC 26 (http://www.lacnic.net/web/eventos/lacnic26), Gabriel Montenegro, engineer at Microsoft, shared with us his views on the current status of Internet of Things development, as well as the difficulties that may appear with the mass connection of devices to the Internet.
What impact will the IoT have on people’s everyday lives?
It’s important to remember that before the expression “IoT” became popular, “sensor networks” was a very common term used in reference to the devices that are integrated into the environment and have logic (computation), control and communication capabilities. A primary goal of these devices is to detect changes in the environment and to react accordingly. We can expect that the environment will have increasingly greater sensing capabilities. This represents a radical change. Just as with the introduction of the Internet, it is impossible to predict all the implications.
We believe that industrial processes and the world of business will gain in efficiency thanks to the increasing ability to monitor processes and solve issues in a timely manner. Farmers are able to follow their crops, irrigation systems and soils much more closely and thus increase their yield and quality. Electric utility companies are already benefiting from increased visibility into home and building consumption, which enables a finer-grained and more adaptable energy management. In the health sector, the constant monitoring of patients by medical staff (as opposed to annual visits) will help identify and prevent diseases before they become critical. Along these lines, senior citizens living on their own will have increased autonomy and the risks of living alone will be reduced. There is also much expectation about the changes the IoT will bring about in our homes. We are already seeing many intelligent devices such as door locks, sprinkler systems, thermostats, security systems, and central controllers that serve as an interface between users and other devices and even use voice and gesture recognition.
What do you think will change with the advent of massive IoT deployment?
So far what we are seeing smart “things” that exchange information with a very low number of devices (typically one or two). Widespread deployment of IoT devices will lead to networks that will be able to exchange information and coordinate the operation of different devices. We would no longer be talking about interactions with a couple of other things, but rather of mass behavior: coordinated movement (e.g., groups of drones) or coordinated information processing. This will take us from an “Internet of Things” to an “Internet of Networks of Things.”
Another essential component is the processing of all this information and managing all these devices. The scale of this issue is so massive that it will push the evolution of the cloud just to be able to handle all the telemetry. Even so, it will be impossible to save all the data and therefore processing rules and local processing (within a network of devices) will be key for obtaining summaries or determining trends as an alternative to permanently storing the data torrent generated daily. This type of processing on the edge can be applied to images or video as well as to edge analytics.
What are the first applications of the IoT in Latin America and the Caribbean?
Frankly, I don’t have much contact with the status of the adoption of these technologies in Latin America. This is one of the reasons I’m participating in the LACNIC meeting this month of September in San José, Costa Rica.
What are the main technical challenges faced by the IoT today? (Security, privacy, etc.)
Indeed, security and privacy are two of the most pressing issues. During the recent DEF CON conference (August 2016), 47 new vulnerabilities were revealed on 23 devices from 21 different vendors. Keep in mind that this does not only involve software but also devices: locks that can be opened, thermostats that can be hijacked, wheelchairs that can be controlled. This type of attacks can lead to serious damages, theft, fire, and the risk of personal harm. And, of course, many devices lack a proper software upgrade policy (software updates).
Another issue is the enormous amount of information produced and stored in the cloud or in centralized servers. How is privacy of this information guaranteed?
We should also mention that interoperability causes issues not only at protocol level but also in the upper layers: the semantic definition of these devices. There is too much variability in this area, which is why projects such as Open Translators allow programmers to have a consistent API, for example, to control thermostats regardless of their vendors or of the semantic definition originally given to their devices.
As an industry we are not being very responsible about this. There is too much fragmentation in terms of protocols and software implementations. There are too many standards organizations, each with their own protocol stack. Each stack increases the attack surface. A key element to improve upon this situation is to reduce the number of stacks and for that we need to reduce the number of IoT standards. We have already seen good progress with the merging of AllJoyn and OFC, but further work is still needed. The IETF is an important part of this effort.
What is the IETF doing to make the IoT more viable? Is IPv6 the solution?
After RFC4919 and the creation of the 6lowpan Working Group (now the 6lo WG) we realized that IPv6 was part of the solution. We’re now starting to see products on the market. There are various reasons for this:
- Enormous numbers of devices require addressing way beyond the capabilities of IPv4
- Need for autoconfiguration
- Header compression (IPv6 is more compressible than IPv4)
Memory constraints mean that there is often space for just one stack, so the IETF has focused almost exclusively on IPv6 for the IoT.
Another interesting fact is that, when we started the group, various representatives from Asian countries told us that their governments required the use of IPv6.
These efforts have resulted in RFCs such as RFC4944 (basic IPv6 adaptation), RFC6282 (header compression) and RFC6775 (neighbor discovery for 6lowpan), as well as several extensions of these RFCs. There has also been much work on routing and the application layer (CoAP).
To conclude, it’s worth adding that IP is not the solution for every technology. For example, LPWAN technologies currently in use employ very small packets, usually smaller than 20 octets. In this case, the IETF is considering working primarily at the application layer, but little is being said about IP support for LPWAN.