I’ve been a manager for 2.5 years and I’ve been too long away from programming. There is something just so wonderful about being able to work again in a world where there is a right and a wrong.1
I decided to start to finally2 teach myself iOS development today for: first, because I’ve never done it before and second, because it’s an opportunity to learn a new language and re-learn an old one3 I haven’t done for over a decade.
We’ll see how it goes. I’m not optimistic.
At my first startup, I had a CTO, who was a year younger than me, used to tell me stories of his time at UUNET and what whiney little cheap turds most of the first generation dotCom billionaires were before their tiny little ISP’s sold for hundreds of millions.
This story isn’t about that.
As a Korean, he was asked to set up the asian version of UUnet. Back then Korea’s networking should have been trivial as its a peninsula and most of the population lives in or around a single city, Seoul. Instead it was terrible because none of the Korean networking companies would set up peering points with each other. “See,” he told me, “When you need to connect two networks to each other, you have to set these up, and while they’re free to run, they’re costly to build. How do you decide who pays for the point?”
You might think 50-50 is fair and propose that. If you’re a new player in this space, since you’re the one asking to set up these points, you probably benefit more than network you are trying to pair up with so they expect you to pay all or most of it. You get nowhere with this offer. This is what happened to him.
Korea wasn’t the powerhouse Internet trendsetter they are today. I remember, back then it was often faster to connect to a US site then to another Korean one because your packets to the latter would have to go through peering in the US since there wasn’t any in Korea. In other words, to talk to your neighbor, your message would be sent from Korea, across the Pacific to the US, and then back across the Pacific to Korea. This was the Korean Internet in the early-mid 90’s, all because of a lack of peering there.
On a lark, a friend of his wanted to set up a BBS or something and he dropped a server on his network for his friend to use for free.
Then one day, months later, companies that had previously refused to peer up were contacting him to peer up. As I’ve outlined just now, being the asker is a huge concession in the negotiation. He wondered what changed and looked at his traffic. It turned out that his friend’s BBS or something had become hugely popular in Korea and a lot of that network’s customers were demanding better service to it.
…
The lesson here is if you have what the other party wants, you hold the cards and the other side has to make the concession.
I bring this story up when I have to explain a basic business principle to others to understand what is going on.
Want to understand why Office is finally out on iOS? Look no further than the peering story.