Its nuts. People are talking about whether we have too much bandwidth to the home?
The blogosphere is in a titter over a recent post from Om Malik where he makes the argument that recent upgrades in last mile bandwidth are currently all for naught.
He’s overlooking elasticity and the classic evolution cycle of technology, particularly computing. I see no reason why it does not apply to networking either. It’s what I like to call the Industrial Accident model.
Few people ever see Industrial Accidents coming, but after the fact lots of people like to pretend they did. A new service or device appears that ‘wastes’ these new resources and changes the status quo. It’s not an accident in the sense it was unplanned. Industrial accidents are the result of an unrelated sequence of events aligned by an individual or group who is unaware of the impact their work will have. Don’t be thinking “Watson! Come here, I need you.” - And think more along the lines of The Terminator and Skynet with a happy ending.
Stage One - Stretch the Elastic
Often when we see or envision a big quantum improvement in computing, the performance exceeds what most people think is necesary. You’ll know when you’re there because you hear the following:
- Why would I want to run Doom at 100 frames per second?
- I don’t need a cable modem to download email!
- Why would anyone need a hard drive that small?
Or more recently…
- 10 Gigabit Ethernet? I can’t even use 1 Gigabit Ethernet!
- How many people really need a 60″ Plasma TV?
- Why would I want a computer in my living room!
- Or you see post’s like Mr. Malik’s
And most famous of all….
“640k is enough memory for anybody.”
Bill Gates, 1981.
This is where we are right now with FTTH.
Stage Two - The Industrial Accident
Meanwhile, unknown to most if not all, someone somewhere is about to create a big mess. Napster capitalized on fast broadband connections in Dorms and DSL/Cable modem penetration. The Web capitalized on the shift in universities from terminals to PC’s and took a geeky service called Gopher/Lynx and made it graphical. The MP3 player took a big capacity laptop drive and got rid of the CPU. Apple wasn’t the first to do it; they just were the first to make it sexy and easy to use.
Stage Three - Breaking the Elastic
Things go upside down. Suddenly a resource that was in abundance is now at a premium. The web showed up and people couldn’t download graphical pages fast enough at 56k. New games stress out GPU’s before they even reach mainstream deployment. Napster for MP3’s? Yeah, well that’s great but it takes days for me to steal share a DVD over my 768k connection!
Often, other related things get broken too. Once the web took off you got network switches that were content aware to do web server load balancing, etc. As in real life, industrial accidents create a big mess in multiple areas that, lucky for us, our free-market system is more than happy to clean up.
Conclusion
WAN Bandwidth isn’t any different than RAM, Flash, Disk Space, LAN bandwidth, CPU cycles, etc. You put a fatter pipe into a house for around the same cost level and guaranteed, someone, will fill it. Who the hell knows with what, but someone somewhere will create the industrial accident that makes it happen.
…. and yes I do think Google is the next Skynet.
The naysayers don’t get it.
Just as mainstream cable allowed the proliferation of “Wayne’s World” type enterprises, so will FTTH enable the next generation of grassroots media content creation and delivery. These fast pipes (especially the promise of symmetrical 100mbs) will allow college students, independent filmmakers and who-knows-who-else to distribute their works to a much broader audience. We’ll eventually see a new type of network, one that more accurately fits the definition of the word.
This is just one example. Remote collaboration will also finally be practical. Erc etc etc.
I’m glad to see SOMEone gets it. I enjoy your writing; keep it up!