The safety related response to all this is discussed briefly in the "piece by Jens Flottau in Aviation Week for August 10, 2009. "
A speed sensing malfunction for any reason was probably judged during initial certification as a minor malfunction due to its low expected rate of occurrence as well as the barriers (crew actions and availability of other instruments) available to make it unlikely that it would progress to a more serious outcome.
The article mentioned above indicates that the additional in-service experience reveals that the frequency of the malfunctions coupled together with the likelihood of significantly increasing pilot workload now creates a more hazardous outcome which needs to be addressed. Somewhat akin to the engine grapple (ice Crystals) problems discussed in the past when aircraft flew over storms.
The challenges to addressing problems like are also similar. If you go for a simple product change like a different manufactured probe the question is immediately how much is good enough and how do you know? Since the original product meets the certification standards.
Other means of addressing the safety risk are avoidance as well as means for reducing the pilot workload (simplified tasks by memory). Again the issue is a practical way for doing this (retraining all pilots takes time).
Right now it looks like a combination will be tried in a subjective manner to at least add more data to the knowledge base. I'm not sure if this gets mandated by EASA or not or even if the FAA will follow.