Re: When IPv6 ... if ever?
If I charge a customer more for IPv6 connectivity than for IPv4 connectivity, to offset the costs of dealing with ships-in-the-night routing (deploying it, training everyone to understand it), do you think my entire customer base is going to transition over to IPv6? Ask yourself, as an ISP, how much more you are willing to pay your transit providers for IPv4 + IPv6 transit, and how you are going to get the money for that and for the deployment/retraining costs. Then ask yourself, as an ISP, what benefit you get from IPv6. My answers: not a chance, none, and zero, respectively. Sean.
On Sat, Sep 02, 2000 at 10:40:32AM -0700, smd@clock.org wrote:
Then ask yourself, as an ISP, what benefit you get from IPv6.
My answers: not a chance, none, and zero, respectively.
At some point, when ARIN just stops issueing IPv4 address space, I would say it would be in a {I,N}SPs best intrest to switch. As an ASP, if my uplinks offered ipv6 address space/routing, I'd get it and start working with it _now_ rather then later. -- Bill Fumerola - Network Architect, BOFH / Chimes, Inc. billf@chimesnet.com / billf@FreeBSD.org
Bill Fumerola wrote:
On Sat, Sep 02, 2000 at 10:40:32AM -0700, smd@clock.org wrote:
Then ask yourself, as an ISP, what benefit you get from IPv6.
My answers: not a chance, none, and zero, respectively.
At some point, when ARIN just stops issueing IPv4 address space, I would say it would be in a {I,N}SPs best intrest to switch.
As an ASP, if my uplinks offered ipv6 address space/routing, I'd get it and start working with it _now_ rather then later.
--
Envision yourselves in a position where the company you work for has allocated all 1918 addresses and a new application comes along. Let's say it's wireless devices that perform a function to streamline 20 jobs and must communicate with the supplier's network. This supplier also supplies all of your competitors, so global end-to-end connectivity is required. Let's say there are 4000 devices per remote site with 4000 sites. Am I going to get that ipv4 space from ARIN? NOPE! They want extremely sensitive information, inlcuding purchase orders, from me to justify it. However, how can the project proceed if the addresses are unavailable? We cannot commit to the order until we know if it is deployable. Enter v6 with a slightly better policy (not the one I had hoped for, which was free or low cost v6 space to those who could justify it using the same justifications [save the purchase orders] as for v4 space) and we may just consider it. IPv6 is not saleable to anyone until the allocation policies come into line with reality. The reality is: there is NULL commercial demand right now so there can be no justifiable charge for the space. This environment does not in any way foster the startups that can prove a new application or business model. Note that I am specifically not stating that we must institute a land grab for v6 space. It must be justified, but the current justification and pricing models keep business out while we all wait for the inevitable crisis. -Nathan Lane
smd@clock.org wrote:
If I charge a customer more for IPv6 connectivity than for IPv4 connectivity, to offset the costs of dealing with ships-in-the-night routing (deploying it, training everyone to understand it), do you think my entire customer base is going to transition over to IPv6?
Ask yourself, as an ISP, how much more you are willing to pay your transit providers for IPv4 + IPv6 transit, and how you are going to get the money for that and for the deployment/retraining costs.
Then ask yourself, as an ISP, what benefit you get from IPv6.
My answers: not a chance, none, and zero, respectively.
Sean.
Now we're hitting the nail on the head. Well said Doran !! dave
On Sat, Sep 02, 2000 at 10:40:32AM -0700, smd@clock.org wrote:
Ask yourself, as an ISP, how much more you are willing to pay your transit providers for IPv4 + IPv6 transit, and how you are going to get the money for that and for the deployment/retraining costs.
Then ask yourself, as an ISP, what benefit you get from IPv6.
As yourself, as an ISP if you can afford not to be testing IPv6 today. I don't think it's ready for deployment on any wide scale, due to a number of factors. However, I do believe it will happen, and sooner rather then later, due to the demand for addresses. ISP's should not be offering it as a service yet, or charging extra for it, but they better be working on figuring out how it works, so when the day comes they can convert quickly. -- Leo Bicknell - bicknell@ufp.org Systems Engineer - Internetworking Engineer - CCIE 3440 Read TMBG List - tmbg-list-request@tmbg.org, www.tmbg.org
On Sat, Sep 02, 2000 at 09:17:42PM -0400, Leo Bicknell wrote:
As yourself, as an ISP if you can afford not to be testing IPv6 today.
Absolutely... IPv6 is not so much different from IPv4 that we need to do years of testing, particularly before many vendors provide solid IPv6 and IPv6 routing support. IPv6 routing will change drastically before it becomes production, and therefore, any operational experience collected now will be nearly useless; particularly with the lack of any volume of IPv6 traffic. All this assumes that it will become production, which is going to be a very slow process if it happens at all. As an ISP, I would have to be insane to collect expensive and useless (for several years) IPv6 allocations when I have my (useful) paid for or grandfathered IPv4 allocations. It doesn't make any business sense. The model is broken, and as long as it's broken, large-scale deployment of IPv6 will be extremely slow if it happens at all. --msa
Sean;
Then ask yourself, as an ISP, what benefit you get from IPv6.
My answers: not a chance, none, and zero, respectively.
Don't you think it a benefit that you are allocated a large block of IPv4 address space if you support IPv6? <draft-ohta-address-allocation-00.txt> says: Usage Based Address Allocation Considered Harmful The current usage based IPv4 address assignment policies might have prolonged the useful lifetime of IPv4 address space but this has to the detriment of the the end-to-end architecture of the Internet. This memo proposes the adoption of an address assignment strategy that releases large blocks of IPv4 address space into the Internet. The objective of this policy is to encourage healthy Internet deployment models with end-to-end transparency and association of permanent connectivity with a stable IP address. This is intended to encourage provider support for open transparent Internet service environments that can be sustained with the adoption of IPv6. and there will be some experimental assignment "soon".
| At some point, when ARIN just stops issueing IPv4 address space, | I would say it would be in a {I,N}SPs best intrest to switch.
This is not "soon".
That's why usage based address allocation is considered harmful. However, deployment of the policy in the draft above makes it real soon. Of course, more serious problem is in multihoming but the solution, as usual, is to rely on the end to end principle as described in: draft-ohta-e2e-multihoming-00.txt Masataka Ohta
On Sun, Sep 03, 2000 at 06:14:54PM +0859, Masataka Ohta wrote:
Don't you think it a benefit that you are allocated a large block of IPv4 address space if you support IPv6?
At which point you say Thanks, use the new v4 addresses, and stall on your v6 implementation. --msa
msa;
Don't you think it a benefit that you are allocated a large block of IPv4 address space if you support IPv6?
At which point you say Thanks, use the new v4 addresses, and stall on your v6 implementation.
Read the draft. The policy do require something. However, the policy do not require IPv6 used. The policy do not even require allocated IPv4 addresses used. Masataka Ohta
smd@clock.org wrote:
If I charge a customer more for IPv6 connectivity than for IPv4 connectivity, to offset the costs of dealing with ships-in-the-night routing (deploying it, training everyone to understand it), do you think my entire customer base is going to transition over to IPv6?
Ask yourself, as an ISP, how much more you are willing to pay your transit providers for IPv4 + IPv6 transit, and how you are going to get the money for that and for the deployment/retraining costs.
Then ask yourself, as an ISP, what benefit you get from IPv6.
My answers: not a chance, none, and zero, respectively.
Sean.
Sean, You speak in such extremes of your vision of what we should do with our applications and usage of IP. Might you expand your vision to include others? How are you going to support a customer who legitimately needs a /8v4 worth of end-to-end connected devices and has the purchase orders to prove it? How would said customer, though, envision such an application if they couldn't get an allocation for it? They wouldn't and your business would never see it. The application would either die or find another provider or protocol. In my place, the application isn't dying. I've been fighting for SNMP proxies (to manage the existing plethora of devices), but these SNMP proxies don't exist and leave the wonderfully developed SQL databases so carefully designed over the years useless (and I hate proxies...I like end-to-end). Management doesn't like hearing you can't know the number of cans of Coca Cola in the machine outside store 359 at this "right now" second. Read about Televend. A $4.50 radio in vending machines. Ouch on the IP address usage for that application. How many vending machines do you cross day to day? We can NAT these to our hearts content, but eventually it must connect to the true supplier and that requires end-to-end. How can ipv4 support that load? I don't think it can. Think ahead. My xeroxed copy of "RFC 1" is a fascinating journey as it was written shortly before my wife and I were born. It discussed the finances and realities of building the network from Santa Barbara and UCLA to Salt Lake City @ 1200 bps across the Mojave Desert. It discussed the costs of 2400 bps and the stations required to make it a reality. Think about the future applications. My own three children do not know what a modem is. They expect and are delivered IP connectivity and if any of them reports network trouble, I know I have failed in my delivery of full connectivity for all applications. (Their access is filtered; but my son's very mention of DHCP sends my hackles rising.) Business is now, indeed. Applications are later. -Nathan Lane
On Sat, 2 Sep 2000 smd@clock.org wrote: :Ask yourself, as an ISP, how much more you are willing to pay your :transit providers for IPv4 + IPv6 transit, and how you are going :to get the money for that and for the deployment/retraining costs. : :Then ask yourself, as an ISP, what benefit you get from IPv6. : :My answers: not a chance, none, and zero, respectively. : Has there been any studies done on IPv6 as an alternative to NAT? Besides IPSec, dynamic addressing, authentication and improved security, are there other benefits to deploying IPv6 instead of NAT? -- batz Chief Reverse Engineer Superficial Intelligence Research Defective Technologies
NAT is more than just a means to ease IP address space use. The use of a dynamic NAT pool allows the hiding of internal IP topology, thereby increasing security. On Thu, 7 Sep 2000, batz wrote:
On Sat, 2 Sep 2000 smd@clock.org wrote:
:Ask yourself, as an ISP, how much more you are willing to pay your :transit providers for IPv4 + IPv6 transit, and how you are going :to get the money for that and for the deployment/retraining costs. : :Then ask yourself, as an ISP, what benefit you get from IPv6. : :My answers: not a chance, none, and zero, respectively. :
Has there been any studies done on IPv6 as an alternative to NAT?
Besides IPSec, dynamic addressing, authentication and improved security, are there other benefits to deploying IPv6 instead of NAT?
-- batz Chief Reverse Engineer Superficial Intelligence Research Defective Technologies
On Thu, 7 Sep 2000, Dana Hudes wrote: :NAT is more than just a means to ease IP address space use. :The use of a dynamic NAT pool allows the hiding of internal IP :topology, thereby increasing security. My questions took for granted that v6 does both. I was just wondering if there has been a comparison done, or there were published studies of one as a substitute for the other. -- batz Chief Reverse Engineer Superficial Intelligence Research Defective Technologies
On Thu, 7 Sep 2000, batz wrote:
Has there been any studies done on IPv6 as an alternative to NAT?
Besides IPSec, dynamic addressing, authentication and improved security, are there other benefits to deploying IPv6 instead of NAT?
On Thu, Sep 07, 2000 at 11:05:50AM -0400, Dana Hudes wrote:
NAT is more than just a means to ease IP address space use. The use of a dynamic NAT pool allows the hiding of internal IP topology, thereby increasing security.
for people that use nat to increase the number of machines that can access network services, ipv6 can do that too, but without nat. for people who want to use nat to increase their security, ipv6 can do that too. or maybe some people just really like using nat? -- |-----< "CODE WARRIOR" >-----| codewarrior@daemon.org * "ah! i see you have the internet twofsonet@graffiti.com (Andrew Brown) that goes *ping*!" andrew@crossbar.com * "information is power -- share the wealth."
Batz;
Has there been any studies done on IPv6 as an alternative to NAT?
Besides IPSec, dynamic addressing, authentication and improved security, are there other benefits to deploying IPv6 instead of NAT?
Are you saying that there has been some studies done on IPv6 that it does offer dynamic addressing, authentication and improved security? Where can I find it? Masataka Ohta
On Fri, 8 Sep 2000, Masataka Ohta wrote: :> Besides IPSec, dynamic addressing, authentication and improved :> security, are there other benefits to deploying IPv6 instead of :> NAT? : :Are you saying that there has been some studies done on IPv6 that it :does offer dynamic addressing, authentication and improved security? : :Where can I find it? I'm assuming you're being facetious. Those features can be done with v4 and v6, using DHCP, IPSec, and ideally some other features. If v6 is going to come into widespread use on the net, it has to be in production somewhere. Firewalled corporate networks are as sterile an environment to unleash it as any. As it stands, NAT was just a hack to conserve address space, and now that there are 'functional' v6 implementations maybe it's time to start thinking of a strategy for deployment. The reason I was asking if a study was done was to find out if there was any good reason, beyond curiosity, to deploy v6 on private production networks. If not, how long should we expect to have to tolerate vendor hubris and bad hacks to get around depleted address space? -- batz Chief Reverse Engineer Superficial Intelligence Research Defective Technologies
Batz;
:> Besides IPSec, dynamic addressing, authentication and improved :> security, are there other benefits to deploying IPv6 instead of :> NAT? : :Are you saying that there has been some studies done on IPv6 that it :does offer dynamic addressing, authentication and improved security? : :Where can I find it?
I'm assuming you're being facetious.
I* (including but not limited to "I" and "IPv6") are facetious, of course.
If not, how long should we expect to have to tolerate vendor hubris and bad hacks to get around depleted address space?
First, vendors of IPv6 address space should seriously tell vendors of Internet service supply IPv6 service. Then, vendors of Internet service should seriously tell vendors of routers that they really supply IPv6 capable routers. And there will be a v6-capable Internet, only after which there will be some good reason, beyond curiosity, to deploy v6 on private production networks. And then, we can get around depleted address space. Masataka Ohta
participants (10)
-
Andrew Brown -
batz -
Bill Fumerola -
Dana Hudes -
Dave Morton -
Leo Bicknell -
Majdi S. Abbas -
Masataka Ohta -
Nathan Lane -
smd＠clock.org