Managing storage growth is a constant challenge. As the capacity of PC hard drives has increased, users expect to store and keep more and more data. Unfortunately, high-reliability enterprise storage hardware is still relatively expensive - 146GB hard drives for our storage array are about $1,100 each and they have to be added in pairs. Shelves for the additional drives have to be added as the numbers increase, and there are additional licensing costs for the storage management software. In addition, we have to be able to back it up.
It is expensive because of its high performance, reliability and manageability. These drives are expected to sustain constant activity and have a "Mean Time Between Failure" of 1.5 million hours of operation. Your average consumer hard drive is only expected to last 500,000 hours at 20% of that activity level.
This is an industry-wide problem - users want a lot of easily available, networked storage, but the storage costs are daunting. The idea of "tiered storage" is making a bit of a comeback. The idea is that older data is moved off over time to less-expensive, lower-performing storage hardware. It is still managed storage, integrated with existing storage hardware or management tools, but the total cost is lower.
Overall, we have seen our storage usage on the network double at an interval of about two years, and that rate may be increasing. As we replace or upgrade storage hardware in the future, we may have to balance spending to increase capacity with spending to increase fault tolerance.