10 hours ago, Tuskin38 said:

Probably an issue on your end, I haven't had any of this.

Exactly the same here:  no stutters, no weird anything.  Why is it I never have this garbage and others do?

10 hours ago, Adamski_NZ said:

You're lucky!!! I don't think it's me ... I have super-fast broadband and the Internet has been performing flawlessly in every other respect. I see that last sentence of mine echoed a thousand times throughout the history of this "sim".

Three disconnects for me tonight Adam and I am in Auckland too. It's not you! 

 

 

 

9 hours ago, irocx said:

If Microsoft want's to develop a sim so dependent on internet connectivity, it's incumbent on them to provide the necessary server capacity/performance to meet the needs of the users - especially when their customer has done their part.  

The situation is too dynamic.   MS doesn't know for certain how many will log in, and for how long.  Maybe in April 2022 a million users will log on at certain times of the day, then in 3 months that number dwindles to 500K users, and maybe in a year only 100K users log in at any one time.  Building sufficient capacity for any possible level of demand sounds dicey and so the idea may be flawed outright because of these uncertainties.   

9 hours ago, LRBS said:

I fully agree, there is something terribly wrong and upsetting with the way of their "progress". 

Besides their "eye candy" scenery, with every update or so-called "fixing issues" they manage to break or disturb something else that was working acceptably.

 With this latest update, I noticed this turbulence behavior to be more pronounced and unrealistic.

They were supposed to tone it down, now with no winds at all (turned off), you can notice every other a few seconds wind kicking in a few knots and direction changes. It feels like a constant light chop.

I don't understand what kind of inputs they get and from whom but all this is very disappointing to the point where it's not worth it anymore. 

At this point I wonder if we really needed that SU9 or what's good about it. Something is going terribly wrong with those guys.

28 minutes ago, Pathfinder633 said:

I said in Practice...in practice who would enact your dumb scenario?

Come back when your reading comprehension improves

Accessing files from distant servers is common, whether you realize you are doing it or not.  While there are a lot of instances where the net effect of this is irrelevant, if the application is very ping dependant as gaming definitely tends to be, server location is very important, both as far as physical distance, and the number of hops and network routers that it passes through on the way is increased, increasing the net effect of any latency induced on any particular hop.  Added to that, with most isp’s the path that each individual packet takes may be different, which can lead to packets arriving out of order (Jitter) which means that the operation has to wait for them all to be in sequence.  All this happens in a number of milliseconds, but in gaming applications this can be significant to the operation.  
 

if you know the ip address of the server you can use the traceroute and pathping commands in windows to see where the bottlenecks might be, but unless it is within the first couple (your pc to your isp) or the last couple hops (Microsoft’s servers). It’s unlikely that anything can/will be done.

48 minutes ago, psolk said:

You may be bypassing some local bottlenecks with a secure tunnel direct to the endpoint with a VPN depends on packet routing.  

Could it be that, instead of a connection to a "local" MSFS or Microsoft server, I am connecting to the one in Europe?

The same thing is happening with DJI and other server

 

We recently had to bring up our server capacaties too (weather, telex, etc.). Cause the number of users exceeded our expectations. Maybe MS is facing the same problem.

 

 

Perhaps if everyone logging in used a very big rolling cache there would be less demand on servers.  Mine is 100Gb.

28 minutes ago, bendead said:

Could it be that, instead of a connection to a "local" MSFS or Microsoft server, I am connecting to the one in Europe?

The same thing is happening with DJI and other server

 

Yes, could indeed be that.  At the other end of the VPN you look like you are coming from the local area you are connecting to.  I will do it to watch UK sporting events that are limited to the UK despite living in the US as the VPN makes my connection appear as though it's coming from the UK where the VPN connection terminates.  Same with DJI and any other online services you access via VPN.  

Even on my work laptop right now because of a VPN client it shows that I am in North Carolina despite being in New Jersey.  I know because it is not 81 and sunny in New Jersey LOL!!!  

Edited April 27, 20224 yr by psolk

30 minutes ago, Watsi said:

We recently had to bring up our server capacaties too (weather, telex, etc.). Cause the number of users exceeded our expectations. Maybe MS is facing the same problem.

I've a bit of a difficulty to understand this. Looking upon the Steam Charts MSFS right now has the nearly exact same number of concurrent users as it had 2 months after launch (after the initial downloads settled). Now that's Steam alone but I don't see a reason why this result (i.e. constant number of concurrent players) shouldn't be extended to MS Store.

Thus, the only net additions were the X-Box users (who don't have access to all the flightsim.to goodies and quite a number of payware addons). Hard to tell if these make such a huge difference, but who knows.

Kind regards, Michael

Edited April 27, 20224 yr by pmb

Now when it disconnects briefly, all the scenery erases and turns blue. Even when it reconnects the scenery is still gone. This just started after yesterday's update. Before that, the scenery would remain even when if the server dropped briefly. 

This ruined my flight yesterday after never having this issue before. Rolling cache was set to 8gb. 

I haven't seen anyone with this issue, having a hard time believing I'm the only one?

Edited April 27, 20224 yr by NightOfDreams

11 hours ago, Adamski_NZ said:

It's been shown many times that the servers themselves cannot take the load. Packets not re-sent etc. Try Googling it - there have been a few somewhat technical posts, where users have run very comprehensive connectivity diagnostics.

As to in-between nodes, there's not really much I can do about that - but I do think Asobo should look at their in-sim ping time thresholds to allow for a little more leeway. I suspect the problem is merely ping time related and not actual throughput - but that's a wild guess.

I agree about the "sprained eyeballs" category (LOL) ... but it certainly does appear to be going that way! BTW - I've been simming since the very first flight sims (yes, I'm that old!).

Yeah, me too, since they were made by Sublogic. And this is the best sim yet despite all that's wrong with it. People are still comparing it to a fully decked out P3d with $1,000 worth of addons which is silly. Stock P3d can't hold a candle to this sim, and Asobo, as exasperating as some of their decisions have been, is a very active and engaged developer.

Bottom line is that the current state of the sim blows any other unmodified desktop sim out of the water, which is why we look askance at those who pretend it is somehow not a legitimate sim.

3 hours ago, pmb said:

I've a bit of a difficulty to understand this. Looking upon the Steam Charts MSFS right now has the nearly exact same number of concurrent users as it had 2 months after launch (after the initial downloads settled). Now that's Steam alone but I don't see a reason why this result (i.e. constant number of concurrent players) shouldn't be extended to MS Store.

Thus, the only net additions were the X-Box users (who don't have access to all the flightsim.to goodies and quite a number of payware addons). Hard to tell if these make such a huge difference, but who knows.

Kind regards, Michael

The Xbox user numbers are bound to have made some difference and that difference doesnt need to be huge to hit tipping point in server demand if insufficient spare capacity was factored in.

Not sure what your point about addon accessibility has to do with this?

I dont really know if it is a capacity problem but whatever it is when a game is provided largely as a service relying on connectivity, that connectivity needs to be very very  reliable

ITT:  A delightful mix of hubris, ignorance, and intolerance, from multiple quarters.  😄

First of all, as noted, it's not Asobo who manages "the servers", it's Microsoft.   So blaming Asobo for any shortcomings is unfair to start with.

Second, based on what I've read elsewhere, most of the data flows over the same network that MS's other gaming services use.   This may not totally be the case, I don't know if anyone outside MS knows, but they do use a lot of their existing gaming infrastructure for this sim.  So it's not just an MSFS problem.

(I know some folks have said in the past that they use AWS for bulk content delivery, but I just downloaded a big package from the content manager, and it was all over an IPv6 connection to a Microsoft server in Seattle, so that, at least, is not AWS).

Next:  No matter how fast your connection, there can be any number of bottlenecks between you and a remote server.   Even if your problems are recurrent, it doesn't necessarily mean that the other end is at fault.  That's just the fundamental nature of a packet-switched network.   Heck, in some situations you don't even need a particularly slow node.  At high enough data rates, and long enough distances, the speed of light combines with the need for TCP to acknowledge every single packet (although not necessarily individually) to mean that, due to round-trip times, there is a maximum possible theoretical speed to a connection between, say, Seattle and Perth that's probably a lot lower than you think it is, especially through copper.

But: That doesn't mean that everyone's performance problem is due to local or network conditions.   It's completely possible things can work just fine for you, but not for another person, with the problem still being at the server end.   As one potential example, assume a geographic load-balancer that directs your traffic to the east or west coast depending on where you are.  Based on the logic it uses and your geography, a person only miles from you could be shunted totally to the other side of the country from where you are.  If the system at that side of the country is struggling, it'll suck for them, but it's not their fault.   Saying "it works for me", unless as a very specific troubleshooting step, is generally not very helpful and can seem dismissive of other peoples genuine problems.

Now, the mention of load balancers brings me on to my next topic.    The folks who are saying "Microsoft doesn't have enough servers!" and those saying "it's not like they limit them to a fixed amount!" can both be part right and part wrong.   This is getting more theoretical and educated-guessy, because while I have a lot of experience in this area I know nothing about the precise details of Microsoft's system architecture.   But, I can talk about how I would design, and have designed, highly scalable content delivery systems.  With modern cloud tech it is extremely easy, with a little bit of knowledge, to build a system that can scale to a factor of 10, or even 100, in mere minutes.   But, unless you have a bottomless budget, cost restrictions can limit performance in two ways.

First is in absolute maximum capacity.  Even though you can go from 1 server to 100 in mere moments, you may not be able to afford having 100 systems running for any length of time, so you may limit your scaling capacity to, say, 50 systems.   If we say each system can support 100 people, then as long as you're one of the first 5000 people demanding access you should, for the most part, be able to get it.   But the 5001st person is going to run up against the problem that the scaling constraints simply won't allow the system to scale any more, so they're going to have to wait until someone else is done.  This can put up an immediate brick wall.  This is also a fairly easy situation to manage if you can afford to.

The second limit is sneakier, and that's the transitional capacity.  Put most simply:  you can't really start a server you don't know you need yet.   This means that, for any given increase in load, you will reach a point where current capacity runs out and you need to spin something else up.   This is very rarely an instantaneous thing.   Taking the example above, the initial configuration may have 25 of those 50 systems active at some point.   When the 2501st person tries to access things, the scalers will say, "aha, we need more!  start up a new server!", and a new one will shortly brought up.  However, that doesn't help Mr 2501, because he can't get in while that's still in progress.  So he may have delayed performance that Mr 2502 doesn't see if he comes along shortly thereafter.

(It's possible to minimize this effect by keeping a permanent over-provisioning, and saying something like "whenever we are at 90% of capacity, scale up", but the more of an over-provision you use the more any given amount of load costs you, so again, it comes down to budget.)

Anyway, as I say:  I don't know how MS's gaming platform works, but I would be very surprised if it didn't use a fair amount of cloud architecture with dynamic scaling, especially given how hard MS is using and pushing Azure.  But using The Cloud(tm) is not a panacea, as described above.   So yeah, they might not have enough servers at any particular moment when you try to access, but that doesn't necessarily mean they have too few overall.  You could have just caught things at a bad time.   Or during an especially busy time.

If I were to further speculate, I would imagine that the content delivery network, from which you download updates, is separate from the multiplayer network, where you connect for gaming.  Those two uses have two different optimal network configurations (with large block sizes and maximal throughput important for downloads, and low latency more important for gaming), so it's completely possible that one of those works fine and the other is totally saturated.

At any rate, this has gotten much, much longer than intended and I've probably lost everyone in my rambling.  😂  For anyone who's stuck along this far, thank you for that chunk of your attention, hopefully you've learned something.  If nothing else, if you're still here, the basic message is:  networking is complex, capacity planning is hard, and not every problem has an easy or obvious solution.  In fact, when it comes to tech, very few do, especially with a limited budget.  Try to be a bit more understanding of those who are having problems -- and of those who are not.  You could both be right.

But whatever you do, don't say it's Asobo's network.  That's just totally wrong.  😄

(My qualifications for talking about this stuff:  25+ years in system and network administration, going on a decade of cloud system architecture and management, and multiple mid to high level certifications from a famous cloud computing vendor.  I may not be an expert in many things, but I'm at least not completely talking out of my bum on this one.)

 

Edited April 27, 20224 yr by DeepestRed

why cant streaming data (eg PG) be offloaded regionally

Edited April 27, 20224 yr by icewater5