Today we’re announcing that we’ve implemented full support for an extremely requested system that we believe the FFXI private server community has needed for a very long time — something many people in the private server space have treated as impossible or impractical.
We’ve been hard at work for a little under a month doing all sorts of backend work to rewrite our core services from scratch so this functionality is truly a first-class feature of what our platform brings to the table.
We believe this will be server-defining! And to be honest, we hope it sets the bar high for what is possible moving forward.
Your Run. Your Group. Your Shard.
What exactly are “Private Shards” you might ask?
If you’ve played other popular MMOs before, you’ve probably seen dungeons or raids that are instanced. Typically, you join with a group of friends and do your best to beat what’s in front of you.
That is exactly what “Private Shards” are; privately instanced content that is specific to & available for you and your friends, on-demand.
That’s cool but….. FFXI is all about group gameplay. Why would you want something like this?
With private shards fully implemented on CalamityXI we have the creative liberty to break free of all the existing problems that we believe have plagued FFXI for so long.
The biggest offender is probably something everyone is very familiar with already: Dynamis.
If you’ve played FFXI before, you’ve probably done Dynamis. And, if you’ve done Dynamis, you’ve probably been a part of a linkshell with scheduled event nights. If you’ve played on popular servers before, you know exactly where I’m going with this.
Ever had your Dynamis night completely ruined because some other group swings in and enters right before you?
Well…
Those days are over.
(cold-start shard boot time down to 3 seconds!)
Sure… I’ll believe it when I see it.
We knew you’d say that.
We got you covered.
The video demonstrates how private shards are allocated on-demand, in real time, as you trade your timeless hourglass for a perpetual hourglass. It also shows that core gameplay features like chat and search continue to work across private shards.
The video is intentionally longer than usual. We wanted to show one continuous stream of everything working from a local dev environment in 4K, so you can follow everything down to the pixel.
What do you plan on doing with this?
We have tons of ideas.
Because we can spin up new private shards in a matter of seconds, and because they don’t conflict with public shards, we’ve unlocked a massive amount of creative freedom:
- Imagine trading a pop item to an NPC outside Dragon’s Aery and having your group warped to a private version of Dragon’s Aery with no mobs — just Fafnir waiting in the middle of the room.
- Imagine working together with other players to complete a zone-wide event in an alternate dimension.
- Imagine custom fights that act like BCNMs, but on any battlefield, in any zone, at any location.
- The list goes on.
Technical Features
Obviously, this feature took a lot of work to implement.
We run a fleet of highly available servers that provide compute capacity to deliver this feature. All shards, both private & public, can be scheduled across compute hosts dynamically and require a lot of supporting systems to work correctly:
- dynamic port allocations & firewall rules
- database capacity & SSD-backed storage
- fast, low-latency networking
- elastic services for load balancing
- shard metrics & health endpoints
Private shards can have whatever lifecycle we want them to have. They can last for an hour, 24 hours, a weekend — it doesn’t matter.
There’s no hard-coded limit to the number of private shards we can run. Limits are obviously based on available compute, and if we need more capacity, we can rack another machine with 0 downtime. The custom operator we wrote takes care of the rest for us automatically.
Are you going to contribute this back to the community?
We would love to.
However, we believe that LSB is meant to be operated on any hardware and it does an amazing job at making this possible.
To make all of this work the way we have, we had to do a lot of specific infrastructure work. That led to very custom changes that probably won’t land directly in LSB because they are architecturally different from what LSB aims to achieve.
We rewrote our connect-server, search-server, and world-server components. No more fragile single-instance core services.
Everything is highly available and built for scale, allowing us to load balance logins, search, and shard communication as demand grows.
Future Work
We’re hard at work building core infrastructure & content that we’re proud of.
We know from our last post you don’t like the “Stay Tuned” answer, but hard work takes time and getting the foundation right matters.
Keep checking in. We have a lot planned and there’s much more actively in the works!