Re posting my comment from the discussion thread prior to the patch going live because clearly it is still relevant (sorry being lazy and don't feel like restating the same things all over again):
They need to implement a hard cap additional to prevent additional load after a certain threshold. Sure overflow instances could theoretically be added on top of such a system, whether they do or not isn't the point here. The point is that right now there is nothing at all to prevent all 3000-7000 people in a server from trying to enter a zone and crashing the world. And this has already more or less happened due to the fates. Therefore they from an infrastructure standpoint need this zone limit, right now. They can't not to do it. It is the only sane thing to do right now.
It also is unlikely to cause problems except during odin and behemoth, and then only in those zones: so it is rarely going to effect people, and only for 20 minutes or so even then. Much better than the alternative where behemoth pops and the servers crash, and nobody can play for hours.
This is why it is a hotfix. Not doing it would be a horrendous mistake. Also, fyi hard caps are a key part of any overflow system also. That is the point of an overflow system: when you reach cap, you overflow. Whether SE does the overflow part later or not is irrelevant to them needing the cap. They need it regardless of what does or does not happen later.
They all do. And it is always necessary. But let me explain in broad generalities:
There is a limited amount of processing any single server can handle. After that point you have to do one of two things
1) Split servers by task aka instance
2) Scale outwards with clustering
Both are very very different design decisions, and both are fraught with different types of technical problems.
A pure open world will go with option 2, and will not have a hard cap per instance, but rather a hard cap on the entire world.
An instanced world with option 1 must have hard caps per instance because each represents a different server (or virtual server but the point remains).
You haven't seen this in other MMO's that go with clustered worlds where there is a world wide cap. But the hard cap is still there
In those cases they either use a pure queue system that prevents people from logging on at all when cap is hit, or they use an overflow server where people are thrown at a different set of servers. But in both cases you still have a hard cap.
This is not something unique to this game, it is something anyone who deals with large scale systems has to cope with in some fashion or another. The only difference here is that for some reason they didn't implement it from day one, and unlike other instanced games released recently, it doesn't have an overflow system to deal with what happens after the cap is reached.
I also am not sure why you are so hostile to this, given you won't even see it except when behemoth or odin pops. Unless you anticipate that 500 trolls will all get together to troll an entire world, which I find rather unlikely.
You seem to think that implementing a hard cap "instead" of channels is bad.
What you clearly didn't catch from my above post is that a hard cap is part of a channel system. If they implemented a channel system, the very first thing they would have to do is implement a hard cap. This isn't "old" technology, or "lazy". It is fundamental fact of any load management design. There is no way around it.
Now clearly you think that they should have implemented channels/overflow servers at the same time. That is great. But they didn't have the infrastructure in place at all (from a development standpoint I was kind of, well, appalled, that they didn't have a hard cap from the get go). Given the nature of this system, my best guess development estimate is 3 months to add and test, and add additional servers etc, to implement channels. That is great and all, if they plan on it. But it is hardly going to happen overnight, much less as part of a hotfix to prevent world crashing issues.
Should they have implemented a hard cap from day one? Yes. That is server load management 101. They didn't, therefore they had to. It is literally that simple.
Also, you clearly misunderstood what I was saying about instance servers vs clustered servers: world cap is not zone cap, because each zone is its own server. World cap is based on distributed load across all servers in the world. Each world is 30 servers. Each server has its own max cap which based on the last time a behemoth overload crashed a world, this maximum possible threshold is something around 750 per server. Given my own experience with servers, that makes sense. I currently work with high end servers (last replaced in 2012) that max out at 5000 connections due to processing overhead, the overhead in this game is much higher due to constant positional updating and enemy management. In order for World cap to be instance cap, they would have to cap each world to the max possible in each zone. That is simply not feasible from a physical server management standpoint, and also would make population density very low causing additional problems in game.