Quote Originally Posted by Kirisunna View Post
Normally I wouldn't post three times after myself but I have an answer for you guys. I talked to tech support about it and here is the conversation. Basically if you already reset the router and modem and contacted your ISP you are screwed. <3

The answer to the dc issue.
PSN has had issues in general for quite some time now. Apparently, it has also been a recent target for attacks as well:
https://downdetector.com/status/playstation-network

Sony even put up tweets and such about them having issues over the weekend as well. It would appear things haven't completely been settled yet. Even though gameplay on the PS3 is less dependent on the PSN connection than the PS4 once you get pass initial login, your connectivity in general can still be impacted in a peripheral way if they are still under attack and your routing is taking you through the same corridors as the attack vectors (flooding a sever via a DDoS attack affects more than just the server since it is getting attacked across multiple attack vectors--ie, from multiple networks).

It keeps going back to connectivity across the internet. Not just your local connectivity to your ISP's nearest headend, nor SE's connection to their ISP's nearest headend. More often then not we are seeing issues with interconnectivity between these two endpoints.

If you don't know the actual IP your client is using (PC users can get it from resource monitor (resmon from run/search line) or netstat (run that command in a CMD prompt) while fully loged into the game, console users might be able to find it by looking at the current connections in their router's admin pages if it provides such data--the Canadian IP's all start with 199), you can at least run traces to the lobby servers to get a better idea of where you are going and if there are any signs of potential elevated congestion and/or packet loss that could be impacting your connectivity. This will be important information to pass along to your ISP's Tier3 support (sometimes referred to as Engineering for the help desk guys that are clueless).

Tier3 support will be your best advocate for getting this investigated properly, but they need these details. Even if you don't run/submit trace data, at the very least they will need a valid IP to research the routing... providing them even just the lobby server addresses will get them looking in the right direction. They are registered with DNS names, so it may be helpful to provide both just in case (the game actually calls the lobby servers by the DNS names initially). Here are the names, followed by the actual IP addresses.

neolobby02.ffxiv.com (199.91.189.74)
neolobby04.ffxiv.com (199.91.189.93)
neolobby06.ffxiv.com (199.91.190.106)

**Note these are the current addresses assigned for the NA/EU lobby servers. This may change once the new EU datacenters go online. The names should still be valid though, they just may get reassigned to new regions or new IP's down the road. The 3 JP lobby servers follow a similar naming scheme, but use the odd numbers 1,3, and 5 in the name and the IP's currently start with 124.



For those unfamiliar with running a traces:

You can run a trace from a command prompt by typing the command "tracert", a space, then the name or IP address (in some other OS's, you would use the command "traceroute" from the terminal for the same thing). If you want to forward the trace results via email or post them it a forum, you can grab all the text from the command prompt window by simply right-clicking in the window and clicking "Select All" to highlight everything. To copy specific text, you can right-click and click on "Mark" and then click/drag to highlight specific text. Once the desired text is highlighted, pressing the <ENTER> key will copy it to clipboard for pasting into an email, text editor, or forum post and such.

Ideally, if you are going to gather this data, you want to have samples of good connections and bad. Either capturing the game in both states, or traces to other known good services like another game or hulu, google, netflix, microsoft, etc. will help to demonstrate the issue for them, as well as building a stronger case for them to get involved. I would provide a sample with a break down of things to look for, but TWC-SC has currently got me running pretty tight tonight, so I don't have any examples of what to look for with my connection atm. The usual tell-tale signs are a wide variance in response times at the same hop, or hops that intermittently return a * result in the mix (when it goes too long, it gives a * result for that packet). There will be slight variances in the response times at each hop, ideally less then a 10% variance between each response... but you don't want to see big jumps. For instance, three results of 40 45 48 actually isn't much reason for concern... but 40 45 98 is a sign that utilisation is starting to run on the high side, and results like 40 129 43 or if you see something like 40 * 48 is bad sign that congestion has exceeded acceptable levels and needs to be addressed.

When packets get delayed/dropped in these situation, it wreaks havoc on this game for several reasons. The game depends on timely, orderly delivery of packets in order to process all player and non-player activity in the game. Your client doesn't progress a battle until it recieves the data from the server. When packets are getting delayed/dropped like this, the error correction in the TCP/IP protocol will try to compensate for the out of order/missing packets by requesting them to be retransmitted. This causes the skipping/rubberbanding and such. If it gets too far out of hand, it will start to narrow your transmit windows, effectively slowing your transfer speeds, which can actually exacerbate the problem to some degree (like at an S rank hunt---it may prevent your system from keeping up with all the player data properly). If it gets really out of hand and TCP/IP does in fact throttle you down to slow start status because it is detecting poor line quality/congestion, the session can drop and you get the connection lost errors. And then you get stuck in the 90k/2002 error loops... and you may remain there until the congestion subsides.

All of this can and is often triggered by the last-mile ISP's (the local guys we pay for internet) overselling markets and undermaintaining the exchanges amongst the ISP's that must work together to provide interconnectivity across the internet. This is something your ISP is largely responsible for maintaining in partnership with higher level ISP's---thus they are an ideal candidate for researching and <hopefully> addressing the problem on your behalf. But you need to reach beyond the helpdesk...the guys that answer that 800 number are typically the lower tiered support. If they can't take your information and escalate the matter for you, then you need to push past them to get to a higher tier. You may have to ask for a supervisor to get it... but it is there, regardless of how they try to convince you otherwise. Until you get past their own NOC and move on to the top tier ISP's (like Level3, TATA, Cogent) and their Admins... there will ALWAYS be a higher level of support that can work on these problems.

Bottom line is, you are paying your last-mile ISP for their service and support... need to hold them accountable to provide it.