IoT workloads use numerous sensors to gather and process massive amounts of data to make meaningful and economical business decisions in real time -- and the cloud is typically the hub for those activities.
Public clouds are often the foundation for IoT projects because they can handle the enormous bandwidth, compute and storage scalability demands. However, there is no guarantee of ease or success. Organizations can face many obstacles when they move, store and secure data produced by expansive IoT deployments. Take a closer look at three of the biggest IoT challenges in cloud computing.
Organizations must move data from each IoT device across a local area network (LAN), onto the internet and then to a cloud storage instance. For optimal results, the network needs to provide enough bandwidth to accommodate the flow of real-time data from the IoT devices in service. In some cases, the local network and internet connection must be upgraded to accommodate the total bandwidth requirements of IoT deployments.
Evaluate your network resiliency requirements and factor in the potential for LAN failures or an internet service provider (ISP) outage. In some cases, you can ease the risks of internet outages or congestion if you distribute the load across multiple network links from multiple ISPs. Alternatively, look into direct connection services from cloud providers, such as AWS Direct Connect and Microsoft Azure ExpressRoute, to establish a dedicated link between user locations and the cloud provider's facility.
Enterprises also should deploy data storage and compute resources in the cloud region closest to their sensors. A shorter geographic distance means fewer network hops, lower latency and less chance for disruption.
The public cloud is well-suited for IoT deployments because storage is readily available and highly scalable. However, cloud storage costs rise with increases in capacity and traffic.
IoT data is real time and temporary and it requires little to no long-term storage retention. With aggressive data retention settings, you can keep data just long enough to make decisions and then delete it to make way for more data. This minimizes capacity requirements and associated costs over time. If you need to keep IoT data longer, consider a less expensive, archival storage tier, such as Amazon S3 Glacier.
While it typically costs nothing to put data into cloud storage, it can be expensive to move it out later. It's usually good policy to put compute and storage resources in the same cloud to process the stored IoT data, rather than process it elsewhere and deal with egress fees. The only storage that might be subject to departure costs is the result of IoT compute activities, which is typically minimal.
IoT is still in its formative years and has yet to emphasize security as a design feature or selling point. IoT vulnerabilities are traced to the weaknesses inherent in wireless data communication, along with the limited compute power in each device. This leaves IoT devices notoriously insecure, so enterprises are responsible for properly securing IoT deployments. Core IoT challenges with security include:
- IoT configuration management: Enterprises need to configure each IoT device for optimum security settings. It's critical to set up each device properly when it is first deployed, then maintain that configuration, prevent any unauthorized changes and report any change attempts to administrators.
- Authentication and authorization: Each IoT device should require clear credentials to access the device for configuration, reading and writing. This ensures that any access to an IoT device takes place from a trusted point or application.
- Encryption: Although you can encrypt data once it's stored in the cloud, IoT data is vulnerable while in transit. Most IoT devices don't encrypt data in transit, but expect more offerings to make this a fundamental capability going forward.
IoT security cannot stop with individual devices. Security must extend throughout the IoT chain to include web interfaces and portals, mobile devices, and cloud services used to configure and manage IoT devices.
Edge computing considerations
Businesses can use edge computing to locally collect, store and preprocess data before it is sent to the cloud. This gives an organization more control over the raw data and how it's protected because it limits what's transmitted. It also provides long-term cost savings if huge volumes are retained for extended periods.
Edge computing deployments also strengthen IoT security. It implements a perimeter between the IoT deployment, LAN and the WAN. Unencrypted IoT device data can be collected locally and encrypted before being sent over the WAN to the cloud.
Cloud IoT services
Public cloud providers provide a range of services designed to offer support and capabilities to increase adoption and address myriad IoT challenges.
AWS offers the most diverse and comprehensive IoT service menu:
- AWS IoT Core supports connectivity and security between IoT devices and cloud applications.
- AWS IoT Device Management focuses on IoT device management so organizations can set up, monitor and manage huge arrays of IoT devices.
- AWS IoT Device Defender monitors and audits IoT deployments to enforce security best practices.
Azure also offers a variety of IoT services:
- Azure IoT Hub connects, monitors and manages huge IoT device deployments.
- Azure IoT Edge promises to bolster edge computing by supporting cloudlike intelligence and analytics in edge devices.
- Azure IoT Central seeks to simplify IoT deployment, and promises simplified connectivity, monitoring and management to IoT devices.
With an emphasis on AI and machine learning, Google currently provides the Cloud IoT Core service to help users connect, manage and exchange data from huge IoT deployments. Google is also developing additional IoT services, such as Cloud IoT Edge and Edge TPU.