To reach that conclusion, we must first start by defining what High Availability is. Ultimately, HA is the level of availability that an organization requires from a particular application (or set of applications). If you only need your systems five minutes a year, and the system is up for those five minutes (and no other), you have high availability. On the other hand, if, when your systems are down, airplanes fall out of the sky, no downtime at all is acceptable. Most likely, your critical systems fall someplace in between those two extremes.
You may decide that the cost of a full-blown DR solution is too high for your organization, and you'd be better served not implementing one at all. Ultimately, HA is a business decision, where you balance the cost of downtime against the cost of implementing the solution.
Hope this is helpful.
This was first published in June 2004