Higher ping values on Amazon-hosted sims

[Shakeno Tomsen]

Hello, I don't really post much often here, but this is concerning me enough to actually begin a discussion in the forums.

I've noticed that some sims I used to go to are migrating to Amazon. One of my biggest worries was the ping sim going higher due to higher distances to the server, and my worries ended becoming reality.

I'm located in Spain, and with the Linden Lab servers I used to have between 180ms minimum to 260ms peaks in ping sim. I averaged around 220ms~230ms. Skill based action games were still somewhat doable, I was just at a slight disadvantage but things could still feel responsive if scripted properly. Now, since they moved to amazon, my ping sim has increased to between 240ms minimum to 320ms maximum. Games/minigames in Second Life or Second Life combat in general is almost unplayable due to the high disadvantage and the general clunkyness SL gives off now. I can mouseclick, and release the finger, and my action still hasn't reached the sim.

Now, before anyone goes on to blame my internet connection, I have a solid 1GB/s connection, and it has been working pretty well before the switch, save for certain stutters that happened for different causes. I believe it is safe to assume that something that has worked for so long (10+ years) isn't to blame when the problems began after one of the variables changed (the switch over to Amazon)

Of course, this is very likely to be a distance related issue. People from other parts of Europe will have different pings. Probably people in the UK will have a lower ping sim. People in the US may not notice much of a difference at all. However, for those who are further, this is going to drive them away, and reduce the community of people who are interested in the fun videogame potential SL had to offer. I'm trying to teach people that SL is more than a visual chat room, and up to this point it used to be, it offered a lot of fun things to do, despite the "bAd GrApHiCs". But things like these which makes the core of the game feel clunky kills the incentive of actually playing when there are other platforms getting the attention of former SL residents.

So, this is a somewhat long post, so let's conclude it here:

What am I asking to LL: Please consider requesting a migration to a more central location in the US: more eastern so European residents can get a ping similar to what it used to be, and more southern, so people from South America also can get a better ping, similar to what they used to have.

What am I asking to residents: Please share your thoughts, experiences, and if possible, see if we can raise awareness for this to reach LL. For people in the United States who are considering criticizing this proposal, please consider that maybe getting +40 ms ping in your end might mean switching the game from extremely clunky to playable for people in other parts of the world. There are people who have a much worse ping than I do who can't do anything other than chatting. Simply going from 230 to 320 has made the base Second Life avatar movement feel much clunkier. I don't want to imagine those having a ping higher than 350...

[Aishagain]

@Shakeno TomsenI am not seing a significant increase in sim pings anywhere in SL during this stage of the uplift, if anything most places aerw showing a slight decrease in ping.  Bear in mind that the sim ping displayed in the viewer always includes a render-time component and is therefore variable depending on your client FPS.

Oz Linden has already said the one aim of LL post uplift is to have regions hosted on cloud servers closer to many non-USA user populations and European hosts are only one of their stated aims.

You suggest that Spain is more affected than my country (UK) though I can see no logic in that. Also bear in mind that users in Australia have been making do with a normal ping of somethinglike 300ms for years and having a host for some regions in Australia would radically improve their SL experience.

I re-iterate my initial comment:  I see no reason why migration of a region from the Phoenix datacentre to AWS cloud would increase your sim ping at all for a given client frame rate.

[Shakeno Tomsen]

@AishagainThere is logic, because you are much closer to the relays where the intercontinental internet cables are connected to. A connection to the US must travel through many more relays and travel a higher distance when it is started in Spain than from the UK, same reason as why someone in the US would have a much lower ping than you do.

And, no, it isn't a FPS based issue. As I said, I've been in SL for longer 10+ years, and I am speaking on averages, but if you insist this is a FPS issue, even empty sims where I have 60+ FPS exhibit this problem.

[EDIT]: Sorry, I didn't reply to your point for people in Australia. It is a very good point, I never posted anything that asks to not to host sims in other parts of the world (In fact, I wasn't aware of this). But what I was asking for was, if they are in the US, for them to be moved to places where other parts of the world could benefit much more.

Edited November 14, 2020 by Shakeno Tomsen

[Candide LeMay]

Yes I can confirm that for me the connections to the aws hosted sims have about 30-40ms bigger latency compared to the old sims (similar numbers to what OP posted). I'm in central Europe. Our loses are Asia's wins *shrug*

[Aishagain]

To both @Shakeno Tomsen and @Candide LeMay being in the UK I usually saw 160ms ish very variable (from as little as 140 to as high as 205ms on occasion) from datacentre regions.  Now I see 175ms steady with little variation from uplifted regions.  So, OK there might be a small increase, though quite where the AWS cloud servers being used are located I don't know.  It is possible that the signal passes through more relays and hence latency might increase accordingly but I have not measured this presently and I must admit I have seen no issues from it. 

The connection to the Transatlantic cables issue I had not considered so perhaps I am ignorant there.

[Fritigern Gothly]

During one of the weekly interviews, meetings or lab gabs, @Oz Linden said that at first, SL will be living in a single AWS data center but that in due time, and once the worst wrinkles are ironed out, the plan is to have SL live in datacenters distributed all over the globe.

I will see  if I can find the interview in question.

 

 

Edited November 14, 2020 by Fritigern Gothly
EDIT: Can't find the article, to tagged Oz in the hopes he will see it and chime in.

[Ardy Lay]

 

[arabellajones]

There's at lest two different measure, with similar names.

The viewer ping which can be obtained by the viewer and is shown in the Statistics Window includes any current delay from the servers.

The internet ping is almost entirely the internet

There are various services which try and test connections to AWS data centres round the world. This is one test of AWS servers and the numbers at the time I used that link looked scary bad, over three times typical internet ping numbers

I don't know what to expect from the Amazon servers, I am used to the internet ping times. https://www.speedtest.net is a reliable service, and they have a server in Phoenix which is a useful check on the general internet. England to Phoenix is around 150ms, and I don't think that is far wrong for anywhere west of Chicago. On what I have seen today, the internet is running OK, but the Amazon servers are looking dodgy. But I know that some of these web-page stats services can be misleading. It's possible that I may have tried to use them at just the time (early Monday morning) when maintenance is being done.

It's worth remembering that a lot of the internet ping has nothing to do with distance. My results give 20ms for about 150 miles, 150ms for 5300 miles. It's the IP switching at the nodes which dominates the timing. The number of nodes may be the same in a connection from Europe. A web page may give slightly different answers to the classic commandline utilities such as ping or traceroute

I have a feeling that Google Search is misbehaving. One check can never be certain.

[arabellajones]

Tiny update: I just did a reboot on my internet modem/router after 28 days uptime. I have good hardware running OpenWRT 19.07.4 and it doesn't show any problems. Some of these boxes have had problems with SL in the past and, a bit like the servers, a reboot can sort things out. The downside for things such as ADSL and VDSL is that dropping the connection too often can be interpreted as a line-fault, and the connection speed is dropped to improve reliability.

My last reboot was a momentary power cut.

[Shakeno Tomsen]

@arabellajonesIt is very likely the DNS is giving you a different relay when trying to ping the Amazon servers. I know Google does it, because www.google.com and www.google.es give me the same ping result. Back then, I was able to get different results and have the ping application return me what is expected from a US location, using the US IP from Google, but now I always get the same result, expected from a Spanish location from both the Spanish and US IP from Google. Could Amazon be doing the same? Really weird stuff.

Regardless, my ping values were given from the Statistics window, from both Phoenix sims and Amazon sims. It also makes sense, because I get a response expected with those ping values if I make a click and wait for the sim response to happen. A bit after clicking and releasing the mouse button, the sim response comes through with a noticeable delay. It was shorter back then, although still noticeable, just not as long as now.

[bigmoe Whitfield]

typically i'm at 36ms or 67ms depending on mainland sim I'm tinkering in.  today I'm 73 to 107, which is not bad.   I used to have a bad hardware bottleneck when I first joined sl in 07... 1200MS! omg those days are long gone for me.