What is Digg?49 Comments
- 350Zed, on 10/06/2008, -2/+181. network hardware costs
2. network management systems costs
3. staff training costs
4. network transition costs
There... 4 reasons.
This is a NETWORK problem, not an OS problem. - incubusbeatsall, on 10/06/2008, -1/+8This is pretty huge. I'm surprised that there aren't more comments on it.
We're going to have to switch to IPv6 eventually, and I'm really curious to what solutions they're going to come up with. This is something like the problem with analog/digital TV's. Millions of people are going to most likely have to get some sort of hardware upgrade. It is inevitable. - mooninite, on 10/06/2008, -0/+7We already have a standard. It's IPv6 tunneled through IPv4.
However, the majority of network infrastructure cannot handle IPv6. All cable companies are using Docsis 2.0 or earlier, which did not call for IPv6 support. Just recently 3.0 was announced with IPv6 support and a new "2.0 + IPv6" standard was released as well. Cable companies have just announced they are going to start rolling out the 3.0 standard next year. Google for docsis 3.0 and your respective company to see when you'll have it. - borez, on 10/06/2008, -0/+5You'd need a lot more then one central system.
- gabacho2, on 10/06/2008, -0/+5I'm quite sure that you did not understand the word "working" from the title.
- ExRe, on 10/06/2008, -0/+5GTFO troll.
- 350Zed, on 10/06/2008, -0/+5And where do you do this? At every egress point to an IPv4 infrastructure?
Scalability FAIL - werries, on 10/06/2008, -0/+4IPv5 was Internet Stream Protocol, or ST. It was meant to be the connection-oriented complement to IPv4 but never saw public usage.
There is this thing called Google, you should try it. - swordedge, on 10/07/2008, -1/+5network hardware cost is NOT an issue. IPv6 is software. Update the rom and BANG, you have an IPv6 router.
- CarzorStelatis, on 10/06/2008, -1/+5Am I just being facetious, or could they just handle this by sticking a whole load of leading zeros onto an IPv4 address when it transits an IPv6 system?
- TehDoctor, on 10/06/2008, -1/+4If "Router software updates" were so easy a solution, we'd probably be on IPv6 already.
- SuperOmegaSlack, on 10/06/2008, -1/+4Don't think so... If you have servers translating IP's as well as servers translating Domain Names... the amount of traffic would be ridiculous and having to process all that, the latency would be crazy!
- dalittle, on 10/06/2008, -1/+4This really seems like the solution. Put root servers out there for IP6 / IP4 translation and have it work like DNS.
- velocity92c, on 10/06/2008, -1/+4The faster we get IPv6 rolled out the better. It's a shame IPv6 wasn't the original implementation, by our standards it would have never needed updating (for IP space, anyway).
- atgmac, on 10/06/2008, -1/+4No, the problem is that IPv4 wasn't designed with the future in mind and so IPv6 has to clean up its mess.
- Nerys, on 10/06/2008, -4/+7Can't they do something like DNS servers? They translate between WORD names and IP Numbers so why not a DNS server to translate bidirectionally between ipv6 and ipv4?
- santaliqueur, on 10/06/2008, -1/+3The problem is, you never posted this solution, so we just have to direct them towards your post, problem solved. Thanks dude.
- borez, on 10/06/2008, -0/+2lol, no worries.
I think Cisco may just have beaten me to it. - borez, on 10/06/2008, -1/+3Router software updates?
- atgmac, on 10/06/2008, -0/+2Does it matter? It's just a stopgap measure. In fact, if it's slow it might encourage people to move to IPv6 to get the full benefits.
- 350Zed, on 10/06/2008, -1/+3Tunneling is not scalable, that is the problem. Static tunnels require huge manual effort to maintain, even with management tools.
- swordedge, on 10/07/2008, -0/+2They already are talking about sticking the 32 bit IPv4 address on the bottom of an IPv6 address. The problem is going the other way... properly routing a 128 bit IPv6 address while using only the 32 bits of an IPv4 address.
- poet, on 10/06/2008, -1/+3That would make the transition to IPv6 pretty redundant.
- thedez, on 10/06/2008, -1/+3They didn't figure that out already?
- swordedge, on 10/07/2008, -0/+2The problem is when you are carrying an IPv6 packet with its very extensive address from one network to another via an IPv4 address. The address space is so vast that they can't properly put an IPv6 destination into an IPv4 packet.... yet. What they are working out is a method of giving IPv4 packets destinations to IPv6 data that will fit into the available address space.
- Nerys, on 10/07/2008, -0/+1No at every point where IPV6 is used. the ipv4 equipment can talk to each other fine its the new ipv6 equipment that will be out of the loop.
- sadGuru, on 10/06/2008, -0/+1you get an A for brainstorming but the two are not related.. the problem here is routing, nat and switching packets. but there already are similar technologies in place so I dont think it will take them too long to figure it out.
- geekmansworld, on 10/06/2008, -1/+2You'd THINK that, wouldn't you...
- geekmansworld, on 10/06/2008, -1/+2That really would have made more sense, but when developing the IPv6 standard it was decided to go with a 32 character hexadecimal sequence: a 128-bit address. This was routing and other network information can be encoded into the address.
Unfortunately, IPv6 requires an overhaul of lots of physical equipment, at a time when the telecomm networks are already whining about infrastructure costs. - esc27, on 10/07/2008, -0/+1IPV6 is bloated, slow, compatible with existing technology, etc. We should just all wait for Win.. I mean IPV7.
- Nerys, on 10/07/2008, -0/+1well could it be any slower than DNS is now? I mean it only needs to be "as fast" as dns. its essentially doing the same thing?
Eventually everyone will be ipv6 and the problem will solve itself. - sancho, on 10/06/2008, -0/+1I really wish that we'd get around to serious implementations of a service-based DNS system. Yes, you can include SVR records, but they aren't used by many protocols. If we could indicate that "there's a web server listening on port X of this machine" using DNS, a lot of the problems with using NAT would go away.
It wouldn't make IPv6 unnecessary, but it would have staved it off for a little while longer, all the while being a pretty useful feature. - ispshadow, on 10/07/2008, -0/+1I was stunned to see that China has been using IVI for two years now. Really?
- swordedge, on 10/08/2008, -0/+1Use to do a piece of that. If a company has a working stack for one of their routers, then what they do is port that to the other routers. The benefit is that code updates and new routers generally have a great, reliable stack. The down side is that if there exist a code problem, all routers in the line have the same problem. In the bigger router vendors, there is ALWAYS someone writing and testing the code. It isn't static. They are always working on improvements.
- Drizzit, on 10/06/2008, -0/+1They could have stuck with IPV4 with making the backbone IPV6 only. Then each country could use the entire IPv4 footprint, NATing V4 ip's statically to V6 ip's to communicate between systems.
OR
Take all current ipv4 ip's and map them to existing ipV6 ip's. So all legacy IPV4 is mapped to a certain block of IPV6 ip's. There's enough room to do that :)
It would be no different than ipv4 private subnet spaces.
so 0.0.0.0 would map to :::::::::0 or somesuch and 255.255.255.255 would map to ::::::FF:FF:FF:FF
Format may be wrong but you get the idea. - swordedge, on 10/07/2008, -0/+1The problem is NOT mapping IPv4 addresses into IPv6. The problem is mapping IPv6 addresses in IPv4.
If you give every country its own IPv4 address space, then how do you route a US based packet to a server in Germany for example? Besides, even if you did that, we are talking about a future where you could have hundreds of devices in your house with an address. Even 4 billion addresses per country might not be enough. IPv6, on the other hand, can handle that. - dougbarrett, on 10/06/2008, -8/+9This is awesome! Vista, modern Linux distros and OS X all support IPv6 so there is no reason IPv6 shouldn't be rolled out at a quicker rate.
- Nerys, on 10/07/2008, -1/+1and bang. Oh wait. Yeah wee problem there. Someone needs to WRITE and TEST that rom :-)
- kingfoot, on 10/06/2008, -2/+2you mean one central system that knows ipv6 that can read it, and transcode it so ipv4 can read it and visa versa?
that could work but how efficient or fast will it be? - geekmansworld, on 10/06/2008, -1/+1It's about time someone looked into this. IPv6 advocates are aggressively pushing the technology, but without practical solutions for the problems it creates.
For starters, we're moving into an addressing system which, practically, is NOT human-readable, while at the same time, we're having big problems with DNS.
Second, IPv6 advocates often fail to recognize the huge installed base of IPv4-only systems and equipment. With no way to talk to growing parts of the internet, all this equipment has to be replaced. That entails a huge cost.
Instead of simply telling up that the TCP/IP sky is falling, IPv6 proponents could help the situation by developing mitigating technologies and strategies to ease the transition and help create legacy-compatible routing. - bud0011, on 10/06/2008, -0/+0w00t!
- balazar, on 10/06/2008, -1/+1I totally agree, IPv4 was not designed with the kind of expansion in mind that we have reached now. However, it is the way it is. This could have been avoided, e.g. if IPv6 had variable length addressing, then IPv4 could have become the 32 bit variant. There are no simple, useful, scalable translation or transition mechanisms.
- borez, on 10/06/2008, -6/+4Encapsulate IPv6 within IPv4 headers and carry them across the existing IPv4 routing network making sure the conversion is implemented in all new backbone routers as standard.
What's the problem? - sancho, on 10/06/2008, -3/+1Just FYI, you don't know what "facetious" means.
http://en.wiktionary.org/wiki/facetious
1. Treating serious issues with deliberately inappropriate humor; flippant.
2. Pleasantly humorous, jocular. - balazar, on 10/06/2008, -4/+1The problem is that IPv6 was designed with no serious thought to operational transition. It is on-the-wire incompatible with IPv4.
- ultrahombre, on 10/06/2008, -10/+2IPv5 is what I would imagine it would be called.
- custangro, on 10/06/2008, -9/+1dougbarrett = fail
- jattea, on 10/06/2008, -10/+1I don't understand a single word in this title and description. you dorks.
- Datix, on 10/06/2008, -19/+3But can it run Crysis?
Check out the new & improved 
© Digg Inc. 2009
— Content created and posted by Digg users is