Ok, firstly I'll try and explain Raid 1 and Raid 10.
In terms of disk resilience, the most resiliant is a mirror, in its simplest case the contents of 1 disk mirrored to the other (2 disks), if 1 disk fails the other has all the data and can operate as normal. (all automatically). You can only use 50% of your total disk space. 2 x 500GB disk, on 500GB usable.
In terms of disk performance, the fastest is Raid 0,when the data is written accross all the disk available, which means that if you have 10 disk, the write is written to 10 disks instead of 1, been 10 times as fast (in theory), although, if you loose 1 of these disk, you loose the lot (no Resilence).
The midway between this (although not the cheapest), is raid 1+0 or raid 10, you take example 5 disk, and set them as raid 0, fastest write performance, and the automatically mirror the contents to the other 5 disk, providing resilency, just mean you actually only get half the storage of the 10 disk, but have resilence and performance.
Theres loads of articals online if you want to read further on RAID configurations.
As for the server specification, yeah its better than you current server, although your current servers not bad, and it appears your not having performance issues, but sometimes availabilty issues.
A new server will only leave you with the same issues, as a single server is a single point of failure, ok, the hardware may be more reliable, but if that server fails you have no access to the system, there are work arounds for this, but there not cheap.
Ask your CRM provider how to scale the system for resilency in the even of a single server failure, I can garentee you it won't be cheap, but it depends on how-much the downtime actually costs you.
Generally, these kind of systems scale with multiple front end servers (web access, and application servers), and a shared database cluster back-end. totaling a minimum of 4 servers (told you it would be more expensive). This would mean that you could lose 1 of the database server, or one of the web servers and the system will still be available.
There are other methods of clustering these services now, a link of which I've provided below. But resilience isn't cheap, but you need to calculate the loss of business caused by downtime, balanced with the system costs.
http://www.stratustechnologies.com/products/ftserver/index.htm
Cheers
