Area52 here, it's the usual server issues. Can't say I'm surprised. At least I have fun laughing at people in chat fighting over WoW and FFXIV.
"I can no longer sit back and allow Communist infiltration, Communist indoctrination, Communist subversion and the international Communist conspiracy to sap and impurify all of our precious bodily fluids. "
- General Jack D. Ripper.
Is it really that easy to scale up and then down 100x or even just 10x without additional costs? And is it really virtualized? You still need hardware from somewhere to provide the additional capacity.
I assume Blizzard run their own servers backends so where is that extra HW coming from?
How do you know all this? I am no server engineer so my knowledge is zero.
Everyone and their mom is using virtualized resources. The one who is not is providing such services.
There are some costs of course; my point is, it’s only opex for running a dozen extra servers for a few days and there isn’t a blocking upfront cost that you’d otherwise face with having to purchase hardware.
You click a button and your preconfigured server is running from a template like 15 minutes later, as low as $200/month.
So we are really talking about blizzard not willing to include an extra couple thousand dollars of one-time opex.
- - - Updated - - -
Let me give you an example on “how easy”. A company that is in worldwide top 5 antivirus vendors, is, naturally, testing their products.
To do so, they have to roll out a couple dozen of product editions on a several dozen of OS versions, on another several dozen of language configurations.
They do this every time new major build is delivered by the dev team - so basically at least once a day.
We are talking about close to 10k virtual machines being deployed automatically from templates, product installed there, automated tests being run (sometimes several hundred per product version). Then data is collected and all virtual machines are decommissioned.
Yes, they do it every day. Sometimes more than once a day. It’s all easy and automated.
day 2 of unplayable shitty experience
some people seem fine, for my character trying to loot or turn in quest results in 30+ second lag
Once you get out of the major population bottlenecks things seem fine, more or less. A couple of issues with instability that boots you out now and again, but I've not gotten any anomalous queues or failures to be able to log back in. Yesterday at the transport hub to get to the Isles was a nightmare, but I think that was more or less down to the huge presses of people just causing lag issues and the like - once past that, relatively smooth sailing.
"We're more of the love, blood, and rhetoric school. Well, we can do you blood and love without the rhetoric, and we can do you blood and rhetoric without the love, and we can do you all three concurrent or consecutive. But we can't give you love and rhetoric without the blood. Blood is compulsory. They're all blood, you see." ― Tom Stoppard, Rosencrantz and Guildenstern are Dead
I imagine WoW servers are a little more complicated, Your not talking about starting 100 completely separate and isolated virtual machines, there are a lot more interconnected systems like player data that needs to be shared by all is extra capacity that could cause bottlenecks.
And I wonder if there is enough idle hardware doing nothing in these datacenters that Blizzard can even just get 5x (or more) their usual capacity. Which I assume is pretty substantial at normal load.
It ignores such insignificant forces as time, entropy, and death
You should really read about cloud computing and how it works. It’s a service; you pay for it, you get what you paid for, you get it immediately (within minutes) and you pay as you go.
And saying that blizzard’s servers are more complex that testing 10k configurations several times a day - all while managing these configurations and templates - only shows your total inexperience in IT area.
As I said it earlier, we’re talking about $2-3k savings, one time savings, at blizzard side.
Just pure greed.
they ran some new servers with Classis from the cloud but that was an entire server somewhere else that was later moved to their datacentres.
Expanding existing servers to run both local at their datacentre and cloud for additional capacity is bound to run into lag issues.
Sorry but I'm going to take Blizzard, and basically all other companies that all have issues with massive temporary load increases, trying to fix it for years and years with varying levels of success over some random guy on the internet saying its ez and would only cost 2k.
It ignores such insignificant forces as time, entropy, and death
As this is no longer the case since launch, no real need for this thread to remain. Closing this.
"We're more of the love, blood, and rhetoric school. Well, we can do you blood and love without the rhetoric, and we can do you blood and rhetoric without the love, and we can do you all three concurrent or consecutive. But we can't give you love and rhetoric without the blood. Blood is compulsory. They're all blood, you see." ― Tom Stoppard, Rosencrantz and Guildenstern are Dead