diff --git a/benchmarks.md b/benchmarks.md
new file mode 100644
index 0000000..3816026
--- /dev/null
+++ b/benchmarks.md
@@ -0,0 +1,37 @@
+## DNS Benchmarks
+
+Collection of hints and links which can be useful for people new to DNS benchmarking.
+
+
+
+### Step 0. Prerequisites
+See DNS-OARC 42, DNS Benchmarking 101: Essentials and Common Pitfalls
+- [slides](https://indico.dns-oarc.net/event/48/contributions/1033/attachments/991/1943/pspacek.pdf)
+- [video recording](https://www.youtube.com/watch?v=ZrMLeyu4gTU&list=PLCAxS3rufJ1fHh8oe9UQ0IwlVfHq30_hh&index=16)
+
+- Test design must be different for any combination of
+  - resolver
+  - authoritative server
+  - normal traffic
+  - DoS traffic
+  - server management operations
+
+- :warning: Be absolutely sure to **test your test environment first**.
+  - If you don't the results are probably garbage
+  - An echo server will be very useful
+    - user-space: https://github.com/DNS-OARC/dumdumd/
+    - XDP (UDP-only): https://gitlab.nic.cz/knot/xdp-utils/
+
+- Usual tuning tips apply:
+  - beware of [NUMA domains](https://en.wikipedia.org/wiki/Non-uniform_memory_access) in hardware - you _might_ want to restrict yourself to an _inside_ of one NUMA domain - commands [numactl](https://man.archlinux.org/man/numactl.8), [taskset](https://man.archlinux.org/man/taskset.1)
+  - pick the NUMA domain so it is directly connected to network card in use - tool [lstopo](https://man.archlinux.org/man/lstopo-no-graphics.1.en) can help with that
+  - find out how many network card queues is available - command [ethtool -l devicename](https://man.archlinux.org/man/ethtool.8)
+  - pick minimum value from (# of CPUS in the chosen NUMA domain, number of network queues) and use that for number of threads
+  - these are just a starting point, experiment!
+
+### Resolvers
+- An example with explanation how and why we test: https://www.isc.org/blogs/bind-resolver-performance-july-2021/
+- **:warning: Results cannot be generalized to other data sets** or setups with any certainty
+  - Why? Because every single query (and it's timing) changes _state_ of the system under test. Not to speak of dependency on authoritative server performance ...
+
+- Introduction to DNS Shotgun [slides](https://ripe79.ripe.net/wp-content/uploads/presentations/45-benchmarking.pdf), [video](https://ripe79.ripe.net/archives/video/198)
\ No newline at end of file