Hi,
as Uwe already mentioned, you won't get RTD without CCH, as the feedback via CCH is essential for training the model.
But in regard to WAD I am currently a bit confused about the redbook's statement you posted.
Codeversion 8.6.0 has introduced additional entropy performance stats on a per volume basis (and hence independently from FCM4 Drives); Those statistics were already used in combination with compression stats to present appropriate warnings in SI - which as far as I remember, was initially referred as basic "Workload Anomaly Detection" before the advanced RTD with FCM4 has been introduced.
Since those entropy stats come along with the normal XML performance stats which are collected with the Data Collector, I would have expected that at least that level of information could be available in SI.
/edit: I just found some documentation that mentioned that with the introduction of FCM4 the whole WAD computation seems to be shifted to the FCM4 layer - so that's probably why the "classic" WAD on xml performance stats basis will not longer come into play. So my above statement may not apply in that case.