back to article IPv6 networking: Bad news for small biz

IPv6 is traditionally a networking topic. Yet IPv6 is as much a business consideration as it is a technical one. As world IPv6 day rolls around again, we're going to see an ever-increasing amount of technical IPv6 coverage. Before we do, I think a business interjection is warranted. IPv6 was neither designed for small biz nor …

COMMENTS

This topic is closed for new posts.

Page:

Silver badge
Facepalm

Re: IPv6 isn't happening any time soon

Hmm, I think Facebook already uses both IPv6 and IPv4, so do many if not all Google services. You'd be surprised to see how many services are already available via IPv6.

You also seem to misunderstand that it's not a "migration". Just like companies who got e-mail didn't turn off their fax machines. Of companies that got fax machines didn't stop receiving letters. People will keep their old IPv4 connections, just like mobile phone users still keep SMS and GSM telephony around.

However it's so much easier to start new services via IPv6. For example building a decentralized network is a pain on IPv4, however it makes sense if you want to do things like social networks. With IPv6 this can be done trivially.

Recovering old IPv4 ranges might save you a few months, but that would actually be a migration effort. You'd need to change the IP-adresses which is a _lot_ of work.

1
0

Re: IPv6 isn't happening any time soon

"People will keep their old IPv4 connections, just like mobile phone users still keep SMS and GSM telephony around."

In which case, everything still works on IPv4, then there is no need for IPv6 and companies won't expend resources on it. That's the point - and it is the current reality. Whilst a few companies (e.g. Google) may be implementing IPv6 alongside their IPv4 offerings, the vast majority are not and have no plans to do so because there is no benefit in such expenditure.

Recovering unused IPv4 ranges doesn't involve changing addresses because, we are discussing the reallocation of entire unused address ranges from one organisation (often defunct) to another. What has changed is that increased memory in routers allows for larger routing tables that arise when allocations are made in smaller chunks.

1
2
Meh

The short version

If we move to another ISP under IPV6, private NATed devices would require re-configuring and also IPV6 breaks multihoming.

0
0
Headmaster

Can't wait for IPissV4 to go.

IPv4 was fine for the small internet of the '90s, but its time has passed. Anyone who thinks that any type of one to many type of NAT is a good idea needs beat with FTP, VOIP, and numerous other P2P protocols (does it hurt to get bludgeoned with a protocol?, I hope so.) The number of hacks in all kinds of protocols has wasted more developer time and caused plenty of problems (SSL, Name based virtual hosting/HTTP)

We will go from NAT being our firewall to a 'real' firewall being our firewall which needs to be properly addressed on lower end routers. A lot of home/small business users do dumb crap like disable their firewalls so edge protection is going to be important. By default they should block incoming traffic and a upnp like service from the workstation should say that it wants access to certain ports opened just to it.... With the benefit that every user can get the port they want without it being in use.

As for some of the other issues I'm seeing people talking about... Address length/number of wasted addresses. The first for sets of colons are the important ones, Yes 2001:c3fd:000f:123a will be hard to learn, but after that your router will probably be 2001:c3fd:000f:123a::1, your desktop 2001:c3fd:000f:123a::2, your printer 2001:c3fd:000f:123a::3 assuming like many sme networks that you only deal with one subnet (/64 in this case). Now there are reasons not to address like that too, such as it makes it easier for hackers to scan your network.

As for people saying we should have went with smaller addresses... So you want a Y2K38 or a Y2.1K problem? In the future there will be 100's of billions of devices capable of speaking on the internet, when the IPv4 was created we didn't realize that it would come to that. Seems smarter to use a solution that will keep from running out for the foreseeable future.

5
3
Silver badge

Re: Can't wait for IPissV4 to go.

Well I know it's hard to defend the IPv6 opponents, but their views are entirely different than yours. They want the Internet to be split into 2 parties. One is Amazacegooglebook, the sole provider of services. They need about one half of the IPv4 space for their load balancers, while the rest of the addresses will be allocated to NA(P)T routers to allow users to access the services of Amazacegooglebook. All Information on the Internet needs to pass through them, because of the nature of NA(P)T. Direct communication is of course impossible.

While this reminds anyone who has seen online services like Compuserve, Quantumlink, MSN and AOL (the later 2 having moved to the Internet) of the reasons why the Internet took off while those services gradually disappeared, this seems to be what IPv6 opponents want. That's why they spread FUD about IPv6.

One lucky thing about this, is that the people who already have IPv6 can enjoy a network free of those people. It is, in many ways, a lot like the Internet used to be in the early 1990s, only faster.

4
2
Silver badge

So, when do we migrate to IPv7 (or IPv8)?

Maybe it is time to just regard IPv6 as broken, and start all over with something that will actually WORK. For the most part the first step is to have all the nice consumer grade routers to at least do IPv6 on the WAN side. Then I can deal with the IPv4 on my own local side. Unfortunately, my routers that I use (4 sites) don't have a clue. I use an "obsolete" router that has ALL sorts of nice features, like being a print server and allowing VPNs between the various routers. Oh, one additional thing, there is code in the router to setup dynamic DNS names so if the address on the WAN side changes (which it does if I restart the router) I can keep accessing my internal NAT stuff from the "outside".

Currently most DSL lines here use PPPoE to get their address and this might continue when my ISP thinks about IPv6.

Of course, we can go back/forward to IPv5, but to do that we'll need to talk to academics down the street (Stanford University) to find out what it actually does.

Summary: Oh what a tangled web we weave, when first we practice to deceive.

1
0

This post has been deleted by its author

FAIL

Ci$co and IPv6

I have to call FAIL on the Ci$co's IPv6 support. Recently bought a RV220w which Ci$co claims supports IPv6, when the implementation does not actually support prefix delegation. This makes it impossible to obtain a native IPv6 address from our ISP.

If enterprise hardware companies like Cisco still don't fully support IPv6 properly, what hope do home and small/medium business users have?

/frustrated customer

//purchased Cisco router specifically for its (claimed) IPv6 support.

0
0

Re: Ci$co and IPv6

Sorry chap, but the RV220W is not a 'real' Cisco product. Says Cisco on the box, but this is a continuation of the product development done by Linksys previously. Designed as an easy-to-use (i.e. graphical user interface) product, and not a 'proper' network device.

For less than £100 more, you could have bought a 'proper' Cisco router/firewall, like an 861W, and had 'proper' IPv6 support.

0
0

enterprises dont spend the money either

Its not just small companies that balk at the costs if ip v 6.

Most enterprise it departments have done the analysis. Cost high, disruption to business high, risk of outages high, benefits too small to measure.

0
0

Ok - you aren't living in reality..

IPv6 NAT? Do you actually understand WHY NAT was developed? NAT was developed to stave off the problems of the old 32-bit address space running out, all the way back in 1994!

NAT is not, will never be, and should never be considered by anyone with half-a-clue, as a security mechanism. NAT is easily over-ridden, tricked, fooled and generally countermanded by those who know what they are doing. NAT is about as much as a security mechanism as ITV is to fact-based entertainment.... Pointless.

Ok - so IPv6 has a few flaws. There needs to be more done to make it work. This reminds me of the good ol' days of IPv4 in the 1990's. The whole reason we are getting onto IPv6 now is that the protocol is way over due for replacement. The problems inherent in v4 were significant, most of all there aren't any addressable spaces left now. Oah - how do we deal with multicast again? lets not get started on stateful security and DNS!!

As technologies change to support the protocol (and we are only talking layer 3 and a bit of layer 4 here) then everything will be made simpler.

To a point, many manufacturers are releasing IPv6-enabled hardware. Those who don't will end up having to sooner or later.

This story is more a beat-up than anything else. I've put 43 sites onto Internode by IPv6, and they are performing flawlessly - two sites even have a dedicated AAAA with a www and mail sitting process sitting there.

As the IPv6 infrastructure rolls-in, there will be a few that decide to beat their Neanderthal chests and do the usual "woah is me!" crap. But we hear this with every change.

3
5

o_o

will there be any such thing as public access points? at mickey-D's or anywhere else?

0
0
Megaphone

IPv6 CPEs

The Billion and Draytek routers that do IPv6 are financially out of reach of home users, but fortunately there are IPv6 capable low end CPEs - The Thomson TG582n and Zyxel AMG1202 routers, both coming in at under £40, are 'IPv6 ready' and will have IPv6 firmware released for public consumption very soon now.

1
0
Pirate

The can-do engineers left IETF nearly 20 years ago!

AS one of the Internet (IPv4) pioneers, I had the privilege to work with the pragmatic engineers who deployed and evolved the Internet infrastructure over the first 10 year. By the end of the first 10 years, the IETF and other standards committees were becoming stuffed with "Goers" who's role was to push corporate self interest, rather than the community interest. Most of the "Doers" gave up and retired to do work away from these stuffed committees. (E.g. Vint Cerf went off to do Inter-planetary Internet, ). It is ever thus!

The non-deployment of IPv6 has been in the interests of the corporate ISPs as it meant they could sweat their assets until they lie totally exhausted. (This is still the case!).

The truth is all we needed was to extend the IPv4 address space by 8 bytes! Simple!! But then the network hardware companies wouldn't be able to require all the infrastructure to be replaced. Guess who sits on many of the IETF committees?

I really dislike companies that work against community interests, but I guess it is ever thus

Good Luck

Peter Dawe

Founder and ex CEO PIPEX

8
0

I am sure one can find some usecase for NAT66. Things like proxy servers are not dead either. Or load balancers. It all does the same job of redirecting traffic from one IP address to another.

BUT - the article is dead wrong about the small office with multiple internet connections. The answer in this case is much much simpler: Attach two routers to the network; DONE. Its that simple. It will work. No NAT66 needed.

With IPv6 both routers will announce themselves to the network. All computers will pick up two addresses, one for each router. All computers will continuously monitor the health of each router and within 30 seconds of a failure switch to the other.

And yes, the office printer will also just work. The software knows about the automatic renumbering, it is smart enough to use link local addressing, which never changes, for talking to the printer. And using multicast for device discovery on the network.

What about serving content to the public internet? The author links to a convoluted IPv4 NAT solution which does not provide a good solution for that either. Currently the only way to be true multihomed as a content provider is to use BGP - the same for both IPv4 and IPv6. But this is changing with the LISP proposal: http://www.lisp4.net/

I believe LISP will basically solve everything the author is complaining about. No it is not mature yet, neither is NAT66 really. LISP is a MUCH nicer solution however with many other benefits too.

0
1

Page:

This topic is closed for new posts.

Forums

Biting the hand that feeds IT © 1998–2017