Major Existing Classification Matrices and Future Directions for Internet of Things

Classification method is a formula, logical description generalizing characteristics of objects of related area. Nowadays, billions of smart objects are immersed in the environment, sensing, interacting, and cooperating with each other to enable efficient services. When we think about IoT, classification is a major challenge particularly if our technology is international level applicable. So, this limitation needs clear and deep analysis of the existing classification matrixes and gives some future directions depending on the different researches in the area. The paper surveys the current state-of-art in the classification of IoT. First, we try to explain commonly existing classification matrixes; Second, cooperation of different methods defending on classification matrixes used. Then analyses challenges that IoT faced from classification angle and finally we give some direction for future IoT classification.


Introduction
It has been estimated that the Internet of Things (IoT) will contain 26 billion devices by 2020 (according to Gartner, Inc.),As a result, recent problems and challenges arise spanning classification of this newly fast developing technology.
The major reason is that Classic classification matrixes are not sufficient to solve this unprecedented issue, and need to be revised to address the complex re- The reason why we interested to do this analysis is as we believe solving classification problem will result in big solution for problems IoT technology's suffering nowadays; like, problems in terms of resource capabilities, lifespan and communication technologies, new standard design if needed in the future and security.
This paper surveys IoT classification matrix and classification related issues nowadays IoT faced.To summarize, this paper address the following main points: • Discusses various classification matrixes which are in use now a day.
• Explains state-of-art of classification technics which are very important in internet of things world.
• Discuses in between currently used IoT classification matrices.
• Current major problems in IoT world because of classification reasons are explained.
• Finally, we end by suggesting our basic classification matrix.
For simplification, we divided this paper in to five various parts: Part one is introduction which is explained above, Part two different classification methods used, Section three discuses and comparison among the IoT Classifications in Table 1, in Section four and five discuses some challenges IoT facing due to classification and future consideration during classification standard is design.

Existing Classification Methods
For doing this research we searched in different publications and ended up with limited number of literatures about IoT classification [1] [2] [3] [4] [5], which is indeed still in its inception phase.This is very important because understanding the existing classification method result in understanding the problems with the existing once and give directions how to solve.This section deals with higher level classification matrixes only.

Different scientific community have different view on classification of IoT
generally but as our study indicate that there are two major IoT classification ways are used today's world, namely, the classic (based on history of IoT) and systematically analyzed and studded (factor dependent classification).
In a classic or traditional classification method, it is a classification depending specifically on single factor like real-life application, standard used, application environment, way of data communication, level of smartness, specific devices with whom it communicates, or depending on the end users.In other word, it is a type of IoT classification just only considering single and simple factor.However, the systematically studded and analyzed one is a classification method by which the developer of IoT will analyses from different perspectives before classifying his/her product.Which means classification in which not depending on technique of creating new values and services through utilization of information.
ZIT (IPC,FI and CPC system is not enough) SO [3] classification according to the concepts of creator and purpose (creator and purpose) self-made: personal purpose.

ready-made: industrial company
Smart objects as building block for IoT [16] Awareness, Representation Interaction Activity aware, police aware, process aware objects Design and implementation of framework for building distributed smart objects system [4] Operation based classification SODD : smart object description document , profile description document (PDD) User Innovation for the Internet of Things [17] User innovation and market based innovation (user centered ecosystem) User-led and market based innovation Internet of Things Tectonics [11] Infrastructure ecosystem hardware side (collection of connected devices) Enterprise and consumer application, industrial automation, entire stack of infrastructure beneath those devices Corsaro sort IoT [12] Any application (collect-store-analyses-share) Consumer (CIoT) and Industrial (IIoT) Web [10] Based on user advantage Wearables, Media, Home automation, Smart appliances.
Internet of things ecosystem [13] With or without IP address internet of people.

Internet of people and internet of things with IP based or not
Classification based on human [18] Functionality of the technology, industrial or consumer dependent Embeddable, Wearables, Moldable, Surmountable single specific factor.But, most of the difficulty's facing this classification method are: 1) until today there is no specifically documented material so that every developer can refer before classifying its product.2) even the existing once are designed by specific organization (industries) to meet only their requirements which is very difficult to use as worldwide.It is known that there are several classification methods, but here only two of the above are considered.

2) Factor Dependent Classification
A IoT classification method where experts in the area follow some specific matrices to classify their technology.These specific matrices are designed for specific organization or deigned for international use by international industries in the area.The widespread problem that most of this classification method sparring is highly depending on the classic classification method and every organization have its own matrixes.There are different views on the Internet of Things paradigm coming from various scientific communities.Below we try to analysis Common ones:

SO (Smart Object) Classification Model
Which is also known as IoT management architectures, it is a type of classification of IoT s depending specific matrix called smartness.Here classification is only depending on the traditional way of classification which accept the logic IoT is synonym to smart object.Under this classification there are dozens of different classification methods used.
Some authors call SO's as police aware, activity aware, process aware object [6].They used list of design dimensions where every SO's type is characterized among them: 1) ability to understand events from sounding (environmental or human event) which is called Awareness.2) considering the programing model of SO called representation.3) way of communication with its users called interaction.Major limitation with this classification is it not operational, only design dimension based.Creator and purpose based classification of SO's [3].Creator: an individual creates SOs for personal use (self-made).Industrial company: creates SOs for business (ready-made).However, still considering two dimensions (creator and purpose) but IoT classification needs more factors than used here.In [4] smart object description (SODD) and profile description document (PDD) it is another two-dimension (matrix) classification method.Under SODD meta data of SO like list of name, vendor, and profiles.On the other side, PDD is profile specifier (detector or actuator).Limited to only management and implementation which is specifically FedNet middleware.Metadata model [5] [7] use as a factor type, service, device, and location.Which are generic for division and used only for Smart Search, discovery and dynamic as their main limitation.SO is cyberpysical object (sensing, processing, storing, and network capability) [8].
still in its inception phase.
onomy with the human at center of all.
1) Embeddable: things in the user.

Power Source for Internet-of-Things [14]
Classification factor based on technology applications.To generalize, as mentioned by many studs when we come to types or classification matrices of IoT everyone has mixed feelings which by itself make difficult the way how to classify and name the term "Internet of Things".Some of the developers describe the term "everything and nothing" [10] because of no defined criteria used for classifying.Even if we had many types of IoT just using classic way of classification the problem is most of classifications are mixed and repeated.

Future Directions to Be Considered When Developing or Designing IoT Classification Factors
As we believe that system's complexity doesn't matter in IoT.The important aspect that identifies an object is its capability to connect to the network and exchange information without any defect.But the reality now a day is far different.
As described in our study the solution is beside the above factors we should think about: 1) Security: global connectedness is a key reason for security threats.It is known that in IoT world everything connected affect everything.In fact highly secure islands of very sensitive information are typically not connected at all to IoT world [19].
2) Safety: error in information processing part of the system can spread in to Or how can we give ordinary persons a voice?How can we insure that IoT allows for user-led innovation?

Lesson Learned and Open Issues
It is known that internet of things is a big idea full of complex technologies.
That's why we need a common ground to classify these complex technologies to do so every company working in the area must come up with common ground • Clarifying IoT and M2M: M2M is often used as a synonym for IoT, particularly in the IoT world.But while similar, there are differences in which rungs of the IoT ecosystem ladder they occupy.
quirements imposed by IoT.This problem, classification matrix, led us to analyses how the current IoT can be classified.If someone develops a new IoT technology then how he/she can classify, what are the existing classification matrix and how they are effective.
Type of classification based on available technology's and standards.Communication: (infrared, Bluetooth, radio frequency identification, ZigBee, high speed LAN) Identification: (biometry and object tracking) Location tracing (advancement in RFID and GPS) Sensor (control of temperature, MEMS, motion sensor and image sensor)Devices: (RFID tagging, mobile phone and embedded portable electronic devices) used for classification.Some of open issues are: • Many products play role in multiple categories.For a given technology it is important to ask a user what am I buying this instead of?What does it sit next to on the shelf?Most of us being very market focused.• IoT technologies have Lack of the environment that Clear Strategy in Becoming Smarter.One key obstacle is citizens themselves, who don't see the value in them.

Table 1 .
Major existing IoT classifications and their matrices.
. Smart service: protocols or new techniques which advance use of IoT.F. Smart object: any device from any classifications described above.We all agree that using such naming is not the problem.The problem arises when we think about what are the specific classification matrices, in this case just only application area.And more now a day's millions of new IoT technology immersing the market and this kind of classification not support from different angle.
A. Smart city: if the IoT technology is used for city modernization.B. Smart farm: IoT technologies for farming C. Smart health-care: IoT technologies in health area.K. M. Basher et al.DOI: 10.4236/ait.2017.74008115 Advances in Internet of Things D. Smart transport: IoT technologies in transport.E