Latency/Connection Issues Since the Launch of the EU Data Center

[Moonraker]

Since EU data center move generally ping has been halved BUT now at prime time bad lag issues. Details as follows including ping data collection;

Name of Internet Service Provider (ISP)

Orange France

Country of residence

France

Reference number from your ISP, if you have raised this matter with them

N/A

When does the problem occur (e.g. only during specific times such as in the evenings?)

Prime time generally evenings around 19:00 > 00:00

Frequency of problem (e.g. every 2 minutes)

All the time especially noticeable during crafting (lag of spells etc.) Gathering (lag of gather up to seconds to complete) When partied see rubber banding with others travelling etc.

Game content in which the issues occur (e.g. dungeons, trials, certain areas)

Crafting, Gathering. Grouped.

IP address (optional)

90.30.116.120

Since the data center move to Europe (Frankfurt Germany KDDI) I have noticed normal latency halved from around 150-160 to 50-70. BUT now in prime time lag is an issue as stated above. When crafting and gathering there can be seconds of delay between gather and spell complete etc.

Checking the tracroute over several evenings now it seems at least that one issue is the transition from Orange FR > International network via Opentransit (level3-2.GW.opentransit.net) This node is Orange/ France Telecom hop/ node from France to International network i.e. Germany and Frankfurt to KDDI data server host. Opentransit being Orange/ France Telecom.

This may not be the whole issue as Frankfurt and KDDI internal network may also be an issue but this is evidently the start of the issue in primetime from France and probably other Orange customers from southern Europe like Spain.

So I have been analysing the connect over (slower) 2MB ADSL connection from France to FFXIV data center for Lich (EU) server. This is a typical daytime traceroute record;



Note that packet lose and higher latency for hop 6,7 &8 can be attributed to those hubs giving low priority to ICMP ping requests. Basically it is the final hop/ node that matters and this is around what is expected at 76ms ping from rural France to Germany.

BUT come the prime timer evening traffic, there is a noticeable impact on that node at:

193.251.255.82 level3-2.GW.opentransit.net

From that node the latency climbs from around 43ms ping to over double 110 ms ping and continues for final 137ms ping. Notably and probably more important is the packet loss which will cause hitching/ lag.



The hop 7,8 are probably down to lower priority ICMP ping response so discounted. Yet now, since hop 6 there is over double the latency right to end ping at Lich server.

This happens each evening therefore appears to be a contributing issue for any connection via Orange out of France (via Open transit) to Germany and the data center hosted by KDDI.

There may also be contributing issue after via Frankfurt or internal network of KDDI but this does show it starting from the opentransit hub so certainly a good starting point.

I guess that the popularity of gaming servers now in northern europe including gamers in southern europe, routed via this node is saturating it at prime time, given that it appears to happen each evening (since swap to EU center) Indeed it has been identified previously for example in this post/ thread from Spain.

Either way I have tried to identify the more obvious issue in network from France. Maybe if other can check if they rout via this hub via tracert etc. it can help.

In the end I have real problem now even with data center in Europe which is great for latency but now introduces some lag / jitter issue via the network.

Hope this help for resolving the issue. if you want to research this a bit more try Pingplotter for better data which is Free version and can use the Pro detail free for one month anyhow but Free is quite fine.