It is reasonable to see temperature spikes that go to 90c, but not anything sustained in that range -- and it certainly shouldn't ever get to your processor's Tjmax (Maximum Junction) temperature.
You need to make sure that the BIOS' fan speed control programming is set to have the Processor Fan running at 100% duty cycle (i.e. full speed) if the temperature of the processor is above its Tcontrol temperature. This is usually in the low-80's, but can vary from one individual processor to another. My suggestion is to set the fan speed control coefficients for the Processor Fan as follows:
Minimum Temperature: Tcontrol - 20
Minimum Duty Cycle: 20%
Duty Cycle Increment: 4%
Temperature Smoothing: 10 seconds [Or higher if a lot of temperature spikes are occurring and this results in fan oscillation.]
All-On Temperature: Tcontrol + 5 [This is supposed to ensure that, if the temperature reaches this level, all of the chassis fans will start running at 100% as well.
Unfortunately, few fan speed control hardware solutions implement this feature properly.]
To find out what your processor's actual Tcontrol temperature is, you need a tool that can read the processor's Model-Specific Registers (MSRs). RWEverything (Read-Write Everything) is one such (free!) tool and what I typically use. You need the value of MSR 01A2, the IA32 Temperature Target MSR. Bits 16-23 provide your processor's Tjmax temperature and Bits 8-15 provide the Tcontrol Offset. Tcontrol is calculated by subtracting the Tcontrol Offset from Tjmax.
Ok, all that said, high temperatures like this is something that causes problems long-term. It is usually not the cause for things like Blue Screens. You need to be looking at other sources for this. There have been a couple of cases over the past couple of months where the outcome was that the motherboard's (BIOS') default programming for the processor power supply provided inadequate voltage levels and this played havoc with the processor's operation. In this case, you need to adjust the settings in the BIOS to ensure that yours is supplying sufficient voltage to the processor. In another case, despite passing tests with MemTest86, it was found that a bad DIMM was present and, once replaced, the Blue Screens stopped happening. Finally, there have been reports where the users were concluding that the failure was a bug in earlier versions of the Intel graphics driver. If you are using an external graphics card, this won't be the case for you - but, if you are, make sure you are running on the latest available driver from Intel.
Hope this helps,