A lot of gaming companies are going to the Megaserver model. In essence instead of individual servers hosting a separate instance of each zone, one zone is created for all players and a new one isn't made until the first one fills up. Dungeons are then available to all players from every server and finding a dungeon group is extremely easy by just going to the dungeon zone-in door and waiting til enough people show up to make a group. Other than the loss of revenue for transfer tokens can anyone think oof a reason why this would not work for EQ2? One possibility is the issues of having to make a seperate server for the 10 people (roughly) who still play open world pvp in this game but I can't think of any others.
This has been beaten to death so much that it borders on troll fodder. A quick google search on EQ2 and megaserver will get you the relevant threads. Short version is the existing architectural of the database won't allow it.
We have no idea what the team will do or what will be possible once EQ2 is transitioned from Oracle databases to EnterpriseDB this summer. Megaservers are a possibility. Server merges are a possibility. Although if all the heroic zones are going to be as heavily raid scripted and gear-dependent as they are in Altar of Malice, I don't think any amount of merges and dungeon finder is going to make PUGs viable.
Oh that is interesting, I thought they had moved off Oracle some time ago due to licensing cost. I remember some buzz about this in the dba community. Guess I was mis-informed.
I think it was just SOE as a whole and maybe not every game. http://www.enterprisedb.com/success-stories/sony-online-entertainment-replaces-oracle-enterprisedb Or they just published that years beforehand, lol.
Don't some of the servers have special things that other servers don't? That might be one reason. To me, it doesn't sound like a big deal, but then I am not a programmer, do not know anything about programming and don't want to know. I leave that to the hubby.
It would be costly for SOE to change a set of servers and adapt them to a megaserver structure in hardware and software. that's even before you factor in charater names. for a megaserver to work with names you'd have to change the naming system. there are only so many reasonable and non-silly names to pick chances are one you use is used by someone else on another server. for a megaserver to work with naming you'd have to set it up so each account can only pick 1-2 character names. you can make characters and when you select character to log in you select which name you wish to login as. soe wont do a megaserver for a game in alpha because it would turn out almost the same as the NGE fiacso.
It's along the same idea as if you have 6 boxes of paper files, and need to reduce them down to 1. You will likely have to change the way you file things, and you may also have to prune some of the information that you've been storing. Not an exact analogy by any means, and overlooks a lot of other problems that would be invovled in creating one big server.
lol, this is too funny. just look at all the angry posts in the CS forums when AB went BOOM over the holiday weekend and about the lag that ensued. Then few days later you have brilliant posts like these popping up. The servers dont have the capacity for such a thing and considering how short staffed SOE already is I would rather the team work on developing new content for the game rather than spending time and money getting a ridiculous mega server.
All new SOE games (FreeRealms, Planetside 2, EQNext/Landmark) are on EDB. EQ2 is migrating from Oracle to EDB this year. But it's not just a data migration. They're improving the design and schema so that characters are truly portable, the broker is less of a beast, and to give us cross-server dungeon running. After the change, the only difference between a character on Butcherblock and a character on Antonia Bayle will be a single piece of data called "serverid". Everything else will be completely agnostic. It's technical, but it's just a better design that solves problems and makes cool things possible. Unfortunately they'll never completely get rid of paid server transfers for two reasons: name conflicts, and the 2500SC per transfer they get.
Existing servers are slow BECAUSE of the janky 13 year old data architecture and inefficiencies that won't exist in EDB.
Seems like with the hack of Sony Pictures more open source stuff is not exactly what they should be aiming to implement.
Let's put two completely unrelated things together in a sentence and post it as a reply to a completely unrelated topic and see if it annoys at least one person. If so, mission accomplished?
Wow, that is great information. Was this at fanfaire this year? Not sure how I missed this. Thanks Feldon!
Pieced together plus my conversations with different SOE folks with regards to Census stuff. For example, EQ2 Census runs on MongoDB which is sort of an agnostic open source database that doesn't need to understand how the data it receives is natively stored. In contrast, DCUO and Planetside 2 census use EDB because the actual game servers do, and so it's just a replication of selected data. I wonder what, if any, change we'll see to Census with EQ2's EDB migration.
