On Tuesday night I went to see Vint Cerf speak at Google’s Open House in London. As one of the guys who started the Internet it was interesting to see where he thought it was going next.
How the Internet is Changing
He first covered how the Internet is changing, who’s connected, what they’re connecting with, etc. I was surprised to hear that Asia already has the most people connected, although less than Europe and North America combined. But with most of the World’s population eventually they will have the most people connected so we should expect a shift in culture and content of the Internet.
At the moment there are about one billion people connected to the net. There are about three billion fixed telephone lines and already two billion mobile phones. The number of mobiles is growing while the fixed lines are static so there should be a shift to more mobile connections (the consequences of that are covered later), but this is countered by the speed of the connection. Vint stressed that mobiles are going to become a much more important part of the Internet, but I think the speed and cost of the connection has to change dramatically first.
Broadband is becoming more prevalent, which means more and more people are online all the time with fast connections. He came up with the term ’21st Century dialtone’ to describe how an Internet connection is something we assume to be there. A consequence of this is that you can provide more every day services over the Internet.
A thought one such example of this is VOIP, turn the broadband connection into a cheap phone. But Vint brought up a very good point, the money is made in the Internet/POTN bridge and placing the call, this isn’t always going to be necessary. Once everyone has a connection all the time VOIP is going to become easy, and profitable as email.
Digital Media
Faster connections means you can also start doing more things with digital media. Having warehouses and inventory is expensive, and it gets a lot more expensive if you try to scale it. As a result selections aren’t always the best, i.e. you get the current best sellers. By digitizing media you slash your costs and the marginal cost of supplying an additional title is so low you can start taking advantage of the long tail, i.e. low demand titles, but by selling enough low demand titles you make enough money for it to be worthwhile. He estimates the long tail may increase revenues by 30%.
There is still a way for this though because the connections still aren’t fast enough. He used Netflix, an online DVD rental company as an example. They ship 1.7m DVDs a day. Assuming each movie only has one DVD and a steady data flow, that requires a 740 Gbit/s connection! Maybe the movie industry will want to reconsider the Bittorrent protocol and the DivX codec. 
Vint then went off for a bit about turing online gaming in a potential video conferencing platform. An odd idea, but Joi Ito agrees, so maybe I’m the one who doesn’t ‘get it’.
Mobiles and Mobility
Back to mobiles. There are two billion of them out there, most are a general purpose programming device just waiting for the right software. They’ll stress the architecture of the net, I assume through sheer numbers if they all went online. They have plenty of potential uses. They’re already used for instant message (SMS) and as payment systems (also SMS), but combine them with geolocation services and they get a lot more interesting. They’ll create a great opportunity to monetize information that has been indexed with geographic information.
IMHO there’s so much scope of what you could do with just local search, but throw in a portable handset and geolocation and the sky’s the limit. It might be a bit sacrilegious to post Yahoo! stuff in the post but check out Checkmates and Event Browser, okay the latter isn’t for phones but imagine being able to use your phone to find something happening in your area, then get all your friends to meet up. And that’s just the tip of the iceberg, but enough about that, back to the talk.
New Business Models
I think this is the only point Vint talked about Google. They’ve have introduced new business models, the main one of course is making online advertising work. So much so Google is now known as an advertising company rather than a search company in certain circles.
With their APIs they’re trying to expand their advertising space rather than charging directly for the services. e.g. you have a database of information and make a Google Maps mashup, then place Google ads on the site. You get make a cool app fairly easily and Google gets its value from having more ads out there.
I’d be interested to know if Google’s main business plan is to simply create more places to put adverts, or if they plan to come up with different revenue streams.
Back to Basics
Here’s the bit for the PhD students. Vint brought up potential research areas of how he’d design the Internet if given the chance again today:
- End/end connectivity - Basically get rid of NAT
- Authentication - Get the end points of the network to authenticate themselves
- Flexible VPN membership - Make it part of the Internet rather than a bolt on
- Mobility (and spectrum sharing) - Consider a Boeing plane with Internet access, it’s basically a whole network that is getting connected to different points as if flies
- Confidentiality
- Early vs. Late Binding of DNS - Late binding is better for mobiles
- Broadband everywhere - Assume faster connections
Policy and Governance
I didn’t take too many notes during this bit. Something about a World Summit creating a working group, which then decided to hold a forum, and all the people involved were shocked that something like the Internet was under central control.
But it wasn’t all negative, he reckons with such a forum they might be able to start tackling problems like SPAM and fraud. Maybe.
Internet Enabled Appliances
More and more non-computers are connecting to the net. One particularly cool example was an Internet enabled photo frame that you can give to your family, then update the photos on a website. But as more devices get connected we will have to make the jump to IP6 just to have enough addresses for them all.
RFID also makes things a lot more interesting by letting things collobrate. e.g. your Internet enabled fridge can identify what you have, check a recipe online, then text you what things you need to buy. Or maybe just bypass you can order them from the supermarket itself.
We’ll see a shift in how we use the Internet from accessing it through one terminal to using multiple devices that collobrate.
InterPlanNetary Intenet (IPN)
An interesting project, but ultimately related to the Internet only by name. Currently things we send into space have custom networking hardware and protocols for efficiently reasons. The downside to this is that you cannot reuse equipment left in space for future missions.
Vint is working at the Jet Propulsion Labs on a standard networking protocol for things we send into space. So each mission means we’ll have another node in the network that can be used for things such as communication, sensor readings, etc. It looks like NASA is going to adopt this for their missions for the next 40 years.
Q&A
There was a brief Q&A session, nothing of note except for a book recommendation, The Singularity Is Near. Amazon describes it as:
Inventor and futurist Ray Kurzweil examines the next step in the evolutionary process of the union of human and machine. Kurzweil foresees the dawning of a new civilization where we will be able to transcend our biological limitations and amplify our creativity, combining our biological skills with the vastly greater capacity, speed and knowledge-sharing abilities of our creations. In practical terms, human ageing and illness will be reversed; pollution will be stopped and world hunger and poverty will be solved. There will be no clear distinction between human and machine, real reality and virtual reality. “The Singularity is Near” offers a view of the coming age that is both a dramatic culmination of centuries of technological ingenuity and a genuinely inspiring vision of our ultimate destiny.
I might pick up a copy, but I’ve been warned it’s quite long and goes into a lot of detail.
After the talk there was drinks and food. As I mentioned in an earlier post I didn’t get a chance to meet any one from Google but did catch up with a few people from Imperial. Cheers Google for hosting the night.
Here’s my friend Simon’s take on the evening, and here is another writeup of the same talk given at Imperial College early that day.
They filmed the talk but I haven’t been able to find the video online. If I find the link, I’ll post it here.
