Add the random trace id flag

spec/20-http_request_header_format.md

@@ -138,8 +138,6 @@ boolean sampled = (traceFlags & FLAG_SAMPLED) == FLAG_SAMPLED;
 
 ##### Sampled flag
 
-The current version of this specification (`00`) only supports a single flag called `sampled`.
-
 When set, the least significant bit (right-most), denotes that the caller may have recorded trace data. When unset, the caller did not record trace data out-of-band.
 
 There are a number of recording scenarios that may break distributed tracing:
@@ -176,6 +174,28 @@ There are two additional options that vendors MAY follow:
 - A component that makes a deferred or delayed recording decision may communicate the priority of a recording by setting `sampled` flag to `1` for a subset of requests.
 - A component may also fall back to probability sampling and set the `sampled` flag to `1` for the subset of requests.
 
+##### Random Trace ID Flag
+
+<!--
+
+TODO: how many random bytes are needed?
+7 was chosen as it can be efficiently represented as a 64-bit signed or unsigned integer.
+8 would require an unsigned long which is not supported by some languages (like Java and Go).
+63 bits would be possible, but would require a more complex description that may be more difficult to understand.
+
+TODO: Which specific bytes should be random?
+The right-most bytes were chosen because some tracing systems are known to use the left-most
+portion of the trace id for non-random data such as a timestamp component.
+
+TODO: Do we want to place any restrictions on the randomness or is saying "MUST be random" enough?
+As an example, RFC4122 (UUID v4) simply states "Set all the other bits to randomly (or pseudo-randomly) chosen values."
+https://datatracker.ietf.org/doc/html/rfc4122#section-4.4
+
+-->
+
+When set, the second least significant bit (second from the right), denotes that the right-most 7 bytes of the trace ID MUST be random (or pseudo-random).
+When unset, the trace ID may be generated in any way that satisfies the requirements of the [trace ID format](#trace-id).
+
 ##### Other Flags
 
 The behavior of other flags, such as (`00000100`) is not defined and is reserved for future use. Vendors MUST set those to zero.

spec/60-trace-id-format.md

@@ -22,6 +22,9 @@ concerns](#privacy-considerations) of exposing unwanted information. Randomness
 also allows tracing vendors to base sampling decisions on `trace-id` field value
 and avoid propagating an additional sampling context.
 
+If the `random-trace-id` flag is set, the least significant 7 bytes of the
+`trace-id` MUST be randomly (or pseudo-randomly) generated.
+
 As shown in the next section, it is important for `trace-id` to carry
 "uniqueness" and "randomness" in the right part of the `trace-id`, for better
 inter-operability with some existing systems.
@@ -50,7 +53,9 @@ systems with these existing systems, the following practices are recommended:
    compliant 16 bytes `trace-id` from a shorter identifier, it SHOULD left pad
    the original identifier with zeroes. For example, the identifier
    `53ce929d0e0e4736`, SHOULD be converted to `trace-id` value
-   `000000000000000053ce929d0e0e4736`.
+   `000000000000000053ce929d0e0e4736`. If the resultant `trace-id` value does
+   not satisfy the constraints of the `random-trace-id` flag, the flag MUST
+   be set to `0`.
 2. When a system receives an inbound message and needs to convert the 16 bytes
    `trace-id` to a shorter identifier, the rightmost part of `trace-id` SHOULD
    be used as this identifier. For instance, if the value of `trace-id` was