- Joined
- Nov 14, 2012
In 2015 I'm going to start releasing a monthly PDF explaining the financial situation of the board. Right now, the financial situation is "there is none". Kiwi is paid for until, like, the middle of 2016. Orange is $10 a month.
However, I'm very certain that, this month, Jace will want me to start selling T-Shirts. If that happens, there's going to be some drama. Namely, more ethic quandaries. "Is this moral? Is profiteering off Jace acceptable?". To deal with this, I've decided Jace will get about 25%~30% in royalties. Whatever doesn't go towards taxes (and paying the Chinese) will go towards the server costs. I won't be scraping anything out of the pot, and if I did, I'd let people know in advance.
And don't worry, if that happens. I'll have someone design a cool, non-neckbeardish general purpose graphical T-Shirt and probably a Hoodie. It'll just look like any article of clothing and won't have any ebin maymays hidden on it.
What weighs on my mind right now is growth. In short, we've grown a lot. Since March we've gone from 300 unique visitors a day to 1,400 visitors everyday, or about 900 new unique sign ins a week. I do not have enough fancy text decorations to describe this so here's a graph with numbers.
Right now, at prime time, we get 350 active users at one time. That is a number we previously only saw after the housefire or during the trial.
I also want to point out that this is unique sign-ins, and not guests. When we talk about page hits and sessions, it looks like this. This isn't 100% reflective, it doesn't include guests below the age of 18, people not running JavaScript (esp. Mozilla users with ScriptBlock), or anyone opted out of Google Analytics.
When Tyce and Jace did a wombo-combo stream on that Sunday, we had 440 concurrent users. This shit on the server. I've since tuned our Apache and MySQL installations to be more greedy, but regardless, that is the first time (without the assistance of a 500 bot DDoS attack) we've seen the server slow down or halt due to traffic. We're approaching an event horizon where Kiwi alone will not be able to support the forum.
This is especially true if you look at my projected traffic for 2015.
And that's just sustained growth, what the forum has been pulling in by just being a forum. With editors like @Jaimas chugging away at http://lolcow.wiki articles and the people we have working on that producing consolidated, high-quality content not seen anywhere else on the Internet, I predict a spike. Not just a spike, I really think it'll be a big deal. We'll see, not just more people, but a fuck of a lot more people.
The next step, I feel, will be to upgrade laterally. The way a server works really has two parts. MySQL is a database. It stores information. Apache is a request handler, it turns http://kiwifarms.net into a content response using a combination of what information you've supplied, requested, and what's in the database. By splitting these two entities apart, it achieves at least a few things:
1. The server has more resources and space to work with.
2. Installations are cleaner, so servers have less shit going on as it caters to one sort of process.
3. You can continue to upgrade laterally with LoadBalancing.
LoadBalancing, in short, is another box that looks at your servers. If, for instance, we grow so much we need a third server to handle a second installation of Apache, it can look at both and say "Server A is busy, so this request goes to Server B". This allows for a huge number of servers to run off one database. A loadbalancer can do this with any number of options to pick from and is how massive companies deal with their traffic.
The way Linode does pricing, CPUs become less cost-effective at the 4 core mark. The next step up for Kiwi would be the 8GB package, which is the first package to have fewer cores than gigs of RAM. Check it out:
That means that, building laterally is also the more cost-effective way to handling future upgrades, considering (atm) CPU is becoming our biggest bottleneck.
What to expect in 2015:
That's all I can think of now. I'm feeling a bit anxious and wanted to go on a text-wall adventure. Let me know if you have any thoughts.
However, I'm very certain that, this month, Jace will want me to start selling T-Shirts. If that happens, there's going to be some drama. Namely, more ethic quandaries. "Is this moral? Is profiteering off Jace acceptable?". To deal with this, I've decided Jace will get about 25%~30% in royalties. Whatever doesn't go towards taxes (and paying the Chinese) will go towards the server costs. I won't be scraping anything out of the pot, and if I did, I'd let people know in advance.
And don't worry, if that happens. I'll have someone design a cool, non-neckbeardish general purpose graphical T-Shirt and probably a Hoodie. It'll just look like any article of clothing and won't have any ebin maymays hidden on it.
What weighs on my mind right now is growth. In short, we've grown a lot. Since March we've gone from 300 unique visitors a day to 1,400 visitors everyday, or about 900 new unique sign ins a week. I do not have enough fancy text decorations to describe this so here's a graph with numbers.
Right now, at prime time, we get 350 active users at one time. That is a number we previously only saw after the housefire or during the trial.
I also want to point out that this is unique sign-ins, and not guests. When we talk about page hits and sessions, it looks like this. This isn't 100% reflective, it doesn't include guests below the age of 18, people not running JavaScript (esp. Mozilla users with ScriptBlock), or anyone opted out of Google Analytics.
When Tyce and Jace did a wombo-combo stream on that Sunday, we had 440 concurrent users. This shit on the server. I've since tuned our Apache and MySQL installations to be more greedy, but regardless, that is the first time (without the assistance of a 500 bot DDoS attack) we've seen the server slow down or halt due to traffic. We're approaching an event horizon where Kiwi alone will not be able to support the forum.
This is especially true if you look at my projected traffic for 2015.
And that's just sustained growth, what the forum has been pulling in by just being a forum. With editors like @Jaimas chugging away at http://lolcow.wiki articles and the people we have working on that producing consolidated, high-quality content not seen anywhere else on the Internet, I predict a spike. Not just a spike, I really think it'll be a big deal. We'll see, not just more people, but a fuck of a lot more people.
The next step, I feel, will be to upgrade laterally. The way a server works really has two parts. MySQL is a database. It stores information. Apache is a request handler, it turns http://kiwifarms.net into a content response using a combination of what information you've supplied, requested, and what's in the database. By splitting these two entities apart, it achieves at least a few things:
1. The server has more resources and space to work with.
2. Installations are cleaner, so servers have less shit going on as it caters to one sort of process.
3. You can continue to upgrade laterally with LoadBalancing.
LoadBalancing, in short, is another box that looks at your servers. If, for instance, we grow so much we need a third server to handle a second installation of Apache, it can look at both and say "Server A is busy, so this request goes to Server B". This allows for a huge number of servers to run off one database. A loadbalancer can do this with any number of options to pick from and is how massive companies deal with their traffic.
The way Linode does pricing, CPUs become less cost-effective at the 4 core mark. The next step up for Kiwi would be the 8GB package, which is the first package to have fewer cores than gigs of RAM. Check it out:
That means that, building laterally is also the more cost-effective way to handling future upgrades, considering (atm) CPU is becoming our biggest bottleneck.
What to expect in 2015:
- http://kiwifarms.net becomes official. The old domain is downgraded to CWC only boards to encourage moving.
- http://lolcow.wiki launches. I'll mod the forum so that you have to have an account here to edit the wiki.
- Lemon, our third server, onlines with the possibility of Lime being added as a load balancer node in the future.
- A monthly PDF with a full, detailed explanation of money going in and out.
- Drama! :^) and maybe T-Shirts
That's all I can think of now. I'm feeling a bit anxious and wanted to go on a text-wall adventure. Let me know if you have any thoughts.
Last edited:
