Differing scrambling parameter generation methods in Secure Remote Protocol

Question

RFC 2945 (The SRP Authentication and Key Exchange System) states,

The parameter u is a 32-bit unsigned integer which takes its value from the first 32 bits of the SHA1 hash of B, MSB first.

whereas the original publication describing SRP states that u is random.

What benefit is gained from using bits from a hash of B and restricting the size, rather than straight pseudorandom data?

Answer 1

The original article also states that: "For example, both sides can compute u as a simple function of B", which is what occurs in RFC 2945. u could be generated randomly and also transmitted, but deriving u from B (which is already transmitted) is harmless (security wise) and saves a few bytes (which is always good).