Urgent.News

What's breaking now, across thousands of outlets.

Tech

I put every smart device on its own VLAN, and found out which ones were phoning home constantly

A flat network has vulnerabilities.

I put every smart device on its own VLAN, and found out which ones were phoning home constantly

In today's world, many smart devices like wireless plugs, IP cameras, and robotic vacuums rely on continuous internet access for vendor cloud connectivity. However, these devices can generate unsolicited telemetry, DNS lookups, and unencrypted data streams. When all devices share the same network, vulnerabilities arise. Network segmentation, implemented with VLANs and strict firewall rules, can isolate untrusted hardware and monitor outbound traffic.

DNS query logging and sinkhole integration can help identify persistent phoning-home behavior. Some devices hardcode DNS servers or use DNS over HTTPS, potentially missing local cache solutions like Pi-hole or AdGuard Home. Devices may ignore DHCP-advertised NTP servers and contact external time servers, which is usually essential but worth noting.

Devices stream device state, uptime metrics, and local environment diagnostics, creating multiple connection attempts. Network segmentation relies on managed switches and routers configured with 802.1Q VLAN trunking to separate physical switch ports and SSIDs into distinct subnets. Within each VLAN, dedicated DHCP scopes ensure that broadcast traffic doesn't leak across boundaries.

Firewall rules between VLANs should be explicitly denied, blocking IoT subnet traffic from initiating TCP or UDP sessions with trusted networks. Enforcing inter-VLAN firewall rules blocks unauthorized WAN access that could prevent firmware checks. By revoking internet access for specific IoT devices while retaining local subnet communication, you can protect your network from unauthorized port-scanning attempts, external firmware-check queries, and unrequested telemetry payloads.

Written by urgent.news from XDA Developers's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at xda-developers.com →

More in Tech

Why standard RAG failed on stack traces until using Hindsight

If you pipe raw production stack traces into a standard vector database, you quickly discover that cosine similarity has no idea what an outage actually means.

  • Standard RAG systems failed with stack traces due to semantic similarity issues.
  • Boilerplate noise and context window limits hindered accurate incident identification.
  • New architecture with deterministic parsing, semantic memory, and LLM reasoning solved problems.

VeloCloud Shows Why Integration Endpoint Trust Needs an Owner

Integration endpoint trust is the part of an orchestrator's exposure that an inbound-reachability review can miss. On November 2, 2026, Arista published Security Advisory 0202 for VeloCloud…

  • Arista published security advisory on VeloCloud Orchestrator vulnerability.
  • Memory-corruption flaw in Brotli decoder exploitable via crafted responses.
  • Upgrading to fixed version recommended to mitigate exploitation risk.

Our scraper stored 20 links per page and kept every page in memory

We build and maintain web scrapers that run for hours at a time. In October one of them ran out of memory, and the cause was a detail of lxml that we had skimmed past in its docs.

  • Scraper stored 20 links per page, each 12.6 MB
  • lxml documentation oversight caused memory leak
  • Changing xpath() to return ordinary strings resolved issue

More from Sunday 8 November →