According to many experts; memristors have the potential to change and drastically improve data storage and processing power of computers. But there is one small thing that doesn't fully make sense to me. Yes you could store data on a memristor, but how would you read it? The only way I could imaging reading the data on a memristor is by passing a current through it and measuring that. But that would change the memristors resistance and thus corrupt your data. So how would we read the data encoded on a memristor without corrupting its data in the same process?
I guess the point is that at least ion-diffusion based memristors change their state depending on the total charge passed through it. Thus one can use large current to write and low current to read.
Additionally one can imagine that ions move by hopping so there is a threshold voltage required to write. Employment of sub-threshold voltages will allow one to read the state.