[mfukar]

Both lock-free and wait-free algorithms/implementations offer the same advantage, as a corollary of their definition: latency guarantees. These are central to real-time systems.

PS. The difference is that lock-free algorithms improve throughput _at the expense of single-thread latency_, and thus are more useful for satisfying system-wide throughput requirements, e.g. for a DBMS. Wait-freedom additionally guarantees freedom from starvation.