In my previous article “cognition - When everything will be connected "an insight into how cognition concepts will play major part of the future holistic interconnectivity is provided. I do provide a similar insight with regard to smart objects and the extended internet.
A myriad of new services and applications will be used in the near future to extend the Internet to the
physical world. This is sometimes referred to as the “ Physical Internet. ” Smart Cities will soon provide useful information to their citizens to improve their quality of life and help them make important daily decisions: environmental data such as air quality, real-time transportation information, emergency assistance, risk of attacks, and so forth. All of this valuable information can be provided to citizens via the public Internet. Other applications will provide data exploited by city departments to more efficiently manage the city such as street light management, water/gas leak detection, or traffic management. These data will be not be made available to citizens and may or may not go through the Internet.
The “ Extended Internet ” refers to intermediate deployment models between the Internet of Things and “ autonomous smart object networks ” : smart object networks are partially or completely connected to the Internet with the appropriate security protection. The basic idea consists of introducing in-band (in-network) data processing in the network while still preserving the notion of an end-to-end principle between application servers. Lets cosider the illustration below;
The core IP infrastructure supports a myriad of applications and interconnects hundreds of thousands or even millions of smart object networks that are characterized by their constrained nature. All of these networks will make use of the IP protocol suite and the network may be connected to the Internet via a fi rewall in charge of securely controlling access to private IP networks from the public Internet. In other words, such architecture prolongs the current Internet (thus the reference to the Extended Internet) just enough to provide access to smart objects that used to be isolated from the Internet. Note that at best such networks used to be reachable from the Internet via complex and diffi cult to manage multiprotocol translation gateways: they are now using IP end to end.
As shown in the illustration , these architectures may require the use of “ proxy engines. ” A proxy engine is a router/computer capable of performing a number of application-level processing tasks to improve the scalability of the Extended Internet.
The Extended Internet model is a true IP end-to-end IP architecture with no protocol translation. If an application requires sending information to an actuator or receiving data from a sensor within the Low-power and Lossy Network (LLN), the IPv6 address is not converted along the data path. Furthermore, since IP is used end to end, the associated semantic is also preserved in support of Quality of Service (QoS), management, routing, security, and so forth. In other words, IP is truly used end to end. But LLNs are not exactly comparable to “ classic ” IP networks due to their constrained nature and their large scale with potentially hundreds of millions of connected IP smart objects. Thus, it may be desirable in some situations to introduce proxy engines within LLNs to perform various tasks such as data collection and aggregation or even in-network data processing.
Consider the example of a Smart City equipped with hundreds of thousands of sensors and actuators
to control environmental factors. One model may consist of collecting all data in a data center for further data mining and processing. Data analysis could then trigger a set of actions and commands that would be sent to actuators. Although fairly simple, such a model is suboptimal in many respects. First, the data fl ows and data traffi c would signifi cantly increase, as traffi c gets closer to the sink/data center, which may affect the overall lifetime of the network. Traffi c congestion would degrade the QoS, but even more important, it would increase energy consumption in the network, which is highly undesirable for battery-operated nodes.
On the other hand, in-band data processing would help increase the overall scalability of the network by an order of magnitude. The idea of in-band data processing consists of introducing data processing modules (proxy engines) in the network that interpret the data and potentially trigger local actions. In other words, distribute the “ intelligence ” in the network. The network of proxy engines then forms an application overlay network embedded in smart object networks and in the Internet.
In conclusion Models such as the Extended Internet involving (dynamic) application overlay networks will likely emerge allowing data processing and local action in the network to further increase the network and application efficiency.
