Re: [v6ops] LISP support for draft-ietf-v6ops-rfc7084-bis-00

[Fred Baker <[e-mail]>]

> On Apr 12, 2017, at 10:44 AM, JORDI PALET MARTINEZ <[e-mail]> wrote:
> 
> [Jordi] I disagree here. Adding the actual transition mechanism (all them) in the v-01 of the document (submitted today) means 0,15% of extra flash memory 

no hats

Perhaps true. It also means that the vendor (open source is a vendor, from this perspective) allocates human resources, which cost money, to write and maintain the software. The software has some probability of bugs per volume of code, which means that there is a quantifiable security risk. Now, any business calculates its costs and potential upside benefits, and figures out how to make money (even open source developers do, believe me). Asking them to spend money with little real opportunity to gain saddles them with a cost and a risk with little upside. It's not the volume and price of the flash; it's the support costs.

If the vendor is looking at 7084 for guidance and saying there is no market for a given technology, it's because the operators in this working group collectively said there was no market in their networks. In the end it isn't about an RFC, it's about a market. No market, no money, no product. It's that simple.

If I'm hearing anything in this discussion, it's that the end game isn't found in having all of the transition technologies in product. It's to be found in native deployment that doesn't need the transition technologies.

You might find these two links interesting. One is from the IPv4 Market Group, and makes prognostications on the value of an IPv4 address; the other is a public presentation by Swisscom about the traffic measurements in its network. In both cases, they are asking when they expect IPv6 usage, whether in traffic or in users, to exceed IPv4, and as a result when the need for dual stack deployment starts to wind down. Both, by different avenues, pick 2019 as the turning point, and Swisscom suggests that by 2024 it might not have enough IPv4 traffic in its network to justify having IPv4 enabled - even in an overlay.

The issue of transition mechanisms has to do with networks being predominantly IPv4 and having to reach them using IPv6. Both are saying that there is a light at the end of that particular tunnel, and it's not an oncoming train.

http://ipv4marketgroup.com/q42016-update/
http://www.ipv6conference.ch/wp-content/uploads/2015/06/B10-Swisscom-Status_Roadmap_and_Outlook_IPv6.pdf