Skip to main content

IoT Architecture: Designing Network Communications

Published Date
Author Cadence PCB Solutions

Key Takeaways

  • The three major keys to IoT architecture are edge sensors, gateways, and cloud storage.

  • IoT sometimes gets thrown around overzealously; it’s important to be able to recognize what is and isn’t IoT.

  • Depending on what aspects designers want to emphasize, there are a host of different IoT protocols available.

Cloud connecting to the various IoT devices.

Designers should expect the prevalence of IoT to expand in the coming years.

The world is becoming more connected, and devices are both driving and mirroring this development. With the untapped data available to companies and customers alike, IoT architecture is rapidly evolving to capture market inefficiencies and produce innovations. To best capture this technological zeitgeist, designers should be cognizant of the format of an IoT system as well as the advantages and disadvantages inherent to common network protocols.

The Three Pillars of IoT Architecture

IoT architecture communicates across three hubs: edge sensing, gateways, and data management/analysis. Depending on the specific industry application of IoT, certain parameters may require preferential treatment in the design. Large-scale data acquisition demands high storage and low power transmission costs for continuous transfer, while real-time systems need to prioritize high processing speeds and transfer rates.

The flow of data involves nodes (which indicate sensor/actuator systems), gateways, and the cloud. The gateway disseminates info from the cloud to the nodes (as indicated by the data packets) and collects data for outgoing storage and analysis in the cloud. Each level of design needs to be optimized within the scope of the greater system:

  • Edge sensing - Sensors form the perceptive layer of IoT integration. By probing internal and external processes, systems can become more attentive to minute fluctuations in processes that could represent inefficiencies or worse. Arguably the most important outcome of improved diagnostics is predictive maintenance or the ability to detect faults in systems at their earliest development when the cost of repair or replacement is at its minimum.
  • Gateways - Sensor data needs to talk with other interrelated systems within the organization’s network. Roughly, the system must handle three essential tasks: a hardware and software description (including read/write frames) of how the device is implemented within the system, addressing for packet delivery, and a structured system to handle data conflicts and overall network reliability. 
  • Data systems - Managing the data that flows into the system acts as the brains of the system, guiding adjustments in real-time for optimal efficiency. Queries, searches, and structured elements like links help sort the data and enable rapid and powerful mathematical analysis.

IoT has become a major point of emphasis in many fields, and it's important to recognize what is and does not apply to IoT design:

System

Description

Is it IoT?

Dissemination

One-way transfer of data (sensor to system, system to the cloud, etc.)

No

Periodic or continuous transfer and updates between sensor, cloud, and system 

Yes

Rapidly Actionable

Data is collected for future expansion or correction

No

Data can be acted upon in real-time (or close to it) to emphasize the performance of one or more key attributes

Yes

Connectivity

A local or distributed network that is isolated from outside objects or networks

No

A network that integrates with additional objects or networks for greater functionality and flexibility

Yes

Topics like big data or industry 4.0 are not necessarily IoT systems, though they may operate alongside or within IoT applications to unlock greater system performance.

Determining Factors for IoT Protocols

Equally important as the structure of data flow are the protocols used to communicate between different layers of an IoT system. There are tradeoffs between the different protocols, with the major balance concerning power consumption, speed, distance, and bitrate. Generally, an increase in any of the latter three characteristics is positively correlated with greater power draw, though particular implementations can minimize the increase. Different IoT protocols see usage in numerous industries depending on the system needs:

  • LPWAN - A low-power wide area network system is something of a holy grail for appropriate IoT applications, offering tremendous range and low-power consumption. For sensors and actuators that can subsist on low data fidelity and transfer rates, LPWAN is a near-perfect solution to wireless communications. 
  • Cellular (4G, 5G, etc.) - While IoT systems can use LPWAN protocols over cellular networks, more familiar in-place protocols may also be of use. These networks require far greater bandwidth than narrowband IoT architectures but can offer greater bitrate transmission as well.
  • NFC- Near-field communication can be utilized for distances up to four centimeters/1.5 inches. While this would be a poor choice for a long-distance setup, it shines as a security measure by reducing the distance sensitive information has to travel. Antennas can typically only detect transponder equipment signals in immediate proximity.
  • RFID - The superset of NFC, radio frequency identification uses a system of passive or active devices to communicate/update information; the active format has the two communicators generating their fields (overall more costly in power terms), while the passive method has the target device communicate by acting upon the field of the transmitter.
  • BLE - Bluetooth low energy operates within a relatively short range (< 100 meters), with communication flowing through some central nexus to the periphery system. In this structure, the nodes do not need to constantly transmit or receive status updates, which allows them to be placed in low-power or sleep modes until the next transmission. For systems that require infrequent updates, this is an excellent option.
  • Zigbee - For a slight reduction in bitrate, Zigbee can double the range of a BLE and does not require a star configuration to operate. However, power consumption increases as sensors need to operate continuously.
  • Wifi - The gold standard of high-speed, high-bitrate communication; it is rarely best suited for IoT due to its high power load and marginal range.

How Designers Can Rise to the Challenge of IoT

IoT architecture is as varied as the industries it’s responsible for interfacing with, and both software and hardware teams need to be well aware of the overall structure of the system as well as the pros and cons of specific models. Being a communication mode itself, it is perhaps unsurprising that design teams need to incorporate device and board needs to meet their project goals. The increasing rollout of IoT systems means more components at the board level and an emphasis on signal integrity and EMC within the larger electronic environment; layout designers must continue to hone placement and routing to optimize these subcircuits without sacrificing critical performance elsewhere.

Fortunately for IoT or any other prominent board feature, Cadence offers a wealth of PCB design and analysis software to expedite the design phase and catch errors prior to production runs. OrCAD PCB designer users can enjoy a comprehensive toolset that is able to handle the dense and complex layouts of today’s boards, with functionality that looks towards the future of PCB design coming down the pipeline.

Leading electronics providers rely on Cadence products to optimize power, space, and energy needs for a wide variety of market applications. To learn more about our innovative solutions, talk to our team of experts or subscribe to our YouTube channel.