The panel was moderated by Leslie Daigle, Chief Internet Technology Officer, ISOC. Panel members were John Brzozowski from Comcast Cable, Erik Nordmark from Arista Networks, and Chris Palmer from Microsoft. IPv6Now’s Professional Services Manager Michael Biber has provided this summary:
LD: Good news – in June 2012 there were 69 major networks with measurable IPv6 deployment, but as of October 2013, there are 197. IPv6 is now part of regular business. The five most visited website in the world are accessible over IPv6. Major ISPs understand IPv6, but enterprise networks still don’t – what’s stopping them?
JB: Comcast has now 60% IPv6 deployed and expect 100% in 12 months. Customers using IPv6 are now 25%, we expect 50-60% in 12 months. We are converting our internal corporate networks to IPv6. We use network modernisation as a means of implementation. When we turned v6 on, we found that 5% of the (12M+) customers had IPv6-capable CPE. This was a surprise – 5% isn’t that high but it isn’t 0% either. In a few months we plan to trial an opt-in IPv6-only service.
EN: Arista sees it’s hard to get enterprises to adopt IPv6, they have been very reluctant. BYOD (bring your own device) would seem to be a key driver. Someone using IPv6 at home or on mobiles may find things break, underperform or have security vulnerabilities at work, if they don’t have IPv6 running there. Large-scale virtualisation is a natural for IPv6, simplifies virtual networks. We recommend the virtualised underlay be IPv6, with IPv4 legacy apps supported on top. All software development should support ‘IP’ without regard to v4 or v6. All new development should be IP-agnostic.
CP: Microsoft sees the Internet as the key issue. ‘Success’ is that the Internet just keeps working. Without IPv6, universal end-to-end Internet connectivity and services will cease to exist. Islands of connectivity will emerge and ubiquity will break down. We see 30 times the present number of Internet-connected devices by 2020 [MB: that’s only 6 years away now!]. Issue – how to get away from the stepping-stone transition technologies – Teredo, isatap, 6to4 etc – and go straight to IPv6. T-Mobile now has Android devices with IPv6 only – IPv4 is handled by translation.
Leslie Daigle then asked for questions from the floor to talk about the v6 Business Case.
- The business challenge is getting approval from finance management to make the conversion. The techies get it, the business managers don’t.
- Treat it as the IP business case, not v4 or v6 … IP is a given – v4 for legacy, v6 for the new.
- CP: Asia-Pacific ISPs are strangely using carrier grade NATS rather than v6. Hard to avoid the tidal wave that is IPv6 transition, and it will lead to local, regional partitioning, degraded services. They need to get serious. Although raw Asia Pacific numbers are good, %v6 traffic is still low compared to Europe and the US.
- Singapore – incumbent ISP declared ‘no islanding’ – forced widespread v6 deployment.
- CP: Microsoft made the point that Skype is not v6 enabled and it’s their product now … getting finance management to approve the expenditure is the key barrier and even they are struggling. They get it; Xbox has v6 for performance and security … why not Skype?
- Emergence of large-scale virtualisation and BYOD data will drive business to IPv6.
- Microsoft commented there’s no point in waiting for customers to ask for IPv6. Except for the gaming community, no other customers care whether it’s v4 or v6. If the ISP community waited for customer demand, v6 wouldn’t be universally deployed by 2020 – it will be more like 3020!
- The business case for customer requirements is for a globally routable secure unique address. This is what will drive v6 – not the technical debate of v4 versus v6.
- LD: the IPv6 Launch was meant to crack the chicken/egg problem, getting content and network providers to turn on IPv6 at the same time.
- We need to create milestones – public, industry, vendors. Next milestones? Moving IPv6 from the home to work without problems. Using it in business travel.
Question from floor: Is IPv6 inevitable?
Panel: Yes. No alternatives. It’s inescapable. Nothing else is possible. IPv6 enabling is not the end goal, that’s just a painful period to get through. Success is when IPv6-only is viable.