Quote:
1:- Is it feasible while keeping a single hosted physical server and shared bandwidth to have 4 separate "shards", meaning by shard an independent "Atys instance"?
Not with the current architecture, no; we need 8-10 big physical servers to host one full shard.
Quote:2:- we don't know the details on what was implied by the pre-ring shard unification process. Is it tied to the first question? What changed, what's the relationship between the terms shard/server/service/UNIX process/etc...
- A shard is a set of 10 physical servers
- A server can contain one or several services
- A service is made of one process
The Ring didn't changed that. The only thing is that Ring sessions, from all shards, are hosted on dedicated servers, so people from all shards can connect to the same sessions, and shards communicate, for example for the tells.
Quote:3:- If the client/server architecture is done in such a way -- at least at the beginning -- that it seems non trivial to have separate instances (e.g. fixed-static vs dynamic service ports, unable to launch several instances of the same process, etc), is a "virtualization" approach possible? considered?
No; for the simple reason that the servers just haven't enough resources to host several shards.
Quote:4:- Some posts in the ORG project, referring to "multilanguage" seem to imply a "Shard merge". Is this correct? (see question 1) or by the evolution from Jolt hosted towards a more economical solution the shards will be kept
separated?
Shard merge is now only considered for the case where the user base would drop substantially. Currently the shards have 300 to 700 simultaneous players at one time, which is enough to have the feeling the shard isn't empty. When a shard goes down to 100-200 players, it isn't good anymore.
Quote:- Cho being conceived as a fresh start, with trading and character transfer limitation, suddenly becoming a shard as the others
I don't know the answer here, but I guess this was more a "uber player" limitation at the start than a technical reason.
Quote:- Scalability issues, such as number of players per server, number of ring instances, etc.
Again, I don't know the answer. But from what I know, I think this was more related to Nevrax *expecting* scalability issues, and the fact that it is good publicity that you need to open new shards.
Quote:- What will be the policy if the ORG server gets "too crowded" to make the server unplayable? how will you manage opening new servers and transferring chars?
I think this would have to be discussed with the overcrowded community then. There is no perfect solution in this kind of case. Splitting a server in two, closing subscriptions in the first one, allowing voluntary transfers are all likely options. None of them is perfect, and this would have to be voted.
Quote
ifferent server status (e.g. built temples, economies, accumulated currency and wealth)
Except for the temples, these were already managed when there was to merge Windermeer to Artispotle; so there is no reason why it shouldn't be done again. And, again also, this could be discussed among the community.
Quote:- Multilanguage clashes, specially in official events, lore articles, Customer Service ("Hello, guide Gobs at your service!..uhm, wait a second, I need to forward your ticket to a German speaking Guide... hold! " )
True, it wouldn't be easy, and there would be side-effects. Possible solutions include having language-specific events, or integrating the different languages directly into the background.
But, remember something important: this would be a community-driven project. So it means nobody should expect a Godly Company to bring the perfect answer, et voilà! On issues like that one, I know a lot of people would have interesting suggestions. That's even what *is* interesting in this project; finding solutions together, instead of relying on one or two decision makers, always in a hurry - they invariably overlook a large part of the possibilities, which a community doesn't.
Quote:6:- Can you elaborate on these issues? is it wrong to assume that the answer to these questions will be "uhm, yes, shard will be merged and although we are aware of these issues, we will deal with them one at a time... till , and we'll see.... this is what we can offer, leave or take it"...
See my previous answer.
http://www.ryzom.org/forums/images/smiles/icon_smile.gif There is no "we" and "you" here. This is not a commercial company project. This is a Free Software, community driven project.
Quote:7:- Is it unreasonable to think that a random player X may prefer a buyer that, without GPLing the code and with less open/community driven model can stick to a more classical MMRPG service/business model and assure a seamless continuity with 4 different shards and a professional CSR service that covers 24/7?. Please elaborate
The Ryzom.org is more risky, true. Something like this has never been done. I personally think this could work, but you can think otherwise. Until it has been tried, no one will know for sure.
_________________
/Aprak on behalf of Ryzom.org community
)
