The history of IPv6 begins in 1992. It was developed for solving worldwide ip address space problem and some other problems. It's decided that IPv6 address space distribution will be carry out by IANA (Internet Assigned Numbers Authority) and its regional representations that will cares for ip address delivering in their regions. This distribution won't be irreversible. IANA will always allow redistributing address space in case of mistake. In other words it's all done to not to assume old mistakes of IPv4 address space distribution.
Address space will be extended from old 4 billion+ of IPv4 to 340 282 366 920 938 463 463 374 607 431 768 211 456 addresses and it will be a 128-bit. It will add some routing features (especially for multicast translation). A new address type has appeared in IPv6 - "anycast". It will lead to the nearest interface form address list.
IPv6 addresses also are capable to context less autoconfiguration. Its will essentially simplify routing.
Beside address space benefit IPv6 also have following innovations and improvements:
IPv6 improvements will not only solve address space problem, it will improve all Internet structure and makes it more logical and technically thought out.
Autoconfiguration of ip addresses for final hosts will hold in two ways: using improved DHCP protocol and without it.
IPv6 allows packages to be fragmented and to be assembled by routers. Sender must at first find out maximum size of package (Maximum Transmission Unit, MTU) supporting on all way to destination and fragment it independently if necessary. This time it's stipulated that MTU must be larger than 576 bytes but probably in future this value will be extend to 1500 byte.
Removing fragmenting cares from routers also promotes increasing of their work's efficiency but may cause complicating of final system's work.
IPv6 authors also declare that it will highly increase network security. It will be impossible to carry out Dos attack and network port scanning.
Transition on IPv6 is inevitable anyway! But this time it goes slowly because of low evidence of its economic benefits. Mainly migration to IPv6 going in those countries where ip address space lack is felt.
It's not so easy to refuse IPv4 but it's worth its expenses! It's more reasonably to change Internet structure once than left its issues and just upgrade main functions...