On Sun, 2009-10-18 at 10:55 +0100, G T Smith wrote:
i can see from the list of issues covered in the novell CLP 11 It is MUCH easier to configure than IPv4. Netmasks: Gone. Multicast: Just Works. Auto-Config: Just works. No addressing issues for point-to-point links. And it is faster for really big local transfers and consumes less router resources on the WAN. Why not use IPv6? It is just a better protocol. did you really get to a point where IPv4 simply couldn't do the job anymore? There is just no reason not to begin phasing out IPv4. QWERTY effect strikes again... The earliest I remember talk about IPv6 was about 15 years ago, and
Adam Tauno Williams wrote: mainly in the context of a solution to a shortage of address space with IPv4. When NAT became perceived as a basic solution to the latter problem interest waned,
NAT is not a solution. It just creates more problems. Look at the whole space of NAT traversal strategies. This is the problem with IPv4: it is long overdue for being taken out to the sheds, it is one hack on top of another (NAT, multicast, all the chicanery of point-to-point connections, tunneling). And people say "my IPv4 network is simple and works great." It isn't, and it doesn't. Pain you are accustomed to doesn't cease to be pain just because you are accustomed to it.
Many would be wary of a mixed IP protocol approach (adds something else to go wrong in unpredictable ways, especially in environments which already have a mixture of protocols). Most IT technical teams are under resourced for what non-technical management expect them to perform, and there is, at least among the better technical managers a general unwillingness to commit their people to support additional responsibilities without the required additional resources (which are usually not forthcoming).
None of which is an argument against IPv6. That is just an organizational issue. The above situation applies to many things both IT related and not.
Many admins adopt an 'if it aint broke dont fix it approach'. Therefore for something to be adopted it needs give a perceived significant benefit against any cost, (or an advocate which tells people to adopt X by date y or else ).
This is a failure of those familiar with IPv6. The argument has been about address space, which is a small portion of IPv6. IPv6: * reduces significantly router congestion (it can be routed much more efficiently, at higher rates given the same computation resources). * tunneling is a natural part of the protocol. Microsoft has even started requiring IPv6 in its new services like Direct-Access for supporting mobile workers. It is very sad that Microsoft is going to be way ahead of the curve on this and drive the superior IPv6 protocol while UNIX/LINUX communities continue to snub it and argue that there is no benefit. * is faster for local transfers (gone are MTU/MRU & fragmentation issues) * multi-cast is built into the protocol (much like tunneling) * addressing on Point-to-Point connections is a non-issue * in many cases devices can self-configure * mobility is supported for persistent connections for roving devices * no more crazy netmask nonsense. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org