Multisite Scaling Strategy

[goldkey]

I'm wondering if anyone is willing to provide some input and suggestions, as we're wrestling with some strategy decisions.

We are launching a WP multisite system. There won't be any community features or other interaction among sites. Each user's sites will be individual and stand alone.

With this in mind, we're debating three strategies:

1) ONE GIANT MULTISITE (set up a big, scalable installation and let it grow)

Advantages:

• Only updating one set up plugins, themes, etc.
• Only paying for one LiteSpeed license, etc.

Disadvantages:

• Complex, expensive setup (especially since it will take a while to build)
• All eggs in one basket...if one server goes down or gets hacked, all clients are affected

2) INDIVIDUAL SERVERS (individual servers, each hosting a set number of sites...when one server reaches the set resource level, we just add another server to the cluster).

Advantages:

• Not putting all eggs in one basket
• Modular approach
• Ability to offer clients a choice of different geographic server locations

Disadvantages:

• More maintenance required (although possibly offset by not having to maintain complex giant multisite)
• Paying for more LiteSpeed licenses, etc. (however we'll use large servers to minimize this)
• Need to update multiple servers when themes and plugins updated (not a terribly big deal)

3) MULTIPLE SERVER CLUSTERS (a few smaller clustered multisite setups, maybe spanning 3-4 servers each)

Advantages:

• Maintaining fewer systems than in strategy #2 (although managing this might be more complex overall)
• Paying for fewer LiteSpeed licenses, etc., than strategy #2

Disadvantages:

• Fairly complex, expensive setups
• Still putting all eggs in a few big baskets

Also, no matter what, it is always best to have the database(s) on a separate server, or is that only optimal after the installation reaches a certain size?

Any feedback would be appreciated!

Thanks,

Mark