Many IT managers are now facing the challenge of adapting their existing enterprise campus and branch networks to the new demands of connecting IoT devices. The Internet of Things is designed take data from large numbers of distributed devices and aggregate it to provide actionable outcomes.
IoT systems are extremely complex and system design is further complicated by the lack of IoT network standards and sheer number of architectural and supplier options. Given the number of standards in the various stages of development, flexibility and adaptability will be critical.
The connectivity of things rather than people or PCs is an existing technology, generally referred to a machine-to-machine (M2M). Industrial controls are an example of this technology. However, for most IT managers, the buzz is around how new ways of connecting things/devices can impact their business – specifically by making processes more efficient, reducing costs, and improving customer service.
The challenge for IT managers is to adapt their existing network architecture at large campus and remote branch locations to the new demands of IoT. Depending on the type of device and information system, IoT demands will vary considerably, which will dictate the type of network(s) required.
Here are some factors IT managers should consider when planning for IoT connected networks:
- Amount of traffic per device/thing
- Aggregate amount of traffic (amount per device X number of devices)
- Are there significant spikes in traffic?
- Frequency of connection per device
- Range of connection per device
- Bandwidth costs per device
- Is the traffic primarily upstream (from device to controller) or more bi-directional?
- What is the lifespan of the device/sensor/thing?
- Will I need to update software at the endpoint?
- What are the security threats to the devices and the data transferred?
IoT connectivity options
There are a multitude of IoT connectivity options and each has specific pluses and minuses. The leading connection options that have near universal adoption are WiFi, Bluetooth, and cellular wireless (4G). But each of these have disadvantages: Wi-Fi uses a lot of power; Bluetooth has limited range; and 4G also is power intensive and can be expensive, especially for heavy data usage.
Network innovation is focused on addressing the challenges of connecting devices over long distances and accommodating low-power requirements, e.g., devices running on batteries. A number of newer IoT network options exist, including:
- Sigfox -- In 2009, Sigfox, a French startup, introduced a low-power, ultra- narrow band network running on unlicensed spectrum. Based on service model like wireless operators, Sigfox is only available in specific parts of Europe, major US cities, and selectively in Latin America.
- LoRA -- The LoRA Alliance is a collaboration of industry partners building a low-power protocol called LoRaWAN that runs on unlicensed spectrum.
- IoT cellular -- There are several standards based on the 3GPP specification, such as LTE-M, NB LTE-M, and NB-IOT. These standards are cellular networking modified and optimized for IoT devices.
In addition, there are other IoT network standards in process. Sound confusing? It is. There are too many competing options, making it difficult to bet on a long term winner.
IoT networks provide the infrastructure to deliver potentially massive amounts of data between the end devices and IoT control systems. Picking the right networks for the bandwidth, power, and cost requirements of your IoT system is challenging due the sheer number of network options. As we are very early in the evolution to IoT, IT managers should pilot small projects to get experience with IoT requirements. The DevOps principals of build fast, fail fast are applicable for IoT. Redundant security systems must be built in to the IoT architecture, not added on later. Picking the right partners with system integration expertise is critical.