INS is the saponification value minus the iodine number. Iodine number is a measure of the unsaturated bonds in the fats. In other words, more liquid oils -> more unsaturation -> higher iodine number OR lower INS number. The more unsaturation, the greater potential for a softer soap and greater chance for DOS. Obviously, a high iodine number doesn't absolutely mean the soap will be soft or get DOS; it's just an indicator.
Both iodine number and INS are ways to evaluate a fat mixture for ease of saponification, initial hardness for ease of unmolding, and the shelf life of the soap. They are values originally created for quickly evaluating blends of fats for use in large scale commercial soapmaking. They are not quite so critical for the handcrafted soapmaker who is less concerned about churning out mass quantities of soap.
Rather than focus on the iodine or INS numbers, some people keep the sum of linoleic and linolenic acid under 15% to help guard against DOS and softness. Or they have a rule of thumb about limiting the polyunsaturated oils such as canola, corn, soy, sunflower, safflower, etc. to a modestly low % in their recipes. Some watch the iodine number or INS value instead.
See also: http://www.soapmakingforum.com/showthread.php?p=320350
ETA: I made five batches of soap this weekend. The INS ranged from 137 to 155. Linoleic + linolenic amounts ranged from 4% to 6%. All were firm enough to unmold in 12 hours, and they will be fine to cut today. The little sample bars I poured from one batch didn't go through gel, so they're somewhat softer than the main loaf made from the same batch, but I still was able to ummold these little bars just fine.