The DNS root is the highest level in the DNS chain. Let's say a user wants to visit www.example.com.au. When they type it into their browser and hit enter, it sends a question - where is www.example.com.au? - to their corporate or ISP's resolver. If this is a new question, the resolver will start asking a known list of other DNS servers if they know where the www.example.com.au IP address is. The resolvers' job is to scour the Internet searching for the answer to the user's question. Each time a resolver finds a valid answer it stores it in memory for a defined period, making the search for the answer next time it's asked much quicker.
Domain names are organised in a hierarchical structure and searched starting from right to left.
The top of the domain name structure is the root (.) zone. The resolver will ask a root server, "where is www.example.com.au?" These root servers only contain answers about the location of the next level down in the structure, in this instance '.au'. When asked, they respond with "I don't know where www.example.com.au is, but I do know where '.au' is and here is the answer." The process repeats until the example.com.au DNS servers return the IP of the www record.
This process makes the root servers very important. Without them, there is no place for a resolver to start looking for resources on the Internet. The Internet is too vast for one server to know or to store all the answers, much like one phone book listing every single phone number in the world. Servers rely heavily on referrals for more information.