Loading docs/source/faq.rst +25 −42 Original line number Diff line number Diff line Loading @@ -7,11 +7,9 @@ Frequently Asked Questions (FAQ) .. note:: You may also enjoy the :doc:`glossary`. How many hops are the circuits used to perform the measurements? ------------------------------------------------------------------ ---------------------------------------------------------------- .. todo:: see :ref:`XX` Two hops: the relay to be measured and the :term:`helper relay`. Two hops. How are relays selected to be measured? --------------------------------------- Loading @@ -22,54 +20,39 @@ recent results. In this way, relays that have just joined the network or have just come back online after a many-day period of being offline will be measured before relays that have been online constantly. Is it the sbws scanner or server that gives the v3bw file to the directory authority? ------------------------------------------------------------------------------------- Technically, neither. In the suggested setup, the machine running :term:`sbws scanner` continuously will also periodically run :term:`sbws generate` to produce a :term:`v3bw file` for the :term:`bandwidth authority` to read. .. todo:: see :ref:`XX` How do sbws scanner results end up in the consensus? ---------------------------------------------------- The :term:`sbws scanner` runs continuously to gather fresh data. Is it the sbws scanner or server that I need to run close to a fast relay? ------------------------------------------------------------------------- The :term `sbws generate` command takes the fresh data and generates a :term:`v3bw file`. The :term:`sbws server`. The Tor :term:`directory authority` parses the v3bw file and includes bandwidth information in its vote. Why doesn't sbws just use a web/file server instead of custom software? ----------------------------------------------------------------------- The authorities take the low-median of the bandwidths for each relay from all of the :term:`bandwidth authorities <bandwidth authority>` and use that in the consensus. To lower protocol overhead and to allow :term:`sbws scanners <sbws scanner>` to request a wide range of bytes. Does sbws need any open ports? ------------------------------ Sbws essentially has no overhead, with only about 70 bytes used in a handshake at the beginning of each connection. A connection can be used to perform multiple measurements of one relay. No. At the time of writing, sbws scanners are allowed to request from the :term:`server <sbws server>` between 1 byte and 1,073,741,824 bytes (1 GiB). That's a lot of possibilities and a ton of storage space. How much bandwidth will the sbws scanner use? --------------------------------------------- .. todo:: .. todo:: answer this Look more into the viability of using an HTTP(S) server using HTTP basic authentication and range requests to control how many bytes to download. Determine if many requests (and therefore measurements) can be made over a single stream. Determine how much overhead HTTP adds and make a judgement call on whether it is worth it. How much bandwidth will the webserver use? ------------------------------------------ Why is there authentication between sbws scanner and sbws server? ----------------------------------------------------------------- .. todo:: answer this So random people on the Internet cannot discover an :term:`sbws server` and ask it to repeatedly send large amounts of data or otherwise abuse it. Should I run my own webserver? Use a CDN? Something else? --------------------------------------------------------- What ports does sbws use by default? ------------------------------------ It's up to you. Sbws is very flexible. - **31648/tcp**: The :term:`sbws server` listens on this port by default. It does not need to be reachable from the Internet, but only from the :term:`helper relay` near it. .. todo:: better answer. Loading
docs/source/faq.rst +25 −42 Original line number Diff line number Diff line Loading @@ -7,11 +7,9 @@ Frequently Asked Questions (FAQ) .. note:: You may also enjoy the :doc:`glossary`. How many hops are the circuits used to perform the measurements? ------------------------------------------------------------------ ---------------------------------------------------------------- .. todo:: see :ref:`XX` Two hops: the relay to be measured and the :term:`helper relay`. Two hops. How are relays selected to be measured? --------------------------------------- Loading @@ -22,54 +20,39 @@ recent results. In this way, relays that have just joined the network or have just come back online after a many-day period of being offline will be measured before relays that have been online constantly. Is it the sbws scanner or server that gives the v3bw file to the directory authority? ------------------------------------------------------------------------------------- Technically, neither. In the suggested setup, the machine running :term:`sbws scanner` continuously will also periodically run :term:`sbws generate` to produce a :term:`v3bw file` for the :term:`bandwidth authority` to read. .. todo:: see :ref:`XX` How do sbws scanner results end up in the consensus? ---------------------------------------------------- The :term:`sbws scanner` runs continuously to gather fresh data. Is it the sbws scanner or server that I need to run close to a fast relay? ------------------------------------------------------------------------- The :term `sbws generate` command takes the fresh data and generates a :term:`v3bw file`. The :term:`sbws server`. The Tor :term:`directory authority` parses the v3bw file and includes bandwidth information in its vote. Why doesn't sbws just use a web/file server instead of custom software? ----------------------------------------------------------------------- The authorities take the low-median of the bandwidths for each relay from all of the :term:`bandwidth authorities <bandwidth authority>` and use that in the consensus. To lower protocol overhead and to allow :term:`sbws scanners <sbws scanner>` to request a wide range of bytes. Does sbws need any open ports? ------------------------------ Sbws essentially has no overhead, with only about 70 bytes used in a handshake at the beginning of each connection. A connection can be used to perform multiple measurements of one relay. No. At the time of writing, sbws scanners are allowed to request from the :term:`server <sbws server>` between 1 byte and 1,073,741,824 bytes (1 GiB). That's a lot of possibilities and a ton of storage space. How much bandwidth will the sbws scanner use? --------------------------------------------- .. todo:: .. todo:: answer this Look more into the viability of using an HTTP(S) server using HTTP basic authentication and range requests to control how many bytes to download. Determine if many requests (and therefore measurements) can be made over a single stream. Determine how much overhead HTTP adds and make a judgement call on whether it is worth it. How much bandwidth will the webserver use? ------------------------------------------ Why is there authentication between sbws scanner and sbws server? ----------------------------------------------------------------- .. todo:: answer this So random people on the Internet cannot discover an :term:`sbws server` and ask it to repeatedly send large amounts of data or otherwise abuse it. Should I run my own webserver? Use a CDN? Something else? --------------------------------------------------------- What ports does sbws use by default? ------------------------------------ It's up to you. Sbws is very flexible. - **31648/tcp**: The :term:`sbws server` listens on this port by default. It does not need to be reachable from the Internet, but only from the :term:`helper relay` near it. .. todo:: better answer.
