Cover Traffic

Cover traffic is the practice of sending dummy messages that carry no real payload, purely to obscure the timing and volume of genuine communication. It is a foundational defense against traffic analysis: even if an adversary cannot read encrypted content, the mere pattern of when and how much a person transmits can leak who they talk to, when they sleep, and when something important is happening. Cover traffic blurs that pattern by making real and fake messages indistinguishable on the wire — the observer sees a steady hum instead of a story.

Why Encryption Alone Is Not Enough

Encryption hides content, but it does not hide that a message was sent. An observer watching a connection still sees bursts of activity, packet sizes, and timing, and correlating bursts across two endpoints can reveal a conversation with high confidence — no decryption required. This is the core weakness of low-latency systems: even routing through relays or an onion service leaves timing correlations intact for an adversary who can watch both ends. By injecting a steady stream of dummy packets, a system ensures the observed traffic rate stays roughly constant whether or not the user is actively communicating. The dummies are discarded at their destination and are bitwise indistinguishable from real packets in transit, so neither a relay nor a wiretap can tell which is which.

Loops and Unobservability

Advanced anonymity systems such as Loopix and Nym use "loop" cover traffic, where a node or client sends dummy messages that travel through the network and return to the sender. Loops generated by mix nodes guarantee a baseline volume of traffic flowing through the system at all times, so real messages always have a crowd to hide in. Loops generated by end users go further: because the user's transmission rate never changes, an observer cannot even tell whether that user is communicating — the property researchers call unobservability, which is strictly stronger than anonymity. Loops also serve a second duty as active measurement: a node whose own loop messages stop coming back has detected that someone downstream is dropping traffic.

The Cost Is the Point

Cover traffic deliberately trades efficiency for privacy. Constant transmission consumes bandwidth and battery even when nothing is being said, which is why mass-market tools mostly omit it and why systems that include it meter it carefully. The trade-off surfaces anywhere bandwidth is scarce: on a LoRa mesh like Meshtastic, where airtime is limited by duty-cycle rules, meaningful cover traffic is largely unaffordable — a reminder that privacy properties must be engineered within a link's real budget, not assumed. On broadband, by contrast, a few kilobits per second of constant cover is a trivial price for hiding your communication pattern, and the main obstacles are engineering and adoption rather than cost.

Why Sovereign Users Should Care

Bitcoiners already understand that metadata is the attack surface: chain surveillance works by pattern analysis, not by breaking cryptography, and network surveillance works the same way. A node that broadcasts transactions only when its owner spends, from a static IP, is leaking a timing signature no encryption fixes. Cover traffic is the general answer to that class of leak — flatten the pattern until there is nothing to correlate. It is a core mechanism inside a mixnet, complements sender-hiding designs like Sealed Sender, and supports the broader objective of metadata-resistant messaging: systems that protect not just what you say, but the fact and rhythm of your saying it.

When evaluating any privacy system, ask where its cover traffic comes from and who pays for it. A design that promises unlinkability without spending bandwidth on dummies is almost certainly leaning on latency or batching tricks instead — legitimate tools, but weaker against a patient observer. There is no free lunch in traffic analysis; there is only choosing, explicitly, which resource you burn to buy silence.