Error 90000 & lag

[Raist]

Sounds like the typical issue of congestion happening at the choking points like everyone else--either during the hand-off to their routing partner, or between there and the hand-off into Ormuco (SE's ISP in Canada). Where you are seeing the big intermittent lag spikes is what needs to be looked at first, as that is your first point of congestion that can compound additional congestion that may be occurring further down your route.

This is why generating a tracert to your actual server is important. It gives a roadmap of just what segments you are coming through, and if you can get both a good and a bad snapshot trace, it gives clues as to where things are going downhill so they know who to contact and what action to take if it can't be resolved--as in they can change the route if necessary to avoid the congested corridor. There are roughly less than 20 paths to take to get to Montreal from the major players, so it's necessary to periodically change up the routing partner sometimes when one of them is having congestion issues.

Have to also consider that it isn't JUST FFXIV traffic going through the same corridor that you are using. Streaming traffic in general more than doubled this summer, at times the volume has tripled. That's all streams considered as a whole--gaming, netflix, VPN's, financials, shopping, just generic curiosity and You-Tubing/Twitch and such. ALL internet traffic has been on a steady rise for some time now, and it has really ramped up in the last 6 months. This is leading to a huge congestion (and thus stability) issue in the major corridors across not only the US, but worldwide as well. It's been covered extensively on various forums/blogs and not just here. But if you browse/search the threads here about the lag/stability issues, you will see lots of maps/charts/quotes/links to more info on the matter if you are interested. There's been a series of posts on Level3's blog about the issues with peering where the ISP's are not operating under best practices, and we the consumers are suffering from it. I'll provide a link to one of the posts that may help you understand the mechanics at play... there are links within the post and in the comments that can shed more light on the issues if you are wanting to read more on the matter:
http://blog.level3.com/open-internet...net-middleman/

For just a quick visual of the impact from all of these peering/transit congestion issues, you can view maps for some popular services monitored over at the downdetector.com's companies list. Just for an example, I will include an image of the map for my ISP from them. It updates periodically, so the areas with issues will change throughout the day:

https://downdetector.com/status/time-warner-cable


Right now, it is pretty clean with just mild issues at some of the big bottleneck spots on the East and West Coasts, with some small blips in Texas. During prime time, these regions frequently go into the red territory, as well as some hot spots in the North East and around Chicago/Ohio. We have been on them like rabid dogs for about a year now to clean up their act, and it is slowly paying off. They're in the process of rolling out fatter pipes in the metro areas, and it's allowing them to split off more nodes and spreading the traffic out more. An unfortunate side effect is we are getting slight increases to our average latency, but we are also seeing dramatic reductions in the spikes--making it much more manageable. Hopefully the momentum continues... it took a lot of work to get the ball rolling on this, hate to see it fall apart in the end.

Just for comparison sake, here is also a map for AT&T--note that some of the same areas are experiencing issues, often at the same time. This is because there are issues at the peering/transit that need to be addressed--that is between our ISP's and third party ISP's that are providing the interconnectivity between all the different networks that make up the world wide web:

https://downdetector.com/status/att


And, just for completeness sake.. here is an online suite of tools provided by Akamai that show stats on web traffic. Lots of tools built in to this--different drop boxes and filters to pick different types of stats and to filter by industry and/or region. It's quite an eye opener on some of the pages when you see how much traffic ebbs and flows daily, but when you look at the history over a year you can see how there was a sudden jump in usage. For an industry that sometimes only adjusts policies quarterly or semi-annually, the sudden spikes that are sometimes displayed in these stats can be overwhelming at best, devastating at worst:
http://www.akamai.com/html/technology/dataviz1.html

TLDR:
There are known, well documented issues amongst the ISP's that must collaborate to maintain the complicated web of networks we refer to as simply the "Internet". There are lots of articles and tools to demonstrate the problem. But, in order to get it addressed... we need to collectively put pressure on our ISP's to escalate the issue to higher tiers to better manage the congestion cropping up with their routing agreements. This can't be managed by your local techs by the way... it needs to move up the chain to at LEAST their Tier3 guys (sometimes called engineering). They are the ones with the resources to identify the potential issues more accurately and pull in the proper people to investigate further and <hopefuly> address the issue more directly. At the very least, they can be instrumental in altering your route in the short term to help bypass the congested corridors once they can be identified. One of the best things you can do is provide them a trace or at least the IP of your actual server your client connects to in order to conduct a more proper investigation. You can use netstat or resmon while logged fully in game to find the IP--Canadian server IP's all start with 199.