How about more control over the software & configuration side of ELB as well? If you could control more things that nginx or haproxy let you control, I think there would be a lot less need for another routing mechanism behind ELB.
In addition to weighting, we need an nginx layer for different app pools, custom routing options, max connections limits, request queueing, url rewriting, static content serving for specific requests. The list goes on but these things could easily be brought up into the ELB layer.
Actually that a great thing. Shows to hackers that even with billions of dollars, you can still make one stupid mistake that ruins your website and reputation.
Someone that will look on this on the business side could say: if VW isn't meticulous on their website, where else do they make mistakes?
In theory, i like the idea but I can't see paying for an API like a service would generate as much money for them in the long run as their current ads strategy. API Services tend to cap out at a certain rate so then you need to start offering unlimited packages, etc. Companies are not going to be paying millions of dollars for API services, I cant think of one example where this is remotely valid.
Have you thought about how Twitter would actually structure something like that? Twitter is monetizing FAR FAR better than any of these other small Twitter API clients. Something like Tweetdeck might have generated hundreds of thousands of dollars, but thats a joke compared to the kind of money Twitter is dealing with.
You'd be surprised, "scaling" a database is something that more and more folks are having to do lately, with the onslaught of mobile apps and facebook apps and web apps, it's just a whole lot easier to create an application that gets hundreds of thousands of users - millions of users.
And more importantly, you can have an application that has hundreds of thousands of users and not really make a ton of money, so you kinda need something that is easy to use because you don't have the capital
Although I agree with the Facebook App sucking (and seeming to get worse and worse), I do feel like the article is a bit misguided. I don't give 100% blame to UIWebView and no-Nitro simply because the entire way Facebook has developed the mobile app has been hacky and misguided.
The code itself is poorly designed, the API's are all over the place, there are inconsistencies everywhere and the performance is clearly lacking. On the flip side, take a loop at the LinkedIn app for iPhone. It's super slick, easy to use but still has a ton of features and the performance is far better. Yes, it's not as slick as Path but its miles ahead of the Facebook app.
So if I were Facebook, I'd rely a bit less on the 100% platform agnostic approach, take it back a bit and build the things that make sense cross-platform with HTML and build the parts separately that make sense to build using native code. This approach to me accomplishes much of the same cross platform success without creating such a crappy and laggy app.
I don't think anyone is disputing the fact that guys in coffeeshops can't make successful startups. I don't think that is the point the author is trying to make. The ecosystem, much like the rest of the world, is centered on averages and unfortunately, it becomes a LOT easier to start a company and not FAIL when you have something to leverage. In the startup world, there is the deadpool and avoiding the deadpool is very unlikely when you have a plethora of experience to play.
Everyone starts with nothing and people succeed based on merit, there is nobody who is creating a shitty product and succeeding just because they are Kevin Rose. BUT, Kevin Rose can start a company and get millions in funding with the snap of his fingers and average Joe has to work and work and grind and grind before that happens. And probably, there are average Joe's out there that do all of this and still don't make it because of some other reason, that may even be out of Joe's control.
So conclusion is that success in the startup world is still about making a great product and building a good business but getting funding in my opinion is just as equally about product as it is about selling the founders to the investors and if you have something good to sell, like 'I created Flickr', it's incredibly easy to raise money. And one could argue that with money comes another level of huge advantage, and the sooner you raise money, the more advantage you have. SO in the end, having experience is a huge differentiator when looking at averages across the board.
In addition to weighting, we need an nginx layer for different app pools, custom routing options, max connections limits, request queueing, url rewriting, static content serving for specific requests. The list goes on but these things could easily be brought up into the ELB layer.