IoT: Benefit or Headache for IT?
The business issues may be even more difficult to resolve than the technical challenges.
The Internet of Things is arriving in many environments. Connecting to and supporting IoT can be performed by an enterprise or through cloud services. In either case, the challenges will be the same.
IoT endpoints will be measured in the billions. But these billions of IoT endpoints do not include PCs, tablets, and smartphones, which IT will still have to manage. The traditional endpoints will also contribute part of the traffic that networks will have to continue to carry in addition to the IoT endpoint traffic.
IoT traffic falls into two categories: passive data collection; and active control and response to the data collected. When it is passive and no immediate response is required, then the network does not need to be that fast and response time is not an issue. On the other hand, when responses are required, then the network will have to be fast enough so that the responses can control the endpoint rapidly--like stopping an escalator, changing traffic lights, sending out notifications for emergency services, turning off failing equipment, and predicting when maintenance is required.
An example of responsive IoT was presented at the Microsoft World Partner Conference in Washington D.C. where participants discussed the IoT implementation for the London Underground rail system. A short video is available, "Bringing the Internet of Things to the London Underground," which explains what this application of IoT means for the rail system.
The traffic that IoT generates will come in many forms. One form is a continuous stream of short messages, traffic which is less common on IP networks today. Another form of traffic will be from wearable devices that communicate over wireless connections. This traffic will impact both Wi-Fi and cellular networks. The continuous stream of short messages will either mean that these wearable devices will be always connected or that each message will require a separate connection. I think that the wireless traffic on cell networks will require some network changes as well as a new charge structure to support IoT endpoints.
Organizations are already pursuing IoT solutions, but there is limited experience. Performing tests in the lab delivers a proof of concept, but making IoT work in the real world is another matter. Many articles and blogs have focused on the promises of IoT, but there will be technical issues to surmount. And the business issues may be even more difficult to resolve.
Since IoT will collect information on individuals, the privacy and security issues will probably be more problematic than the technical issues. What we in the U.S. consider vs. how the rest of the world approaches privacy and security will make it more difficult to deploy IoT. Even if an organization is domestic, employees may travel outside the U.S., and domestic products and services can be operated internationally. How does an organization implement a system so that it complies with all the international regulations? This may foster the growth of IoT cloud services where the privacy and security differences can be supported by a third party.
The list below covers many, but not all, of the technical issues that IT and networks will have to consider in their design to support IoT. The list demonstrates that a wide range of considerations must be taken into account to deliver a successful and efficient implementation.
• Standard interfaces for APIs
• Secure access control and endpoint/user authentication
• Identifying the endpoint accurately to avoid errors and spoofing
• Presence detection
• IP address management--most organizations will have 3 to 6 times more addresses to manage
• Cost implications of the move to IPv6 for all organizations as well as the service providers
• SIP may be the signaling protocol--but that is not guaranteed
• Standard message format(s) will be needed
• Standard data codes will be needed
• Adequate bandwidth to handle the increased traffic, also the high overhead encountered with short messages, and the keep-alive messages for presence and endpoint management
• Endpoint power consumption will be an issue, especially for wearable technology
• Endpoint software management and revisions
• Software error resolution
The organization that chooses to implement IoT will have a number of business issues to contend with, and that will vary by industry and geography:
• Security in access, transmission, and storage
• Privacy regulations which are different in the U.S. and other countries
• Safety concerns when the IoT endpoint is part of an emergency/failure notification system
• Liabilities that the endpoint vendor, network service provider, and/or cloud service will or will not accept (they'll usually attempt to accept as little as possible)
• Bring Your Own Thing (BYOT) policies and management; employees may own some of the wearable endpoints
• Who owns the data produced
• Will the data collector ensure responsible use of the data?
• Service level agreements (consider the recent 9-hour failure of the Microsoft cloud; could this produce major costs and liabilities?)
The above issues may slow the adoption of IoT, but they will not stop its implementation. More issues will be discovered as IoT implementation expands. Managing billions of IoT endpoints will have benefits as well as much vulnerability. Think of a hacker that decides to penalize the London Underground for some perceived infraction by changing the operation of the escalators. That would be a major news item.
The Internet of Things has stimulated the formalization of ideas, concepts, and standards at the Institute of Electrical and Electronics Engineers (IEEE).The IEEE held the "First IEEE ICCC International Workshop on Internet of Things (IOT 2013), Xi'an, China".
The article presents a number of ongoing efforts that the IEEE members deem important to resolve.