Throw a hand grenade into the lobby of the office of Google Ireland tonight and you could have blown every software project schedule in Dublin by six months. The lifeblood of the irish software industry, its young geeks, were out in force to see the father of the Internet, Vint Cerf and to discover whether all the rumours of the Nirvana inside Google were true. We had the full complement of pony-tails, anoraks , tee shirts and beards. Pity the two poor saps who turned up in pinstrip suits.
Google put on a full court press, with everything from free wine (Faustino VII red, and an unidentified white) to the legendary Google food (the samosas were excellent). It was in faith, a recruitment night for Google and Vint played his part like the true gentleman he obviously is.
At 6.55pm Vint took the stage and we all settled into our pale blue chairs (so web 2.0) to listen to the great man speak. He is an easy speaker who is very comfortable in front of an audience. He rolled through the history of the Internet and peppered it with anecdotes which although some may have heard before carried extra weight due to the speakers history.
His discussion ranged widely over a number of issues facing the Internet in the 21st century, what follows is my precis of his talk. All errors are my own.
IP destroys the telephony business: He has little time for the Telcos who scorned packet switching and still have no sensible standard protocols that allow application level interoperability. He stressed that the IP protocol just moved packets and this is the flexability that allowed WIFI, VOIP, FTP, SMTP etc. to flourish.
Presence in the dial tone of the 21st century: Presence is our ability to be contacted by others. In the 21st century Presence will be controlled by the user some will get email, some will get messenging, some video, some voice.
Mobiles creates issues for the Internet: With mobile devices initiation is easy, but discovery is a problem.
IPV6 fixes some of this and Boeing have done wonderful work putting networks into Aeroplanes.
Convergence: Opportunities for applications that converge GPS, mobile internet and location based data. For example, I am driving, I ask (using a microphone) for the nearest ATM. Voice recognition software converts this to a command, searchs a location based service for an ATM giving it my coordinates, gets the coordinates for the ATM back and uses the navigation system to guide me there.
Broadband Symmetry: As we create more data than we download (pictures, video etc.) the need for symmetry becomes more urgent, ADSL won’t cut it in the 21st century internet.
Governance: Policy is much harder than technology, legislating for privacy, security, cross border domain adminstration
Governments like to have somebody in charge and are distrustful of distributed management or decentralised control.
IPN, InterPlanetary Network: The IPN is move by NASA to standardise the network protocol used in all of its space vehicles and satellites. To date each space mission has used a purpose built protocol designed for a specific payload. With a standard protocol so they can communicate over long lifetimes (for outer planets, radioisotope power sources are used that last for many years). Problems are long delays between transmission (high latency) and slow links (low bandwidth). The new protocol will be “mail like”.
Turns out this protocol is also very suitable for mobile networks.
Vint then took questions from the floor:
What is the best research environment?
Well it took 22 years of government funding to development the internet. Who in the private sector is likely to make that investment. The Internet pioneers worked under extreme constraints. Sometimes the formulation of the problem holds the solution.
Will mobile devices supplant the Wired Network?
All networks generate high core traffic, its unlikely that mobile networks can support this core intensity in the near future. Wires will still exist at the core.
Will the scarcity of IPv4 addresses generate a new market the way domain names did in 2000-2002?
Commercialisation generates loopholes. He hoped this wouldn’t occur. He described a process I hadn’t heard of called “Domain Tasting” where a company waits for a name to become available and then puts up a page and measures who visits the page over the first 5 days of operation. If they get lots of hits they keep the name and fill the page with advertising. If they get few hits they abandon the name at no cost, because under the current protocols, you can return a domain name within five days of registration at no charge.
How is Google Different?
70,20, 10 rule: 70% of work is allocated to a specific project, 20% is dedicated to work related to project and 10% to any project you like.
Google is focussed on small teams, very little hierarchy, high bandwidth communication and is very project oriented.
What is your opinion on software patents?
Vint and his co-worjers didn’t patent Internet work and TCP/IP to encourage vendors to integrate. Thinks patents are problematic but companies need to patent defensively so the other guy doesn’t patent the same idea.
A fascinating talk all in all from a Man who work will impact all our lives for many years to come.