Considerations and use case for massive IoT

IoT is touted as one of the next big things in the wireless technology industry.

IoT devices are in general characterized by low power consumption and bandwidth requirements. They may need to transmit only a few bytes of information sporadically either in case of certain events or to keep alive. The battery life requirements correspondingly can protract to years and even decades.These devices could be embedded in the appliances of day-to-day use or be planted standalone along with sensors,etc.

Following are a few current wireless technologies that can be used for such IoT requirements. Each of them is designed for operation only in specific licensed or unlicensed frequency bands. The range of operation of these technologies depends on the frequency of operation as well as other design choices that impact the link budget.

Standard	Range
IEEE 802.15.4	Short
IEEE 802.11ah	Long
IEEE 802.11ax	Short
BLE	Short
3GPP (e)MTC	Cellular
3GPP (e)NB-IOT	Cellular

Managing multiple devices such that they do not interfere with each other requires centralized coordination of resources. Further, whether these devices operate in licensed or unlicensed spectrum determines the guarantee with which transmission resources can be allocated.

One important consideration in how many such devices can be supported in a given range is the minimum time/frequency unit of allocation. The minimum transmission interval also affects the device power consumption. The minimum time unit is a function of the symbol duration of the technology, time scheduling schemes and the cell range which affects the RTT. The minimum frequency unit is a function of the subcarrier spacing as well as frequency multiplexing schemes. The symbol duration and subcarrier spacing are reciprocals of each other and the design choice depends on factors like the delay spread of the channel, targeted tolerance to mobility and frequency errors as well as spectral efficiency.

Scheduling of transmission from multiple users onto these time frequency resources can be done either by transmission of dedicated scheduling information for every scheduled device or by contention-based resource selection by the devices. If dedicated scheduling information is transmitted, the amount of information and its efficiency will also impact the number of the devices that can be supported. In addition, it is not optimal for the device power consumption if it needs to spend a lot of time to decode this scheduling information and then transmit per it.

IMT-2020 defines a new use case called massive Machine type communications (mMTC). In this use case, it is envisioned to support a million of such devices per square kilometre. This amounts to 1 such device deployed per square metre or for a populous city with a population density of say 10000 people per square kilometre, about 100 such devices deployed per person. While the need for such abundance of IoT devices can be reasoned against, this would allow them to enter such articles of day-to-day use as fans, refrigerators, water sprinklers, water heaters, fitness and health care devices, CCTV cameras, door alarms, clothing etc.  Among more critical applications, they can assist in agriculture by sensing irregularities in crop conditions say in terms of moisture, infestation, maturity, etc and cause relevant alerts. Companies have created even scarves and shirts laced with sensors which when empowered with an IoT device can transmit information as configured. Such applications are already in use albeit with the existing communication technologies. However, new candidate mMTC technologies are being developed to provide a solution for the extremely high device density and low power requirements envisioned in mMTC.

So, the mMTC use case would assist the centrally coordinated deployability and operation of a vast multitude of such devices rather than creating extremely novel applications. It might give a huge boost to wearable technology. But the applications may only be an extension of the already prevalent ones.

For example, consider the case of a BLE tracker tag which can be attached on to pets, children or even personal belongings and the proximity of which can be tracked using a smart device. The tracker tag may transmit periodic keep-alive signals with such transmit power that will be detected by the smart device only within a configured range. The absence of such a detection at the smart device for a threshold time duration will cause an alarm to be activated. This application can be enhanced or extended in the following ways:

1. The periodic transmission by the mobile tracker tag can be reduced to a need-based transmission which will optimize the power consumption at the tag.
2. The information that is transferred can be enhanced to include more insightful information such as current location transmitted over a period of time after the trigger is activated
3. The information can be transferred to a network which an distribute it to a much wider audience than a single smart device
4. The management of this problem can be made network-centralized and subscription-based in that multiple such proximity trackers (consisting of one tracked unit and one tracker unit) can be installed and once the tracked unit stops hearing a reference transmission, it can attempt doing a network association and start sending the relevant information which in turn can be distributed by the network or even handled by a central agency
5. The management by a network has other benefits like coordination of resources such that collisions on the same channel can be avoided or efficient allocations on the same channel can be made.

Such enhancements and for all the different existing applications will require support of a massive density of IoT devices and hence, justify the mMTC scenario.

However, one must also consider the drawbacks of making every application controlled centrally by a network operator:

1. Making all information exchange centrally controlled takes away the plug-and-play flexibility that exists with personal devices.
2. It also increases the security and privacy issues since the range of transmitted information and the knowledge of the existence of the device is much wider
3. It is not efficient to have every connection say even that between a television and its remote controller, centrally coordinated.
4. The above may also increase the cost of operation of such devices

Hence, any mMTC technology should combine the flexibility of plug-and-play operation with a capability to integrate into a wider system where interference and resources can be coordinated. In addition, the technology must provide schemes for efficient scheduling of these resources with sufficient granularity and reduced impact on a device's power consumption.

Low power operation of devices also depends on design factors like the maximum transmit power, antenna configuration, bandwidth processing capacity, data rate capacity, transmit/receive duplex operation, etc. These factors cannot be viewed standalone per device but concern the design of the complete technology and impact its efficiency. For example, device cost can be reduced by reducing its maximum power. It will proportionally limit its data rate or/and range. Data rate reduction may not be a potent factor in case of IoT devices. Range reduction will require denser deployment of base stations or eNodeBs or access points and cause the cost of service to increase. In the licensed spectrum, range reduction can be averted to an extent using smaller subchannelization permitting the device to concentrate the same transmit power on a smaller bandwidth and thereby, increasing its power spectral density. In the unlicensed spectrum, maximum power spectral density is limited by regulations. However, smaller subchannelization impacts other factors like susceptibility to frequency errors and scheduling complexity and overhead. All these factors are therefore, interrelated.

To summarize, following are the features necessary for an mMTC technology:

1. Small frequency/time granularity of resources
2. Low control channel decoding overhead
3. Maximization of autonomous transmission
4. Centralized coordination of resources to the extent of cost and overhead justification
5. Low delay from grant to retransmission
6. Options for low cost hardware like single transmit/receive chain, half duplex operation, etc
7. Sufficient range in relation to cost of base station/eNodeB/access point and carrier frequency
8. Mobility is not a concern