Honestly every year there is one popular sever and another that is a ghost town. Just make two severs identical and spread the load. Clearly after all these tlp launches its the same problem that results in upset people.
I don't think they have the resources to do this. Supposedly they have a very limited budget. Plus with the added stress of the coronavirus they aren't as concerned.
I already said it. They don't have a dedicated test environment so the first week(s) of a merge or TLP should be 1) free, 2) beta (lower expectations and expect issues) and 3) with dedicated forums to vent and report issues... until everything stabilises.
That totally makes sense. I know they do not have a dedicated test environment, but perhaps opening up the servers 2-3 weeks for a test run wouldn't be a bad idea. It would be 100% test but they could definitely get some valuable information from the tests. In the latest tlp, they could have tested people getting into the character select screen and make adjustments before launch. I dunno, I'm just bored waiting to log in and wanted to kill some time!
And then what? Wipe it for real launch? Or you're giving some people unrestricted access weeks ahead of others? There were people on these forums demanding a server wipe because others had a couple hours head start.
One. They can't test this, stress test or the likes like other gaming companies. Two. We're in a Pandemic, I'm willing to bet more than half the player base is on unemployment. Creating an even larger playerbase to deal with. Three. They don't consider spending Thousands to upgrade their servers to cater towards the population this server has for the first 1-2 weeks. This happens every TLP. This is the FoTM server, people play it, and the population dies down.
Merge the 15 servers with lower population than the “ghost town”, cause obviously they are pointless.
Those two weeks end up killing the economy and player population. There is nothing standard about this. They could spin this up on AWS, Azure, etc. during this time and then migrate back to normal later.
https://spectrum.ieee.org/consumer-electronics/gaming/engineering-everquest It’s an old read obviously, but a good one. Managing these games is a lot more than just “spin it up on AWS” for a week. Especially after 20 years of adding to an outdated structure that had shifted owners multiple times.
Not relative, technology today compared to when that article was published. Their research was started in late 2004. The author provide no actual facts to backup their findings on the data center. The only facts I seen related to data transfer was 45 megabit cables to the render farm... I understand all the servers over the world and reason of using SoE. Spin it up on AWS, Azure, or similar service that allows the option to turn up services to reach demand.
It’s all still relative as it’s the basis of how the data servers works. Game wasn’t wrote for 2020 data centers, game wasn’t coded to take advantages of huge multi core processors. Everything about it’s design was based around 1999 technology not 2020. But it’s ok to think it’s as simple as spinning up a AWS but they just refuse to do that for some reason. “ More than 1500 servers around the world run EverQuest; collectively the machines have the bit-crunching capabilities of one of the world’s top 100 supercomputers. In just one of the data centers, more than 30 kilometers of wire and cable connect all the boxes. More than 9 million gigabytes of EverQuest data have been downloaded from Sony’s servers in the past six years.” That’s servicing Petabyte level data when the game was much smaller in size.
No it is not relative. I read the article and there is not even fundamental resources to gather from this article. It can be done but it does cost a pretty penny. It is coming down to profit.
Whole game could be rewrote also, but it comes down to profits. What’s your estimate of time, effort, and cost to restructure the entire game to run off AWS. Games that currently use AWS were built from the ground up to use it. Would love to know one example of a MMO retrofitted to run off it. Last I check Blizzard still rents thousands of data centers, Square-Enix also.
Sigh, you obviously have no idea and you are white knighting. I do not feel like busting out the AWS calculator but I can give you an idea of what the cost for 100 users 24/7 365 to upload on average 1 TB per month of CJIS data per year was quoted at 13,500 per year. Since AWS works like a utility company usage, memory, etc. determine your cost. I cannot guess their current users? Since it is not CJIS data it would be cheaper by the hour but they have more users. They are relying on their outsource or not opting to cover the cost of a more beefy solution. I guarantee they are not using 4500 shared servers across the globe any longer. SOE sold off or retool those locations a long time ago. They ran an analyst and probably have all the UX info they need to determine when to upgrade their services or not. DBG/DPG is now a mom and pop like organization for a 21 year old game I get it.
Regular server restarts to boot the afk's, queue and crash grace period to log back in, better hardware to support more pick zones.
Weird definition of white knighting....asked if you know of any MMOs that have been retrofitted to run off AWS. Or any other games in general that were not designed to run off AWS merged over to that infrastructure. But no I do not believe your taking the infrastructure they have built, calling up AWS for a quote, spinning up some VMs to shove it on and pew pew it it now scales and solved all their problems. Sorry that makes me a “white knight who knows nothing”.
It can be done and would save themselves one hell of a headache. They would hardly be charge if they did it months in advance. Since again it works like a utility company. I am no fan of AWS but obviously DBG/DPG no longer has the resources as they did with SoE. Not including anything but the game itself prior to launch could fit one 1 or 2 Snow Balls. They can add resources on the fly for the initial launch and peak of demand. The merger is nothing and Amazon AWS will work with you to drop the line in for the demand that is needed.
I just spent 3 hours waiting in the queue only to realize that once I finally got in, I logged in the wrong account - fml.