What are the main security and networking challenges specific to edge computing with constrained devices?

Protecting data at rest, in motion and in use, on hardware whose resource budget can restrict cryptography and transport choices — which is why lightweight primitives and UDP can be useful design choices.

The data-centred challenges first:

• Data storage — readings must be securely stored both on the device and in the cloud: encryption of data at rest.
• Data transmission — the link must be secured: encryption of data in motion.
• Data usage — data is fed into analytical processing and potentially fused with information from other sources to derive actionable insights. This is the sneaky one: the fusion can create sensitivity that no individual reading had.
• Privacy — preserving mechanisms are needed from the start, not retrofitted.

Then the constraint-driven ones:

• Security for resource-constrained edge devices — the umbrella problem.
• Lightweight cryptographic primitives, where the energy budget and computational power make conventional implementations too costly.
• Software and firmware updates, which must work in the field.
• Networking — UDP can reduce transport overhead; where TCP fits, it remains an option. Choose DTLS or TLS to match the transport.
• End-to-end security and privacy across all of the above.

Tip: the data-usage point is the one people skip. A stream of household electricity readings is dull; fused with a calendar it reveals when the house is empty. Ask not only "is this reading sensitive?" but "what does it become when combined?"

Go deeper:

• NIST — Lightweight Cryptography project — the concrete answer to the lightweight-primitives requirement: the Ascon family, selected in 2023 and standardised as SP 800-232.
• RFC 7925 — TLS and DTLS profiles for IoT — shows how established transport security and credential types are adapted to constrained devices.