There is such a principle — "Is not invented here." If I am not mistaken, met him in the articles of Joel Spolsky.
The point is simple — if there are key technology and you plan to use it as a basis for its decision, you should carefully consider which is better — to use or to write your bike. Because the farther, the greater the probability that the finished solution will not hold and slow down your further development. Of course this is only in the case when you think that you can write at least as well as existing solution, and even better.
Google writes its own file system, designs the hardware for themselves. I'm not saying that mail like Google.
I say that the larger the business the more often we can apply this principle.