World server is down.

[Relapses]

I long for the days of infinitely stacking gyrocopters.

[Breccia]

I can safely say, teleporting inside Brawl'gar and walking out the door is not the clever bypass I thought it was.

[Eosia]

I can safely say, teleporting inside Brawl'gar and walking out the door is not the clever bypass I thought it was.

I tried to bounce of Dalaran and Draenor... Dalaran failed and Draenor just destabilized.

[Izual45]

Good explanation (and to clarify I'm not being sarcastic). If this is the way it goes nowadays, I wonder why Blizzard doesn't do it. Would it be a cost issue?

I can't imagine they actually want their launches to be bad so they get all this bad rep.

There must be valid reasoning. Perhaps building anticipation, charging up players left behind at the small chance of complaints, or purely a very cheap trick to save money. But that's how all the big evil corporations work.

[Visor]

welcome to the world of diversity. I suppose white cisgender males who know something about highload servers was fired. And now this problems will solve by body type 3 ppl. If they are able to do this ofc.

[vincink]

This is the cloud era, not your hardware servers. A company such as blizz coupled with AWS has the capability to increase and decrease its resources on a timeline of seconds. Hell, all of this is automatable based on traffic patterns.

In conclusion, yes Blizz isnt filled up with idiots but accomplished and well-trained guys, however, this practice of not anticipating high traffic on launch days and hiding behind the guise of "nothing is perfect", while raking in people's money is quite unethical.

Re: the bolded part.

AWS, Azure, and the like are not infinite resources. I have witnessed engineering teams accidentally discover the limits of these cloud hosting platforms. When the CTO of AWS calls the CTO of your employer to tell you to stop beating up "the cloud", it gets very real.

Additionally, the cost to have push-button, instantaneous availability of new computing power to meet demand that is scaling at orders of magnitude runs in the millions very fast. So, no, Blizzard cannot simply scale instantly without tremendous cost. There are smarter, cheaper ways to handle massive scale-ups in legitimate traffic but there are always trade-offs. That is what we are experiencing right now. Unfortunately, there is no way to mitigate customer bellyaching.

[Monteverdi]

Re: the bolded part.

AWS, Azure, and the like are not infinite resources. I have witnessed engineering teams accidentally discover the limits of these cloud hosting platforms. When the CTO of AWS calls the CTO of your employer to tell you to stop beating up "the cloud", it gets very real.

Additionally, the cost to have push-button, instantaneous availability of new computing power to meet demand that is scaling at orders of magnitude runs in the millions very fast. So, no, Blizzard cannot simply scale instantly without tremendous cost. There are smarter, cheaper ways to handle massive scale-ups in legitimate traffic but there are always trade-offs. That is what we are experiencing right now. Unfortunately, there is no way to mitigate customer bellyaching.

Yeah people think 'the cloud' is some magic button to get all the resources you need, when you need them. It is not.

That said, WoW doesn't run on AWS or the like. They have their own ESXi clusters on their own hardware in their own datacenters. The issues tonight weren't a hardware failing, they were a software/planning failure, but this wasn't a hardware problem.

[Izual45]

lmfao what do you think the 'cloud is'? It's all on "hardware servers". I work for Blizzard's hardware vendor, I have helped run firmware updates on their hardware. 'The cloud' is not some ephemeral magic place where you don't have to worry about hardware lol. WoW runs on a cluster of 14th Gen Dell Blades running ESXi 7.0u3.

What is happening isn't a hardware problem. It's not a matter of needing to allocate more hardware in anticipation of high traffic. This is a software issue.

There is no software issue, no hardware issue but a bottleneck created on purpose. Funny how streamers get in but your average job has a slim chance.

I would suggest you read up on AWS/ Azure and the way the resources/servers spin up and are able to be operated on basis of milliseconds and billed at the same rate as well.

Infact, AWS modules even take up the responsibility of scaling up and out based on traffic and then reverting when required, thus eliminating human intervention.

[Relapses]

Yeah people think 'the cloud' is some magic button to get all the resources you need, when you need them. It is not.

That said, WoW doesn't run on AWS or the like. They have their own ESXi clusters on their own hardware in their own datacenters. The issues tonight weren't a hardware failing, they were a software/planning failure, but this wasn't a hardware problem.

To be fair, one of my favorite parts of any launch are the inevitable quadruple PhD in MMO-serverology software engineers who chime in with some weird quasi-technical explanation about how Blizzard's failure to anticipate this launch's issues is because they're deliberately running the game on old outdated software/hardware that costs too much to upgrade.

[Monteverdi]

There is no software issue, no hardware issue but a bottleneck created on purpose. Funny how streamers get in but your average job has a slim chance.

I would suggest you read up on AWS/ Azure and the way the resources/servers spin up and are able to be operated on basis of milliseconds and billed at the same rate as well.

Infact, AWS modules even take up the responsibility of scaling up and out based on traffic and then reverting when required, thus eliminating human intervention.

WoW doesn't run on AWS. For a number pretty solid reasons. What's happening right now is not a resources problem.

[minteK917]

Great launch my server entire kalimdor been crashed for 2 hours. Blizzard arent reseting the server or anything. I can log into eastern kingdom on my alliance toon. So ever since the first zipline left from durotar, nobody else in the horde can get to the dragon isle and we cannot play since we are all bugged in kalimdor since its the only way to get there. Fucking smart design.

[Monteverdi]

To be fair, one of my favorite parts of any launch are the inevitable quadruple PhD in MMO-serverology software engineers who chime in with some weird quasi-technical explanation about how Blizzard's failure to anticipate this launch's issues is because they're deliberately running the game on old outdated software/hardware that costs too much to upgrade.

I have the interesting perspective of having worked on some of Blizzard's servers (not WoW servers, but Blizz servers). I know WoW is running on 14th gen Dell blades in an ESXi 7.0u3 cluster. Not bleeding edge by any stretch, but not outdated by any stretch as well. They're on pretty standard, workhorse, reliable infrastructure. And their datacenter guys run a tight ship.

[CosmicAnomaly22]

I have the interesting perspective of having worked on some of Blizzard's servers (not WoW servers, but Blizz servers). I know WoW is running on 14th gen Dell blades in an ESXi 7.0u3 cluster. Not bleeding edge by any stretch, but not outdated by any stretch as well. They're on pretty standard, workhorse, reliable infrastructure. And their datacenter guys run a tight ship.

Source: trust me bro

[Izual45]

Re: the bolded part.

AWS, Azure, and the like are not infinite resources. I have witnessed engineering teams accidentally discover the limits of these cloud hosting platforms. When the CTO of AWS calls the CTO of your employer to tell you to stop beating up "the cloud", it gets very real.

Additionally, the cost to have push-button, instantaneous availability of new computing power to meet demand that is scaling at orders of magnitude runs in the millions very fast. So, no, Blizzard cannot simply scale instantly without tremendous cost. There are smarter, cheaper ways to handle massive scale-ups in legitimate traffic but there are always trade-offs. That is what we are experiencing right now. Unfortunately, there is no way to mitigate customer bellyaching.

In the engineering world, there is a thing called proof of concept. Which is what the majority of AWS solutions architects work on with the customers. Here a design and potential solutions are built and tested before the go-live day. You will NEVER hear from AWS/ Azure or any major cloud providers that our resources are limited. I know this because I work in this area.

What I do agree with you about is the cost, as usual, the big evil corps do anything to save money. The fact is that company growth is directly proportional to the investment of capital in resources.

[Zayrinoke]

welcome to the world of diversity. I suppose white cisgender males who know something about highload servers was fired. And now this problems will solve by body type 3 ppl. If they are able to do this ofc.

Holy shit the victim complex here was strong enough to get me to log in for the first time in 5 years. What a sad life you must live.

[Monteverdi]

Source: trust me bro

Don't need a source, it's not like these are industry secrets.

The point is, people who don't know anything about the Enterprise IT world like to say "blah blah blah just use the cloud!" "blah blah blah there is no hardware anymore" and the fact is that there is hardware still, the 'cloud' isn't a magic resources solution, and there's nothing about this situation that screams that it's a hardware problem. People want to say Blizz is cheaping out on old hardware and that's the problem have no idea what they're talking about.

[Caprias]

Source: trust me bro

Quite a bit more believable than your sorry ass burner account.

[Izual45]

Don't need a source, it's not like these are industry secrets.

The point is, people who don't know anything about the Enterprise IT world like to say "blah blah blah just use the cloud!" "blah blah blah there is no hardware anymore" and the fact is that there is hardware still, the 'cloud' isn't a magic resources solution, and there's nothing about this situation that screams that it's a hardware problem. People want to say Blizz is cheaping out on old hardware and that's the problem have no idea what they're talking about.

Yeah your source is still as shitty as your explanation. Tell us more about Blizz inside secrets, are the servers running on diesel generators??

[Echeyakee]

Loading everyone on an airship was a bad idea.

[RobertMugabe]

Unpredictable number of player activity, because so many people chose religiously not to preorder (even though they would 100% play the game anyways on launch).
It's funny how I see several people moaning about the launch that insulted people who preordered several months ago.