- last post: 01.01.0001 12:00 AM PDT
Solution
If the host system is to be changed, the question then becomes "What do we do?".
The first, and obvious, idea is Central Servers. The servers would act as host and control the game. Players would always have the same connection, lag/latency would hopefully be minimised and "host advantage" would not exist. So, are servers feasible/viable? There is much debate on this and in my previous thread, I went to great lengths to prove that servers were indeed feasible. I will, again, prove this below.
The questions that need to be answered are how many servers do we need, how much will these cost and where do we get the money from.
How many servers are needed?
In the last 24 hours, there have been 563294 games logged. If you take an average game length of 10 mins, that's 3911 games being played at any one time. An opposer on the thread in Zanzibar has claimed that a server could not handle 1000 games. I'm going to be VERY fair on this and take the number of concurrent games per server to 500. This then would require 8 servers to host Halo 2 servers. In order to head off any stupid objections I'm going to multiply this by 20. So, we need 160 servers to host Halo 3 which I believe would be a massive over-estimation. I'm going to double this figure for redundancy (that means spares). So, we need 320 servers.
How much do servers cost?
At a cost of $100,000 per server, which is not unreasonable for a decent IBM i-series server we have a cost of servers of £32,000,000.
Where do we get the money from?
I have put together a large number of ideas below to fund the servers project. I am NOT saying that any one of these is the RIGHT way to go. I am not saying that any ONE item here could, and should, fund the whole project. I am saying that a COMBINATION of these ideas can all contribute to the project to make servers a reality. Any contribution from these, however small adds to the funding. If want to object to an idea, then do so. If you feel my maths is awry, feel free to re-calculate my numbers. Do NOT, however, dismiss servers as unfeasible by purely opposing ONE of these ideas.
1. Microsoft.
This first option is that Microsoft fund the whole project. It is in their best interests that Halo 3 does not flop. If H3 Multiplayer were to fail, it could potentially bring XBL down with it. Very soon, Sony will have their competitor to XBL up and running and the failure of XBL could tip people in favour of the Sony platform rather than Microsoft. The issue here is known as "Business Risk". Microsoft will have to do a "Business Risk" Assessment for H3 and try to minimise the threats to their product at launch. This Project will, at some point, come to the server issue. If MS feel that the risk is too great to allow the current host system to be used in H3, they will have to go down the server route. Although they would probably prefer not to, should the worst happen, they would lose a lot of ground on the competition. Re-engineering the product could take months. All the while, revenue is going to Sony and the PS3 is running away from the 360. In this case, MS could just invest in the servers as an Insurance against the possibility of the failure of XBL because H3 fails. In it's first year of sales, Halo 2 sold 7 million copies worldwide (income @ $50/copy = £350,000,000) and still continues to sell. 3 years after it's launch, Halo:CE was 9th in the sales charts, still generating revenue for Microsoft (as is H2 currently). Adding together the income from Halo:CE, Halo 2 and Halo 3 would amount to a staggering sum of money and I haven't even included any income from XBL (whose interests are closely linked with the success of H3). It is not asking a lot for MS to fund the entire project from their own pocket.
2. Users
User could pay more for XBL. If you were to calculate how much XBL currently costs you, I'll bet it would be less than ANY other leisure activity you could mention (that you have to pay for). To demonstrate this, I will use myself in a worked example. I will point out that in the early days, I played quite a bit. Since the first update, I have not played as much at all. In the first year, I have 80 pages of games at 25 per page. That's 2000 games. I paid £40 for the game and £40 for XBL. As I only play H2 online, I will attribute ALL of the XBL fee to H2. It has therefore cost me £80 for 2000 games. That's £0.04 per game in the first year. A game typically takes around 10mins to play. That's therefore 6 games an hour at £0.04 per game. Hourly cost is therefore £0.24.
Going to the pub at 2 pints of lager per hour = £5 per hour.
Going to the cinema at £10 per film (call it a 2 hour film which is over the top estimate) = £5 per hour.
Playing Badminton at the sports centre at £5.60 per court (4 players) = £1.40 per hour.
I could go on ALL day with this but I think it's safe to say that Bungie could charge you £160 for H3 and it would still work out as the cheapest form of leisure for even an average gamer like me. Anyone care to dispute those numbers?
On that basis, the extra £120 per user could be put into servers. In the last 24 hours, there have been 312742 unique accounts on XBL in H2. The additional revenue from those alone would be £37,529,040. Is £37M enough to buy your servers and bandwidth and maintain it? The additional revenue from the other players who weren't online in the last 24 hours can go into Bill Pocket for all I care.
If you object to this on the grounds that $160 for H3 would be a rip off, then perhaps we could split the game into 2 games. A H3 campaign game at $50 and a H3 Multiplayer Game at $110. I'd buy it!!!!
Please do NOT flame me for suggesting $160 as this is merely a demonstration of the "cost of living" impact on your pocket of Halo. Make the cost more, make the cost less, it's still a significant contribution to the servers project.
3. Authors.
See No. 1 but substitute Bungie for Microsoft.
4. Reduce Hardware cost by "Prestige" contracts i.e. Sponsorship.
Hardware companies could be "convinced" by MS that winning the supply contracts for the H3 project's hardware is a prestigious feather in their caps and they all "bid" low values to win the supply contract. I have personally built the servers that my Business runs on. The spec of these servers would probably have them retailing at £5k-£8k each. As it is, they cost only £2500 for both. This is with my (limited) buying power with the distributor. The hardware suppliers such as Dell, IBM, Hp etc... could all supply hardware at cost, cost + 1% etc... to win the contract. The lobbies, screenshots, blue screens etc... could then carry the manufacturers logos. Sponsorship happens everywhere. Companies like Vodafone, O2 and Emirates pay the likes of Manchester United and Chelsea tens of millions just to put their logos on their shirts and a few advertising boards.
5. Advertising.
As you log in, during the lobbies (pre-game and post), during the matchmaking screens, Companies can advertise their products. Anything can be advertised, Mortgages, Cars, Kettles, Games, ISPs, Routers, Fizzy Drinks, Clothes and so on. The cost of these adverts pays for the servers. This is not ideal for the gamers but hey, so what! Successful websites are already doing this. Although this option may not be enough to fully cover the cost, it could, nevertheless, contribute to the cost. You may not want adverts but at the end of the day, IT'S NOT UP TO YOU!
6. Usage basis.
Each game you play, the cost gets charged to your credit card. I don't need to go on about this one too much but in the last 24 hours 669,255 have been logged. Games per year (on this figure) would be 244,278,075 games per year. At £0.10 per game, the income in a year would be £24,427,807.50! Is this enough for your servers? I'd be happy to play a game at a cost of £0.10 per game, 6 games per hour, £0.60 per hour.
7. Limit certain playlists to server to reduce size of project.
Perhaps only certain playlists could be server based to cut down the traffic/cost. These could 4+ team games only (Team Slayer, BTB, 6v6 etc...). It could be just the ranked playlists that are server hosted and custom games are hosted by an xbox in the current manner, cheating being rare and irrelevant on Customs.
8. Premium Playlists.
Perhaps there could be 2 sets of playlists, the xbox hosted ones and the server hosted ones. You have to have a higher level of membership to play on the servers. Perhaps you pay as you play (see 6).
9. Don't use expensive servers at all, use xboxes!
I have shown that at any one time, there are around 4000 games being played. Each on is being hosted by an xbox somewhere in the world. The problem with this is that someone (other than Bungie) is controlling that Xbox. How much would it cost Bungie to provide several thousand xboxes to act as hosts? A re-write of the matchmaking code and hey-presto SERVERS!!!!!! I can buy an xbox 360 for £220 (trade price). Let's assume, after removing several mark-ups that an xbox would cost Bungie £150 (it will probably be MUCH lower but I'm erring on the expesive side). For launch day, let's say 80,000 xboxes. At £150/box = £12million. That's MUCH less than the IBM's. As demand for H3 MP dies down, these xboxes could be sold cheaply, give to charity etc... If demand increases, add more xboxes. They could EASILY be distributed around the world to provide "regional server farms". If one breaks down, SO WHAT, they're cheap, replace it. My numbers on this are based on hosting EVERY game on XBL, not simply the ranked playlists etc... Using xboxes could be a fantastic solution for everyone.