Sooo, are we there yet?
2 Likes
So what I want to know⌠were there any major timers or fights happening/brewing?
DDOS attacks generally happen for a specific reason. Maybe itâs someone disgruntled, but more likely (in my mind) a group had assets at risk and launched the DDOS to protect assets they were otherwise expecting to lose.
1 Like
Thanks for the 101 explanation, however lets try to bring it up a level since you are at least using acronyms in their correct contextâŚ
Again⌠weâre talking mitigating the damage of the attack. Not actually preventing it. Anyone worth their grain of salt in any IT/development capacity will know that mitigation of impacted services due to these types of attacks is priority number 1. Once the problem has been mitigated, then you worry about prevention. You literally learn this in any tech training that covers IPS vs IDS.
Some tech savy individual with the emotional age of 12 must be throwing a tantrum about something and has resorted to ddos attack to attempt to repair their own ego. Or Blizzard are waging stealth guerrilla attacks against their competitor.
1 Like
thereâs an easier way to do that, just do what Goonswarm did. Send in a bunch of carriers, launch every fighter wing and cause instant and persistent tidi. And then bitch about the tidi on Reddit like it wasnât their fault.
2 Likes
Since apparently the servers are still accepting connexion when they can receive all the packets I think mitigation works pretty well. You canât mitigate the damage of a nuclear bomb. CCP canât force internet providers to build terrabytes bandwidth networks on all routes going to their servers.
1 Like
Yes thank you for the update
1 Like
Yeah right ddos. Thats totally not a problem related with retiring xml api and crest. Everytime a player opens its hangar tranquility is shitting itself.
1 Like
Works when youâve got a bunch of carriers. What if you donât have a bunch of carriers? One dude can launch a DDOS.
You, sir, are stupid. The API made calls to the database. It never touched TQ and the database never depended on it. +1 point for being that stupid though, youâre reached the point where itâs actually impressive.
Do we already know what CCP did that pissed somebody off this hard to cause a DDoS attack?
Whatever it was, I want to put it onto my bucket list of things to do before I die.
2 Likes
Someone already mentioned a very high level solution in this thread. The actual application servers are exposed as the authentication servers are? The fact that all layers of connectivity were affected is not a good sign and indicates that no mitigation occurred. We were online at the time and everything went down at once (no new connections were being accepted at any layer).
Iâve seen lots of "it canât be done"s, and âit might be possible but god damn that would be nastyâ answers. Iâm in the âit might be possible butâŚâ camp.
Itâs extremely plausible that no mitigation occurred, because mitigation really isnât (feasibly) possible for this particular case and as such it isnât implemented. That everything went down at once would suggest that their edge was the failure point - which is hardly surprising as it would have the highest load.
1 Like
Did you map CCPs infrastructure or how do you come to such conclusions? There are many different possible reasons for what youâve described.
Also the weird image above with the cars and the bandwidth is only a very limited example of what it can be.
I mean, is this a jerk off thread for people who watched a few youtube videos or recently paid a couple of hundred $ for a certificate somewhere.
I really love this kind of people. Theyâll always smart talk how easy the solution is and how stupid everyone else is. The moment they face a real life situation, they are the ones who totally â â â â up, who cannot deal with the fact that things are fluid and that your certificate and ability to be a loudmouth doesnât help a bit, when things need to be solved.
How about let CCP do their job?
(or try to help in a constructive way)
1 Like
Symptoms described by players hint at congestion of the network. Itâs the most plausible to explain why players disconnected after a timeout
Network congestion can break even the most reliable infrastructure when itâs caused by an attack, because it targets the network. Once the pipes are full there nothing you can do, and CCP will not ruin itself to get pipes that are only useful in case of massive DDOS attack.
Ultimately you are connected through internet to CCP servers. If either you or CCP canât send or receive all the packets the game need to work, you will disconnect. You might get enough packets in and out at times, but a disconnection will follow when your packets start to wait again to be transmited. Thatâs why you sometimes see people manage to connect only to see some strange behaviors once logged in and a disconnection that follows soon after. This is either the server which canât process all the requests, or the network which lose packets on the way, or is too slow to transmit them. But if thereâs no one on TQ then itâs not a server problem but a network one. And the network congestion is also highly plausible when CCP says there is a DDOS attack running.
1 Like
PM me and lets compare professional credentials? This wouldnât be so irritating if the outage wasnât indicative of some major architectural issues that allow someone to cause a complete denial of service by only knowing about the point of entry (if we take all of the aforementioned âexplanationsâ at face value).
rofl what a trainwreck thread just lock this already
you donât know how to fix anyoneâs problems
Do you even know how internet works ?!
Unless you are talking about the flawed internet architecture ?
Honestly man, that whole thing is under the assumption that authenticated users are accessing the application the same way a non-authenticated request is used. Furthermore, itâs not taking into account any redundancy that should be put in place. Followed by no mention of identity management? Just stahp
There is no authentication or not, there is network congestion ! Where do you think the packets go in either case ?
Blockquote
Jenna Knor
Not hard at all if your talking about severs, why is the authentication sever the same as the primary gaming server? Why they even on the same connection? Just like DDOS uses proxies as a means to do what they do, they can be equally used to prevent attacks. Eves initial request server should be separated from itâs gaming server, where you get forwarded onto after being authenticated. DDOS should only prevent people from connecting that arenât already ingame.
Jenna put it pretty well in the quoted text above. Again, weâre talking about damage mitigation, not prevention. I donât understand why this is such a heated topic. Literally all of the services went down at once (including people reporting slowness on some of the Web services).
I donât care how they fix it, Iâm here to say that they should take a more proactive approach and not play the victim.
1 Like