Communication and specifically networking has brought a huge change in our lifestyles. People are increasingly getting more and more dependent on networks, networked devices, and the services provided by them. People, in the near future, will get even more connected and dependent on the ICT systems. Higher the dependency, higher is the complexity of the desired services from these systems.
This dependency will drive people to expect the services everywhere, which may not be feasible always. Thus, when a particular service is unavailable at a location, one has to find some form of service that is as similar as possible to the one sought after. We envisage here a dynamic, approximate service composition. It is expected to satisfy the users under the given circumstance though it is not a complete service. Thus, these approximate services may be seen as a conglomeration of different functions provided by different systems, entities or devices. We need to be ready to explore the possibilities of such services and architecture which are not rigidly defined.
Mark Weiser almost 20 years ago provided his vision of the ubiquitous computing: “The most profound technologies are those that disappear. They weave themselves into the fabric of everyday life until they are indistinguishable from it.” This vision is in fact the driving force in the current context of miniaturized technologies and communication substrates. The idea is to enable seamless and interoperable connectivity amongst heterogeneous number of devices and systems, hide their complexity to the user while providing sophisticated services and applications. The challenge is to connect, collect and distribute information amongst devices with different degrees of complexity and capability - from simple and passive RFID tags to multi-front-end high speed networked and mobility enabled Personal Digital Assistant (PDA). Ideally, the Future Internet should be able to classify and represent information of communicating and non-communicating devices, objects and entities, which are present in the physical world.
“1000 devices per person” paradigm is indeed a much discussed topic in recent years. However, there is no denying that the number of devices per person is growing steadfast. The expectation is to bring more efficiency in the use of all the resources that humans use now while boosting the quality of life. Whether it is 1000 or 100 devices per person, the complexity of networking and maintaining them is definitely a need of the hour. Moreover, each of these devices that support a person will not be complete in, however, but they cooperate and support each other to form an end-user service.
Given the explosion of ICT enabled devices, we need to address many tough questions such as, management, availability, accessibility, service provisioning, etc. We envisage that in near future we will see many of the objects would have some form of presence in the digital world enabled by wireless interface on the device, or an RFID tag, or due to (nano-) sensors in the surrounding – thus these resources are decentralized and are in the periphery concentrating around small pockets such as homes, offices, cars, etc. The sensors and many other installations in the surrounding will be capable of providing some services to us. In the past we have seen many projects and research initiatives, for example Personal Networks and Federations, which is context aware, secure and self-organizing.
If we define “function” as the individual support offered by a device, “service” would be a conglomeration of such functions. Some of the functions may be re-used in a way that may not be conceived at its creation. These devices surrounding us and supporting us will be part of the Internet of Things (IoT). With many devices and services around a person, the first question is how to locate a service. Ontologies of devices and services are one method to classify, and locate services. While these are standard services, we conceive of IoT enabling services similar to humans i.e., approximate services. For example, the system should help a user locate a hard object (he needs a hammer) when a hammer is not in his vicinity. It requires a clear vision and also flexibility to provide services similar to humans.
Future Internet research
The paradigm of Future Internet has been investigated in great details by many European FP7 projects, such as FIRE, FIND, GENI, etc. They describe the application of concepts like large scale networking, Cognitive Networking (including Cognitive Radios), network of networks, as well as architectures developed for a converged communication and infrastructure services. The European commission has taken a big step in identifying issues and encouraging new and innovative ideas towards the Mark Weiser’s vision under FIRE initiative. NSF also dealt with the Future Internet initiatives in Future Internet Design , which is focused on designing future networks that are more secure and available than today's Internet or by ensuring that functions like information dissemination, location management or identity management fitting in new design and environment. FIND also investigates how economics and technology can interact to shape the overall design of a future network. The NSF is also taking further steps to address the scalability of the future Internet by various means.
While Future Internet research, development and deployment are going at a full pace, academics and industrialists are asking the question as to how to use promised futuristic networking capabilities. How to make the services offered to users better than before? Without a clear adjudication on this question, industry will not support the above initiatives. A critical challenge is the management of the future Internet infrastructure, platforms and the vast amount of information, which is collected, distributed and stored. We have to evolve from the old Client-Server paradigm and Service Oriented Architecture, to new architectures which are at the same time both scalable and reliable while being flexible and dynamic in nature. In this context, we can clearly see that many cognitive and human decisions like aspects need to be addressed in the near future
Humans can use instruments, objects, and all the available resources - physical or intangible (knowledge, experience, etc.) - adaptively in any situation. Humans derive applications based on the context and availability owing to previous knowledge. Thus, many times they indeed get services which are not exact but still are able to carry on with available and possible services. Let us take an example. A table knife that we use daily has one defined view – that is to cut the bread. However, it could be used by humans, in case required, for many purposes. For example, it can act as a screw driver, and it could also be an envelope opener. In a small area if one could somehow find an apple and if this information could be given to person who is hungry, then the purpose is served. Though this is trivial for us, it is not hard to imagine the complexity involved in enabling this in the digital world. The question that is driving us now:
“Is it possible to use the Internet (of Things) depending on the requirements and the context, just as we humans do?” This sort of service provisioning has an advantage of using the system that is nearby and can resolve the requirements to an extent possible.
Every object (a representation of a physical entity; we shall restrict for the time being an object as a simple Internet enabled device) can provide certain functionality. A service can, thus, be defined as a set of re-usable functionalities with a pre-described interface with constraints and policies as specified in its service description.
The foremost requirement in the near future is to find services and may be even searching physical objects. Further, if the search is for a tangible object or service (not a virtual service in the Internet) searching in the vicinity of localized search space first for the service. However, finding the best and perfectly matching service requires a lot of time and infrastructure support. Most often than not, we tend to get a rough service (close to what we expect). A case in point is that when we search the web, we do not always find an exact answer to what we have looked for. Thus, we define the term Approximate Service, in very simple terms as, a service that is a conglomeration of various functions offered by a single or multiple object to satisfy a user in real-time.
It is expected that, in the near future, multiple devices will offer multiple sets of functionalities and they have to be stitched together to satisfy the exact requirements. However, when such a possibility of completely satisfying a user doesn’t exist what should be done? This is where a grand new vision of approximate service provisioning comes in. Unless the services are provided seamlessly with whatever the level of satisfaction that could be achieved, the future Internet will not be able to evolve into a substrate in our lives. The concept of Approximate Service Provisioning, thus, will make the future Internet a truly invisible network.