Fred Baker, Former IETF Chair, Answers Your IPv6 Questions

deborahs · ‎09-04-2009

Have you ever wondered what’s on the mind of an Internet Guru?
What’s the 411 on IPv6, QoS, Smart Grid, IP Technologies?

Now's your chance to ask the expert. We invite you to submit your burning technology questions for this month’s featured expert Fred Baker (Wikipedia bio)

Fred has been involved in data communications since 1978, and in Internet technology since 1986. At Cisco, where he has worked since 1994, he is a Cisco Fellow, which is to say a senior engineer charged with advising on both business and product decisions. He has contributed to over 40 IETF protocols and holds several US patents and is a former member of the Technical Advisory Council of the FCC.

Fred has worked at the forefront of IP technology and can dish on topics such as:

4G mobile Internet
Role of IP in the migration from 3G to 4G
IPv6 and its deployment
QoS (quality of service)
IP routing technologies
Smart Grid

How do I post a question?
1. Log into the community with your cisco.com ID and password and post your questions in the comment section of this blog
2. Don't have a cisco.com login? Complete this short registration form and then log in to the community (top right of screen).

Post your questions now and Fred will answer them in this mobility community September 15 through September 18.

Download an Outlook reminder to add to your calendar.

DebbieDegutis · ‎09-14-2009

What similarities do you see for consumer adoption of the Internet (circa 90’s) and current market interest in a Mobile Internet? I’m thinking of AAA, HTTP, IP, IPv6?, etc.

fred · ‎09-15-2009

The early 1990's... That was a momentous era. In the 1980's and early 1990's I didn't bother telling people what I did for a living; if I said I was a software engineer, they would ask questions about banking programs (http://xkcd.com/627/), and if I said that I taught computers to talk amongst themselves they figured I had been chatting up aliens. That changed in 1994; I saw a bus drive by carrying a lipstick ad, and the only writing was a URL. Clearly someone was test-marketing the notion that the Internet could be used for information distribution to non-techies. The next time someone asked me what I did, I asked her (IIRC) if they knew what the Internet was; "I think my brother uses that at work". I used that approach for several years, until a young coed shut me down - "that's so condescending!" oops...

My model for consumer uptake is my family. My mom started using the Internet when she discovered that her grandkids would reply to email. My wife started using video on Skype when a friend moved to Israel carrying a computer with a built-in camera. My daughter hates Microsoft because it is so "helpful" and redoes things she has carefully arranged the way she wants; she's the one that discovered evite and the management of wedding/etc registries on the Internet, though. They don't give a hoot what's under the hood; they only care about the services. But they use the services.

I have been hearing about the "Mobile Internet" for about a decade now; it has been a major marketing program for Nokia, Motorola, and such, who have been trumpeting that people will discard their computers in favor of smart phones. I suppose there is some truth in that; the iPhone could be described as a computer with 3G connectivity. But in my experience, folks that can afford iPhones also have computers with larger screen real estate for the applications in which screen real estate is valuable. Also, the reports on the "Mobile Internet" tell me that, like wireline and other voice services, voice business growth is linear. What drives business on the Mobile Internet is access to services like Youtube, Facebook, and Google Maps, and more generally access outside the walled garden. It's the data side that is growing rapidly, and the issues with it that are making the newspapers.

So to my thinking, the Mobile Internet is an access network, conceptually similar to Cable Modem and DSL. It's a way that providers deliver services, and an important service is access to the rest of the Internet.

How does IPv6 come into that? The same way it comes into the Smart Grid. Imagine that you're an electric utility and want to connect to smart meters. There are many ways to do this; companies like Sensus Metering (which is a vendor to South Company, a large electric utility) are using radio technologies to connect as many as ten square miles of home electric meters to a single "pole-top" router. Much of that is proprietary now, but NIST has been challenged to lead the Smart Grid into standards-based networking. It could be - and in some places is - done using IPv4. But think about that problem with IPv4 - how does one provision that network and manage it? It's a lot easier if one can reliably obtain the address space one needs from one's RIR to build and expand the business and if one doesn't have to micro-manage address assignment. Putting 5000 homes into an IPv4 subnet is hard - I have to get a /20 per ten square mile area and assign each address individually. Putting 5000 homes into a single IPv6 subnet is trivial, and the host part of the address can be something meaningful like a MAC Address, an up-to-16 digit BCD telephone number, or a meter serial number. If privacy or identity is a concern, one can use random numbers or cryptographically generated addresses. No prolbemo, baby. It just works, and simplifies provisioning.

In my experience, network providers are an intensely practical lot (http://tinyurl.com/lm9yb7). In the late 1980's, I obtained a Class B (/16, in today's language) for my lab bench. That was an amazingly bad address allocation policy, and in the early 1990's the providers told the rest of us that they were going to use the approach we now call Classless Interdomain Routing (CIDR) - they were going to use prefixes that made sense to them. And they were right. When they saw stupidity, they did something smart. They have been holding off on IPv6 deployment because their businesses work just fine on IPv4, which makes sense from their perspective: "if it ain't broke, don't fix it." Many of the SPs I'm talking with - especially in the mobile space - are now telling me that if there is any question of them being able to get the address space they need to run and expand their businesses, they're moving to whatever technology will get them that address space. They're fussing about it of course, and I would too. But Sprint is today running IPv6-only in at least part of its network, and I'm hearing about that from more of them daily it seems.

Consumers don't especially care whether the network is IPv4 or IPv6. Many of them are unknowingly using IPv6 in their homes, though, because their computers come with it turned on and they auto-connect to each other. When we turn on IPv6 in conferences (as we did at Networkers EMEA (http://www.cisconetworkers6.com/) last year), and especially when we give them browser access to Google, we find them using it. Free, an ISP in France, announced a service last year in which people could exchange videos but had to use IPv6 and had to sign up for it. They have 350K subscribers right now. Netflix deployed IPv6 throughout their network using two engineers for two months from conception to completion (http://everythingsysadmin.com/2009/06/netflix-streaming-over-ipv6.html). IPv6 just works, for those that can get past the FUD and turn it on. Increasingly, the Mobile Internet will in fact be IPv6, and consumers will be using it.

rsynnest · ‎09-16-2009

Thanks Fred. The early Internet was an intersting era indeed (Mobile Internet Odyssey)

When we think about the pre-HTTP Internet, it was really quite limited to academics, governments, and ... well software engineering nerds like us ;-) In it's earliest days, "access terminals" were only available to people with badges and keys into buildings that had multi-million dollars worth of "computers" in them.

In my experience, the advent of HTTP and WWW offered a transformative "service" just as PC's and Mac's were becoming powerful enough to offer the visual process power at a price that was within reach of the upper end of consumer market (i.e., no longer just for academics, hobbiest, and nerds). Oh, and Ethernet started to take hold in the latter part of transition to help fuel the adoption at a price point that pretty quickly put an end to the "ATM to the desk top" efforts some of us at Nortel were considering.

Anyway ... I was thinking that the early Mobile Internet has been similar. At inception, the big radio vendors were pushing it as a concept, but in reality it was limited to ... academics, hobbiests, and nerds.

Well along comes this upstart mobile device .. the iPhone .. offering an "App Store" service, right at a time when mobile device visual processing power is there, and mobile network IP transformation is beginning (still early days!), et Voila! ... ingredients are there to ignite what I'd call a Mobile Internet Revolution.

And because mobile devices are typically a fraction of the cost of computers, and offered at a huge discount as part of a service agreement, and because the whole planet has a population that's been getting used to using mobile devices for their primary collabroative communication device (mobile phones), it feels like a transition is here that will carry us over the hump, and make Mobile Internet really take off.

To make sure it has enough "runway" for take off, I see operators must offer and adopt key technology, hidden under the covers of the visual and collaborative services people enjoy. IPv6 seem like a good one to ease managment of the billions of device to device or machine to machine "consumers" of mobile services, like "Hi I'm OK, don't shut me off yet ;-)"

fred · ‎09-16-2009

The web was certainly a transformational service; the addition of multicast voice/video/whiteboard in the LBL mbone in tools 1992 was transformational also, and gave way to commercial implementations not long after.

There is one issue that I would raise with the implements of pocket computers, which I think we agree is what the Blackberry/iPhone/windows CE/Android/etc are. Two things were critical to making the Internet as we know it happen: deployment and innovation. We certainly have GSM deployment, and a decade after the investment in frequencies are finally seeing some 3G deployment. I'm concerned that closed platforms inhibit innovation, as it's not all that easy to add new applications to them. Android differs there - it is more open. But if you want cool applications, people need to be able to write them easily and test them. Heck, I was looking at an issue last week and wanted a tcpdump run from a smart telephone on a certain network, and there was no way to do really basic diagnostics. If you want runway for the vision, it needs wings and ailerons as well.

DebbieDegutis · ‎09-16-2009

Fred - This question was sent to me by a co-worker: What roles do you see IETF and academic research play in empowering/accelerating/enabling a Mobile Internet Transformation?

fred · ‎09-16-2009

Academic research has always been a huge part of the success of the Internet. Many of the products and features we sell walked directly from the lab to the street - the concept of a datagram network, internet routing, QoS technology, voice and video on IP, the web, SIP, and so on. Solutions that were developed in SDOs without reference to academic research have been important too - H.323 comes to mind, as do BGP, IS-IS, and OSPF - but the vast majority come with strong research pedigrees. This comes back to my previous comment about innovation; if it is difficult for a grad student to put his application on your favorite telephone, he's probably going to innovate elsewhere, and the telephone or its network loses the market potential the application implies.

If you think of the Mobile Internet as 3G and 4G, relations between the IETF and the other SDOs have had issues. 3GPP's initial letter offering to collaborate with the IETF could have been read as saying "we would like you to completely redesign IP in a completely different way and deliver it to us on this date"; you can imagine how that was received. And frankly some of the solutions that the IETF has come up with don't entirely meet their needs. So they have had to learn to work together, and they seem to occasionally need to relearn those lessons. That said, both are looking hard at specific problems, and are in fact collaborating on solutions for problems like the coexistence/transition from IPv4 to IPv6 and the implications for applications running on IPv6-only telephones that need to access the Internet's servers using IPv4.

I expect both to be very much involved.

jc.bogard · ‎09-22-2009

Mr. Baker,

With IPv6 arriving already installed and in operation on many of the operating systems on the market by one of the Major vendors; is it then possible from a security stand point for IPv6 to be exploited while encapsulated with IPv4 headers? By this I am asking if one would be able to open a "back door", gain access to a local LAN or a companies LAN via encapsulating packets with an IPv4 header and having them passed through and while they are being striped with in the sever layers the IPv6 information being able to then exploit the use of it having been pre enabled upon numerous workstations, home computers, or tables used in companies?

From my limited knowledge with IPv6 it has much the same suite of "sub protocols" with in it that are with in IPv4. It would appear from a layman's perspective that those "tools/protocols" would be able to be accessed if the proper security measures wee not in place for packet inspection like an ASA device, Firewall or other robust means of defence against IP Packet penetration to compromise a network.

My cellphone provider is using CDMA and I am hoping they will move on to 4G as I have read of some places Europe. My questions concerning these types being used - and from what I recall reading earlier - enable and deploy and like IPv4, IPv6 will work seamlessly provided the gear in place is sufficient to handle (route/switch) these types of packets. Is this true? With out giving any of the various vendors a free "plug"; are their vendors with in the US who are ready to move (within the next 6 months) to IPv6?

For those of us wishing to know more of the inner workings of IPv6 with QoS can you suggest some/all of the RFCs' documenting how the two will work together. From my certification with Cisco there was very little IPv6 on that "test". I am thinking that this will increase in the near future.

A former mentor of mine had advised me that IPv6 was being used in the "core of the Internet" (not sure if it was only in testing or in full usage (most likely testing)) and that Japan was using IPv6 currently from a SP side (not local routing with in LANs).

I look forward to your replies; Respectfully,

JC.Bogard