Pyramid of Pain
David Bianco's six indicator types, from hash values to TTPs, ranked by what each one costs the adversary.
David Bianco's 2013 model says indicators are not worth the same. It stacks six types from hashes at the base, trivial to change, to TTPs at the top, expensive to change.
A blocked IP is swapped in minutes. A broken technique forces new tradecraft, which is the most time-consuming thing you can make an adversary do.
6 rows.
| Level | Gist | Why it hurts |
|---|---|---|
| Hash Values | A SHA-1/MD5/SHA-256 fingerprint of one specific file. | The most precise indicator there is, and the easiest to lose. A hash digests the entire file, so changing even a single bit (padding with one extra byte, flipping an unused flag) produces a completely unrelated hash value. An adversary doesn't even have to try to evade a hash match; it breaks automatically the moment they touch the file, which is why Bianco notes it may not even be worth tracking in isolation. |
| IP Addresses | A single address or netblock the adversary connects from or to. | Bianco calls this "the most fundamental indicator": almost any attack requires a network connection, and a connection requires an address. But it's also trivial to rotate: a capable adversary can change IPs whenever it suits them, and one routing through an anonymizing proxy may switch addresses constantly without even noticing. Deny them one IP and they usually recover without breaking stride. |
| Domain Names | A domain (or sub-domain) resolving to adversary infrastructure. | A small step up from IPs: a domain has to be registered, paid for (even with stolen funds), and hosted somewhere before it works, and a newly registered one can take a day or two to become visible everywhere. In practice that's a minor speed bump, not a real cost. Plenty of registrars apply little scrutiny to new registrations, so swapping domains stays cheap for anyone willing to burn a few of them. |
| Network & Host Artifacts | Observable side effects of the activity: on the wire or on a host. | This is, in Bianco's words, the level where you finally start to have some negative impact on the adversary. Catch a distinctive artifact (his example is an off-by-one-character User-Agent string their recon tool always sends), and block it, and you force them back to their own lab to figure out how you detected it and fix it. The fix itself might be trivial, but they still had to spend real time finding and patching the leak. |
| Tools | The actual software the adversary brings to do the job. | Burn a tool broadly enough, meaning its distinguishing traits and not just one sample, and you're not asking for a five-minute recompile. Bianco frames it as taking away one of the adversary's arrows entirely: they have to invest in research (find an existing alternative), development (build a replacement, if they even can), and training (learn to use it as effectively as the one you just cost them). |
| Tactics, Techniques & Procedures | How the adversary operates: behavior, not any one tool. | Detect at this level and you're operating directly on adversary behavior, not against a swappable artifact underneath it. In Bianco's own framing, respond to TTPs fast enough and across enough of an adversary's behaviors, and you force them to do the single most time-consuming thing possible: learn new behavior, or abandon their tradecraft and rebuild it from scratch. |