The Internet is now more than 40 years old. It may even be past its ‘best before’ date. Although it’s hard to question the Internet’s value and benefits, technical experts are starting to point out its limitations as a communications platform for future systems. Ram D. Sriram, chief of the NIST Software and Systems Division, suggested in a recent webinar that a new unifying framework is needed to guide the development of the next generation of smart systems.
Existing architectures, systems and networks, including the Internet, will need to be re-examined, re-organized and even re-fitted to fully support emerging ‘Internet of Everything’ applications. Other commentators have also considered this requirement. See, for example, this ETSI report on market drivers for next generation protocols and Martin Geddes’ comments on the Internet in his post on Recursive Inter Networking Architecture.
But do we really need another framework and more ICT architectures? If we do, are the stars properly aligned for success now?
Towards a more inclusive ICT framework
The starting point for any ICT framework is an ontology, which is a formalized set of concepts and terminology relevant to a field of endeavour. Then reference architecture are designed to help limit complexity, organize information and serve as a baseline for solution-specific designs. One example is the NIST and the ISO reference architecture for cloud computing. A similar framework for the Internet of Things (IoT) is currently a work in progress. But, to date, there has been no agreement on an overarching framework that ties ‘everything’ together.
The Internet of Everything (IoE) takes this idea of a framework to this next level – it expands the IoT framework to include individuals and social groups. Dr. Sriram identified three major IoE subsets: Cyber Physical Systems (i.e., networks of things); Cyber Physical Human Systems (networks of things and organizations); and Cyber Social Networks (social networks of people). Dr. Sriram calls the whole “Smart Networked Systems and Societies” (SNSS).
A framework such as the SNSS can be quite useful as a guide for standardization, product development and industry alignment, especially in an environment that is ‘multi’ (multivendor, multinational, multi-technology, multipurpose and multi-tenant). A template for the three SNSS subsets is illustrated in the diagram below.
Key webinar takeaways
Dr. Sriram packs a lot of content into his one hour webinar (the 87 slides are here). His key conclusion, however, was that all ‘things’ need to be smarter, mobile and interoperable.
‘Things’ are getting smarter
Systems of things are smarter when they interact well with humans, when they are adaptive and composable, when they collect data from diverse sources or when they are context-and situation-aware.
Many physical things (from cameras to cars to hyper-scale, multi-cloud data centres) now have the potential to be cyber-physical and smart; that is, they are software-defined, Internet-connected and paired with one or more cloud applications.
‘Things’ need to be mobile and wireless
The days when terminals were tied to a desktop with a thick cable are long gone. Smartphones (or, as Dr. Sriram calls them, mobile personal communications devices - mPCDs) are associated with individuals, not places, and must be able to move with their users. mPCDs can also serve as smart gateways for other personal devices such as watches, fitness trackers or implanted medical devices.
IoT things (a car’s engine temperature sensor, for example) may not be a personal device but must still be connected wirelessly and able to travel.
Portability is a key component of mobility that is enabled through software. Virtualization and containerization allow cyber physical systems to be more easily re-located to different environments.
‘Things’ should interoperate
Cyber physical systems components will typically be developed and operated by multiple different providers. Interoperability and standardization are needed, almost by definition, for highly networked systems. For example, medical systems must support a wide range of devices from implanted sensors to hospital rooms to imaging machines.
The three levels of interoperability are physical and logical connectivity (i.e., a path for the data to be transferred), syntactic compatibility (i.e., the languages/character sets can be decoded) and semantic understanding (i.e., the meaning of messages is correctly interpreted).
In addition, system-level capabilities for access, security, privacy, identity, quality and manageability involve functional interoperability.
Key research challenges
An important benefit of an ICT framework is that it can help identify technology gaps – in interoperability, for example, and provide clearer focus for research and development requirements.
Dr. Sriram highlighted seven critical research challenges for SNSS: security/privacy, interoperability, knowledge representation, data analytics, network behaviours, human-computer interaction and storage architectures.
Under these topic areas, he identified four specific needs:
- Modeling events and their relationships in time and space: life events, for example, that go beyond a simple time-ordered calendar;
- Re-thinking sensors to cover a much wider range of things to sense - for example, sensing in the cyber world in addition to the physical world;
- Knowledge representation – ontology development and testing; and
- Big Data and decision-making tools and techniques.
Benefits of a framework
Hosting non-traditional devices on the Internet is not new - a toaster was connected as early as 1990. However, advance has been very rapid and analysts now predict there will be more than 20 billion connected devices by 2020. A framework that knits all the moving parts together at this scale will accelerate progress and lead to the development of more efficient solutions.
The traditional Internet vision of connecting people has given way to a much more complicated and evolving concept that involves multivendor suites of sensors, actuators, edge processors, smart networks and cloud-based ICT services. Dr. Sriram suggests a paradigm shift is occurring that will involve “symbiotic networks of people (social networks), smart devices, smartphones or mobile personal computing and communications devices that will form smart net-centric systems and societies (SNSS) or the Internet of Everything.” A unifying framework that can guide the architectures, standards and best practices to support this vision would help to ensure global compatibility and interoperability.
Any framework that gains significant industry support can help to realize the Internet of Everything. The real question is whether Dr. Sriram’s Smart Networked Systems and Societies is the right answer in the right place at the right time.