.eof

That doesn't sound inviting. I'm outta here.

Or you can hang around and try to get quantumly entangled

Thank you everyone for joining us today. This ends our session.

Bye for now!

Thank you again!

thanks for the chat - See you on the Net!!

Thank you for meeting with us.

future search engines will have more semantic understanding of the content of the Web; I hope!!!

have a look at the "more" choice for Google Search to get to Google Labs where you will see a lot of experiments under way.

Thanks for your time and intersting chat, Dr. Cerf. Great stuff!

this was a pretty efficient way to get a lot of Q&A done.

Is there going to be any new techs or features added to Google search anytime soon? what would a future search engine look like? Getting more intelligentI guess, right?

any other quesitons quickly from the floor?

Dr. Cerf, thank you very much for a most enjoyable session

we're just about out of time. these were good questions - i hope you enjoyed the answers!

Malware is a never ending battle: one needs to tighten up the browser; make the operating systems less vulnerable, implement DNSSEC, run IPSEC end to end (or TLS at least) for most applications.

that's hard to say but I notice that Verizon FiOS uses IP to stream video from DVR to set top box; ComCast is moving to IPv6. I think we will see a mix of conventional digital and packet switched entertainment communications. I will know we've made progress with IP packets come raining down from satellites!

Is the war against malware winnable, in your view, with the security protocols you mentioned below?

How soon will we see the Internet as the basis for all broadcast entertainment?

among my favorites is NASA with all the great star/constellation protos.

my most frequently used web-site is Google.

when you broadcast on a satellite or radio channel, more than one receiver may hear the transmission. If our protocols took that into account we might be able to implement broadcast/multicast much more efficiently than we do today.

What is your favorite web site?

I do not think it is possible for any one group to literally govern the global Internet - there are too many jurisdictions and authoritative parties including all the ISPs. ICANN has only a coordination role to assure that domain names and IP addresses are allocated fairly and uniquely.

Can you elaborate on better use of broadcast media?

it is generally a very crude measure and makes it easy to cause users to slip into hyper-expensive territory while throughput limits allow users to try new services without fear of a massive and unexpected bill.

Is it possible or advisable for one group to take over governance of the Web? Can ICANN be that group? Should someone else take over?

My experience indiates that such limits - bytes transferred - is more common in North America, and to some extent Europe but not in other patrs of the Globe, i.e. in "less developed" regions

as to starting over - that's partly what the NSF FIND preogram is about. It isn't clear yet that there is a radically new idea although I like exploring better use of broadcast media. As to standards groups, IETF and W3C are two key groups.

some ISP put total byte transferred per month limits on users and charge a premium when that total is exceeded. This is an extremely crude approximation to actual bandwidth demanded by that user - that is really the limiting resource, capacity.

In your view, which standards groups will be most instrumental to the future of Internet Protocol standardization?

I know you mentioned quantum entanglement as a possibility, but the packet  switching technology has been around since beginning on networking. Is it an outdated "technology"? Would starting over from scratch with some smart minds on better methods for handling large volumes of data?

ah, would you not expect ISP from different parts of the globe to have very different views on IPV6 and DPI?? 

and it would be great to ask ISPs what their IPv6 deployment plans are!

In re "I am not a big fan of putting volume  limits on users - only rate of transmission..." Can you expand a bit?

it would be very helpful to hear more from users what areas of internet operation concerns them most.

I think the ISPs will need to think about how to manage user consumption by limiting throughput especially in times of congestion. I do not think that deep packet inspection is needed to achieve this management principle. Protocol and site-independent traffic sharing algorithms and enforcement mechanisms are needed. I am not a big fan of putting volume  limits on users - only rate of transmission.

Do you feel that as "citizens of the Internet" we need to take a more proactive effort in aiding the improvement of this digital community? Example push ISPs to adopt IPv6 or contribute more in efforts to deliver Internet to rural areas.

Is any particular organizaton working on Parametric Bounds Checking?

configuration management reliability is really a tough problem because configurations are often not obviously wrong. Parametric Bounds Checking is one helpful tool.

Are you concerned that recent FCC rules re Net Neutrality, while shaping ISPs'  recent management practices, will motivate them to impose bandwidth limits and tiered services that may inhibit use of the Internet?

generally it is good to be thoughtful about privacy. Google is extremely aware of the importance of privacy to its users. We go out of our way to defend our users from attempts to obtain personally identifiable information; we try to detect malware and steer users away from it. We anonymize information that we capture regarding query usage or ad clicking and discard data after 9 months and in some cases sooner. Our interest is not in individuals but in patterhs that we can match with advertising material.

Do you have any suggestions as to improving configuration management? A couple of years back, everyone was talking about config mgt. databases, for example.

A lot of our readers are worried about Internet privacy, especially in regards to companies like Google keeping a record of what they are searching on and then using that information to target them for other services. Are they right to be worried?

Google continues to look for additional ways to develop revenue producing services although we are very happy with the current advertising model because it has allowed us to implement and deliver so many different applications.

quantum entanglement has been demonstrated over distances of kilometers over optical fiber.

Google has made 55 acquisitions but still makes most of its money from search advertising. What's the next big revenue opp for Google? When will it kick in?

network management is doing moderately well but configuration mistakes are probably the single largest source of failures. We coiuld use some new ideas there.

Quantum entanglement could theoretically allow for instantaneous transmission of data to anywhere in the universe, but from a practical standpoint we've never been able to achieve entanglement outside the confines of a lab, right?

Has Google been recommending to the ISPs the adoption of IPV6?

Can you speak to how well network management is going on the Internet? With DTN, will management advance any?

The ISPs claim there is no market for IPv6. My view is that when we run out of IPv4 they will be looking for a solution and IPv6 is the only one in sight.

regarding strong authentication, I think it would be very useful to have in hand ways to strongly authenticate persons and equipment and even software in the Internet. It would aid in avoiding malware or being phished or pharmed, for instance.

What's the holdup on IPv6 in your view?

DTN is on board the International Space Station thanks to the BIONET team in Colorado. As for overhauling today's Internet, we need to deploy IPv6 and several security measures, notably DNSSEC, IPSEC and secured BGP.

Here's a non-interplanetary question... At the Gov 2.0 Summit you noted that if you could start over you'd put in a much stronger focus on authenticity or authentication into the Internet's architecture... Can you elaborate on why that's important and anything else you'd do differently?

If we set aside the not insignificant issues of budget and time required, how ripe for overhaul is today's Internet, in your opinion?

well TCP/IP has deliberately avoided detailed dependence on any particular technology. One hopes that its sucessors will have a similar resilience and ability to adapt to virtually any signalling method.

Is DTN being used on the International Space Station presently?

More so than TCP/IP does now, perhaps?

as it always has!! :-)

So in the future, advancements will rely on underlying optics and/or hardware as much as software?

mostly used today to transmit encryption keys in such a way that you can detect any interception.

quantum entanglement is an active area of optical research.

whether it is possible to FORCE the state of an entangled particle is not known but if it could be set to any desired state, you could actually communicate over arbitrary distances instantaneously.

Sounds like optical networking advancements... Is that the area where this is being researched?

at the moment, quantum entanglement appears to allow one to examine an entangled photon or electron and to know that a distant entangled counterpart has the opposite state of the particle you measure/examine.

Right... quantum entanglement?

Have you always worked in the networks arena or have you ever ever taken a break from it and worked in another area?

Hm... you lost me on that last sentence, Dr. Cerf!

It may persist for quite some time (TCP/IP) because it has shown ability to adapt to a dynamic range of over 6 orders of magnitude. However, I can imagine some interesting variations such as better use of broadcast media (rather than treating it as point to point as TCP/IP does). Or maybe quantum entanglement will allow for instantaneous transfers.

because of the tolerance of delay it isn't entirely clear whether traffic management is a good application - depends on the average delays. Usually it is bad to build a control system that has highly delayed feedback. As to legal ramifications, if one encrypts the packets, privacy can be protected. Relying on other parties who voluntarily relay traffic works for mobile ad hoc networks operated by an single entity. Cooperation among distinct parties might be a more complex proposition.

We've been discussing specialized applications for DTN. But in the future, will TCP/IP be superseded on the Earth-bound Web as well? When?

no, this would not be a real-time system unless communications could be more readily assured. i agree that this is more likely to be used for longer term observations.

What legal ramifications may be involved with the intervening nodes between the message sender and recipient?  I would think that data ownership may come into question at somepoint, particularly with "indefinite" storage at those nodes.

Also, what about the automotive network? Do you foresee a traffic management for logistics?

I'd imagine deep sea exploration.

Good for tides and temps, but not so good for tsunami warnings, then

What kinds of research would this support?

Data would be stored in the ocean buoys or in sensors under the ocean surface. At intervals data could be transferred to an orbiting satellite- that's what happens with todays rovers on mars.

Can you elaborate a bit on the oceanographic sensing application you referred to earlier and how DTN would be used that way?

Flash storage may be too slow - there are new solid state, random access recorders that are commonly used for space applications that might apply for DTN.

yes, DTN assumes that it can store content in the nodes - that allows it to recover rapidly from link failures but it does imply that packet/bundle content can be temporarily stored. IP does not make that assumption.

In DTN, there's talk about the need for reliable storage. Does DTN require MORE storage than other protocols? And if so, do you recommend flash storage as best?

Programming languages: Python, C+ are the most commonly used for DTN. I don't regret advances that improve on TCP/IP!!

JPL and I (personally as visiting scientist) started the program in 1998, the year Google was born. I didn't join Google until 2005.

As a co-inventor of TCP/IP, does it give you any regrets in any way to see TCP/IP superseded?

Is there a particular programming language that has been used for the applications utilizing DTN?

Google is not exactly working with NASA on DTN - this is just my 20% time effort that Google permits

ok, I missed that

At the moment, there are experiments with rate-based protocols (such as XTP) that might do better than TCP (which still uses slow start to avoid congestion and packet loss). TCP and related protocols above the IP and UDP layers have also done well in real-time operation.

Dr. Cerf has answered the two most recent questions

If the delay is that much important, will DTN be suitable for an earth deployment too?

How long has Google been working with JPL on this technology?

In terms of security, DTN does more than TCP/IP. It requires nodes to authenticate strongly before any packet forwarding will happen. It uses digital signatures and signed hash to assure integrity (or detect failure of integrity)

Ah. Which alternatives, in your view, in addition to DTN, stand out as the best TCP/IP alternatives?

In parts of the globe where weather plays havoc with satellite tx/rx DTN sounds like it may have an application

the Future Internet Design effort of NSF (FIND) is exploring alterative Internet architectures and can use GENI for test bed purposes

How does DTN stack up against TCP/IP in terms of security and network management?

GENI is not an alternative to TCP/IP. It is a testing network to allow exploration of alternatives to the current Internet design.

yes, I agree that rural areas might benefit from this - we have done tests with villagers in the northern parts of sweden (near Lulea) using this technology. Sort of like "data mules"

We've heard about the Global Environment for Network Innovations (GENI) as another alternative to TCP/IP. Is there any relationship being forged between DTN and GENI?

I have an appointment as a visiting scientist at JPL but I am a full time Google employee.

Sounds like rural areas in developing nations could possibly benefit from this technology.

Are you working with NASA as a Google employee or as an independent contractor?

Guarantee is a strong word but TCP is intended for reliable delivery. It does not do well, however, in frequently disrupted networks where end-to-end communication cannot be maintained.  As to spectrum, DTN can theoretically run over any spectrum.

DTN does better than TCP/IP in disrupted environments because it stores data "in the network" and holds it until a link returns to operational status rather than discarding packets as happens in the Internet.

Dr. Cerf, what slice of the wireless spectrum would this ad hoc network operate at? or would this be part of the white spaces?

But we have had guaranteed delivery prior to DTN, no?

SNA is an older IBM protocol that has long been pretty much superceded by TCP/IP. CMIP is the counterpart to SNMP in the network management world.

certainly this would work for oceanographic sensing.

Can you explain -- for a dummy like me -- how DTN differs from other guaranteed delivery protocols, such as proprietary ones like SNA or CMIP?

That's true. I forgot about Google Moon.

Perhaps this could also be applied to generate more information in hard to reach places of the Earth.

that's another possiblity especially for instrumenting automobile performance. It is also possible to imagine a mesh of cars forming an ad hoc network.

Sorry, automotive applications?

well, Google Earth already has Google Moon, Google Mars and Google Sky!!!

right... curious about other, more earthbound applications of DTN and its high delay-tolerant characteristics

What about automative?

We have also tested in military tactical scenarios with good results (not using Android but using military radios)

It sounds like this technology could one day pave the way for expanding Google Earth into Google Solar System . . . as well as a greater information collecting of the solar system for research and educational uses.

We have also ported DTN to Android (the mobile operating system that Google developed) and are also looking at potential testing with Intelsat 14, so there might be some testing and experimentation in civilian environments in 2010.

The delays can be arbitrarily long. Mars is 3-20 min away from Earth but we can make this work across the solar system even with hours of delay or more.

Great. What about commercial deployments? What would those be, and are any planned specifically?

we started deep space testing last Nov with the Deep Impact spacecraft. We will be using the technology on the EPOXI spacecraft that is en route to visit a comet in 2011 (this is actually the same spacecraft, renamed. IE, it is the Deep Impact spacecraft). Later Mars missions are targetted for possible use of DTN.

Also, last year, I read that DTN will support delays of 3 minutes to 20 minutes from earth to Mars. Has that improved, stayed the same?

Are there specific space missions that will use it?

the work began in 1998. We are now at the point where we are doing space qualification testing - at the end of 2009 be believe we will qualify for Technology Readiness Level 8 designation meaning "ready to deploy in operational scenarios"

The last I heard, DTN had been tested at NASA last November.

Maybe you could start by updating us where things stand with this interplantery protocol you've been working on

Here's a first question: Where exactly is Disruption Tolerant Networking (DTN) in its course of development?

I am glad we found the time. How would you like to proceed?

Welcome, Dr. Cerf... thanks for joining us here today

We're delighted you could join us.

HI it's Vint Cerf. The chat room does not work with Safari (MAC) so I had to log in with FireFox.

Welcome, Dr. Cerf!

lol... I think we'll try to stay on topic with the future of hte Internet

Ha. Maybe... Depends on how good the other questions are.

Is Nicole going to ask about what is Sergey Brin's favorite breakfast meat? 

Good afternoon everyone.

I just spoke with Dr. Cerf's office. He's about to log on

Afternoon, everyone... thanks for joining us for this live chat with Vinton Cerf

Dr Cerf, Google has made 55 acquisitions but is still largely dependant on one revenue stream (search advertizing). What is the next big business opportunity for Google?

I understand that the interplanetary replacement for TCP/IP is delay tolerant.  To what degree?  Can there be days (earth time) or weeks between packets?  How is packet synchronization handled in those cases?

Less serious question; How did you convince Google to give you such a radical corporate title? Did you think it up or did someone at Google do it? ;-)

Dr. Cerf, you recently urged companies to migrate to IPv6. Can you expand on your concerns? Is it the Internet of things (ala RFID) or the pervasive computing that will be the tipping point?

Thank you Firefox.

I guess Safari doesn't like your chat window, so I have to reask a question:

Is the interplanetary Internet for when we have servers on other worlds (or space stations)? it's not something for now to communicate with the ISS right now, because they seem to be already connected.

Dr. Cerf will select the questions he wants to answer during the live chat session

This is how questions will be asked

Hi Mary

Hi there

Join Internet Evolution for a Live Chat session with Vint Cerf on Monday, Sept. 28, between 2 and 3 PM EDT.